THEE EZ'ETEKMHMTEON 0:5 THE PROPER ELLUMENATEQN FOR THE AFP’LE PACKENG PLANT Thesis for fI'te Degree of M. S. MICHMAN STATE UNWERSITY Marvin Edward Heft. Jr. 1959 THLJL- LIBRARY Michigan State University T'— THE DETERMINATION OF THE PROPER ILLUMINATION . FOR THE APPLE PACKING PLANT BY Marvin Edward Heft, Jr. - A THESIS Submitted tothe-College of Agriculture and College of Engineering of Michiganfitate‘University in partial fulfillment of the requirements for-the degree of MASTER OF SCIENCE Department of Agricultural Engineering 1959 . ACKNOWLEDGMENT The: author wishes to express his appreciation and gratitude to the following: Professor D. E. Wiant, Agricultural Engineering. Department, under whose guidance and inspiration this study was conducted. Professor A. W.‘ Farrall, Head of Agricultural Engineering - Department, for arranging financial assistance for this project. Michigan State Highway Department for the use of their spectrophotometer and to M. H. Jansonfor his guidance in the operation of the instrument. The apple grower—packing plant Operators who cooperated in supplying information and the locations for trial installations. My wife, Margaret, for her encouragement and unending - assistance in the preparation of this thesis. ii " THE DETERMINATION OF THE PROPER ILLUMINATION . FOR THE APPLE PACKING PLANT BY MarvinEdward Heft, Jr. ' AN ABSTRACT . Submitted tothe: College of Agriculture and- College of Engineering of Michigan. State'University in partial fulfillment of the .requirements for the degree of MASTER. OF SCIENCE Department of Agricultural. Engineering -1959 p. ~50. W ABSTRACT Lighting of the apple grading operation requires a light source of the proper quality and quantity to enable the worker to detect defective apples with a minimum of time and effort and with maximum accuracy. The proper quality light source accentuates the maximum contrast between the normal apple surface and the defect. The proper . quantity of light will provide a minimum of 25 footlamberts of product brightness. The prOper quality light source can be determined by theory or by observation. The theoretical method involves the determination of the best coefficient of correlation of a two~variable correlation between the visibility function curve of the various light sources and the contrast between the visibility function of the normal and defective apple surfaces. The observational method is simply observing the apple and its defects under the various light sources. I The proper quantity of illumination can be determined only when the surface brightness of the apple surface is known. Surface brightness is obtained by an equation using the reflectance of the apple as obtained by the spectrOphotometer, the luminosity function, and the luminant energy output, or by a light meter measuring the incident light and the product brightness of the apple surface. A The average surface brightness of the-red apple was.§3.4 percent underathe deluxe cool white fluorescent lamp, the recommended Afluorescent lamp, and 24 percent under the lSO-watt incandescent filament lamp. This means that to obtain 25 footlamberts of product brightness 178 footcandles of lighting is required when a medium duty service factor of 60 percent is used. iv The required lighting of 178 footcandles on the surface of the apple is to be divided between general and supplemental lighting. General lighting should be at least 10 percent of the total, or 18 footcandles, to as sure eye comfort of the worker. Supplemental lighting should provide the remaining 160 footcandles of shadowfree glarefree light on the working surface of the apple grader. Fluorescent luminaires should be mounted in a continuous row three feet above the center line of the grader. The incandescent filament luminaires with ISO-watt reflector type lamps should be mounted three feet apart and four feet above the center line of the grader. The deluxe cool white fluorescent lamp was the unanimous choice of the west Michigan-apple packing plant owners in whose plants test installations of the recommended luminaires were made. The manager of the plant in which the ISO-watt incandescent filament reflector type lamp was used was well satisfied. TABLE OF CONTENTS Page INTRODUCTION ........................................... 1 Michigan and National Apple Production . . .............. 2 Scope of the Investigation ............................. 3 The Apple Grading Operation .......................... 4 Varieties Studied ..................................... 7 Defects Studied ....................................... 7 The Problem ........................................ 8 OBJECTIVE ............................................... 10 REVIEW OF LITERATURE ................................. . ll Illumination for Fruit Grading and Sorting ............... 11 Illumination for Grading Other Agricultural Products . . . . 12 Ulumination in Other Industries ........................ 13 Color Perception ..................................... 15 Illumination and Color Perception ...................... 16 Effects of Illumination on Man ......................... 17 Other Factors Affecting Grading Efficiency ............. 19 STUDY OF EXISTING APPLE PACKING PLANT LIGHTING. . . . 21 Grader Lighting .................................... . 21 General Lighting ................................... . 22 Color of Grader Belts ........... . .................... . 22 Packing Plant Construction .......................... . 22 PERCEPTIBILITY OF APPLES AND THEIR DEFECTS ........ 24 Spectral Reflectance Curves .......................... 24 Visibility Function ................................... 27 DETERMINATION OF THE APPLE PACKING PLANT ILLUMINATION ...................................... 3O Contrast and Artificial Illumination .................... 3O Selection of the Luminant ............................ . 31 Laboratory Determination of the Luminant .............. 34 Determination of the PrOper Illumination Level .......... 36 General Packing Plant Illumination ..................... 39 vi TABLE OF CONTENTS - Continued Page A Practical Method of Determination of Grading Illumination .................... q ............... 39 Effects of the Background on Seeing ................... 39 TEST LIGHTING INSTALLATIONS ......................... 43 Tungsten Filament Lamps ............................ 43 'Fluorescent Lamps .................................. 43 Results on Apple Quality ............................. 44 DESIGN OF THE ILLUMINATION SYSTEM .................. 46 General Illumination ................................. 46 Supplemental Illumination ........................... . 49 SUMMARY AND CONCLUSIONS ........................... . 52 APPENDICES ........................................... . 57 Appendix I - Spectral Reflectance Curves ............. . 57 Appendix II - Apple Packing Plant Survey Data ........ . 66 Appendix III --Relative Spectral Distribution of Fluorescent Lamps .............................. 73 REFERENCES ........................................... 75 III IV VI VII VIII IX XI LIST OF TABLES Page Michigan Apple Grade Requirements. .' ................. 9 Coefficient of Correlation for the Apple Surfaces and the Various Luminants .................................. 35 Surface Brightness of the Apple Surfaces Under Deluxe Cool White Fluorescent and ISO-Watt Incandescent Fila- ment Illumination ......................... . .......... 38 Correction Factors for use in Converting Magnesium Carbonate Standard Block to a Certified Magnesium Oxide Standard ...................................... 58 Corrected Spectral Reflectance Readings for Typical McIntosh and Double Red Delicious Surfaces ........... 59 Corrected Spectral Reflectance Readings for Typical Common Delicious and Spy Surface .................... 60 Corrected Spectral Reflectance Readings for Typical Jonathan and Golden Delicious Surfaces ................ 61 Corrected Spectral Reflectance Readings for Typical Apple Surface Defects ................ . ............. . 62 Average Corrected Spectral Reflectance Readings for Typical Average Apple Surface and Maximumv Defect Readings ........................................... 63 Summary of the Apple Packing Plant Survey ........... 69 Relative Energy Distribution of Some of the 40-Watt Fluorescent Lamps ............................ , ..... 74 viii LIST OF FIGURES FIGURE Page 1. Typical apple grader layout.. ........................... 5 2. Average Spectral reflectance curves for average normal and green apple surfaces and their maximum defects ..... 26 3. Relative spectral sensitivity for daylight adapted human eye .................................................. 28 4. Average red apple and its defects as seen by the eye, visibility function ..................................... 29 5. Relative visibility function of the red apple and its defects as related to the average green surface of the red apple. ............................................... 32 6. Relative visibility function of the contrast between the red apple and its defects as related to the visibility function of the average green surface ............................. 7. Light meter method of reflectance determination ........ 8. Mounting height of luminaire in packing room ........... 9. Luminaire location in packing room ........... _ . . i ...... 10. Location of incandescent filament luminaires over the apple grader ................................. . ....... 11. Location of fluorescent luminaires over the apple grader . 12. Location of luminaires when they cannot be located over the center of the grader ............................... l3. Spectral reflectance curves of the McIntosh apple samples ix 33 4O 48 48 50 50 51 65 INTRODUCTION _ Since the beginning of time man has worked and played in the light of the sun. . Since the beginning of time man has attempted to provide light so that he could extend his working and playing day and move his activities indoors and not be affected by darkness of night and inclement weather. Lighting has developed from a fire in the cave to today's wide assortment of lamps. Technology for the application of the developing light sources has kept pace with the lamp development. The application of this technology to specific applications has been neglected in many areas. It is the aim of this investigation to apply technology to the problem of apple sorting. ‘ Two of the problems that the Michigan apple grower faces in the marketing of his fruit are increasing packaging costs and the demand for a higher quality pack. Lighting has an effect on both. Consumers are demanding a better quality pack and will not purchase produce that is inferior in quality. The demand for a higher quality pack is influenced by the convenient-sized prepackaged products offered in the self-service supermarkets (22)? Modern grading and handling equipment has done much to reduce bruising and other damage .to the apple from harvest until it reaches the vendor's produce counter, resulting in a better quality pack. Research in materials handling has eliminated much labor from the apple marketing program. .Equipment such as fork lift trucks, *Refers to appended references. pallet boxes, and mechanical box dumpers, allows one man to do the work of many, thereby lowering the packaging and handling costs. Researchers have done much to aid the apple grower in reducing costs and improving the quality of the pack, but in all of this progress one factor has been forgotten, lighting. Can the worker at the apple grader see the defects? The human eye has a wide range of adaptation, but it does have limits. The demands on the eye are greater with the new equipment and higher quality requirements because of increased grader Speeds and capacities and the more critical seeing needed to detect the defective fruit. Researchers in pursuit of the answe rs to the apple grower's problems of increasing costs and the demand for higher quality have overlooked two facts; (a) eye strain caused by poor illumination causes fatigue and the lowering of worker efficiency; (is) as the time available for the detection of the defect decreases the amount of light necessary to detect the defect increases in order to maintain the same degree of ~ accuracy of sorting. Michigan and National Apple Production Apple productionis an important agricultural industry in Michigan ~ and the United States. In the period 1947-56 the average annual value of the‘Michigan- apple cropvwas approximately 13 million dollars. - Since then the crOphas been worth approximately 20 million dollars ‘ annually. The average. annual production in Michigan for the period - 1947-56 was more than 8 million bushels. The 1957 Michigan production was 10 million bushels and is estimated at 12 million for 1958. Michigan consistently ranks among the t0p five states in apple production. The 1958 estimate places Michigan in third place (23, 24). Nationally, the annual apple production averaged more than 100 million bushels forthe period 1947-56; the 1957 production was 118 million bushelsand the 1958 production-is estimated at 125 million bushels. The average annual value of the national apple cr0p for the period 1947-56 was 210 million dollars (23, 24). Approximately 80 percent of the apple crop is used for fresh fruit, which makes grading ~ an, important operation. The Jonathan is Michigan' 8 leading variety of apple, with an average annual production for the period 1946-54 of l. 5 million bushels; the McIntoch is second with l. 15 million bushels; the Spy, with 1 million bushels is third, and Redrlelicious is fourth, with 760 thousand bushels. The Golden Delicious, the leading yellow variety, has an average annual production of 110 thousand bushels (23). . Sc0pe of the Investigation The aim of this investigation is to determine the proper com- mercially available luminaire, or luminaires, that will allow the worker on the apple grader to detect the defective apples with the greatest practical Speedand accuracy and minimum of fatigue. In order to accomplish this, three factors must be considered: (a) the proper'level of illumination that will permit a maximum Speed of detection of defects, (b) the illumination of the apple surface so that the-maximum contrast will occur between the normal apple surface and the defect to be removed, and (c) the bringing of the entire apple surface ‘intothe field of vision of the worker. The third factor (c) will not be considered in this study because the majority of the modern apple packing plants rotate the fruit so that the entire surface of the apple is exposed to the field of vision of the worker. The Apple Grading Operation ’Ingeneral, all apple packing plants use the same basic operations and- equipment, but here the similarity ends. Because of the great variety in building design, equipment design, and plant requirements, every grader installation-is entirely different. . A simple layout of the basic operating sections is shown in Figure l. The first operation-in grading is dumping. Herethe apples are dumped ontothe grader either by hand or by a mechanical dumper. The worker at this point is required to remove the rotten apples before he feedsthe' applesintothe cleaning Operatf.<;~n.. This Operationis one requiring. supplemental illumination. From the dumping Operation the apples move to the cleaning operation. InMichigan apple packing plants, the cleaning is done by rotating brushes or rags. The rotating brushes, or rags, make the apples shine by polishing the cuticle, the natural wax layer onthe apple. Since cleaning is a. mechanical Operation, supplemental illumination is not required as}. the general packing plant illumination will be adequate for this operation. . After the cleaning Operation comes the pickout operation. Usually a series of rollers, perpendicular to the direction. of the apple flow, rotate the apples so. that the worker, or workers, stationed there can, view the entire apple surface and remove defective apples. Becausethe- purpose of the pickout Operationis the removal of defective _- apples, the illuminationof this section is of utmost importance. Next in line is the sizing Operation. Here the apples are sorted intorpredetermined sizes. This is a mechanical operation with the - sizing done by belts or chains or by weight. A Supplemental illumination is not required in. this area. “593.4 HmpmHU mammaw 303458 .H ondmwm “03:0 mngomnm $5.an @154 no “8300an UK + + J HMKWOEOMH HU ®MOQ T I ofiormh Bod.“ @3500 ~ _. Evade .msflomna pnm gawk. .msflodnm .. mewswm . udoxomnm MGEMOHU MQHQESQ [The final operation in- apple grading is performed at the packing -. table and- packing outlets. The packing, table may be either of two 7 types. The simplest type is a sloping table down which the apples roll it tothe packer. This type is little used because of lack of capacity. The second. type is a table of length needed for the production of the -. plant. It consists of counterflow belts that move the apples around the -‘table- until removed by the packers. Packers remove the apples through one of several methods. The oldest and simplest pack is the bulk pack, where the apples are-packed‘in. a crate or apple box. ~ Another method is bagging the v apples-into.consumer-«sized plastic bags. There. are several ways of transferring the apples from the table to the bag. ‘Withthe sloping table the: apples-are usually allowed to roll into the bag. With the counterflow table theapples are moved to the bag either by hand or by diverting rods that are placed in the stream of apples. In some cases filling the bagsis an. automatic Operation, withthe operator placing .the empty bagonto the bagger and removing the full one. There is also. the cell-pack method. The apples are removed from the packing .table, by hand and placed-int individual cells: in the shipping cartons. Inadditionto the packaging of the apples the workersin this area .makethe- final check onquality, asrit is the duty of the workers to » remove'any defectiveapples that have beenmissedin the pickout area. This areas studied in. this investigation. Many Michigan. apple packing plants have graders equipped with a belt for the removal of the defective apples. This belt is located over the: center of the grader and a few inches above the working surface, - aloeationthat- allowsthe workersto place the defectiveapples onto it with‘a minimum of time and effort. This location presents a lighting problem as the belt cuts off some of the overhead light causing shadows and low levels of illumination on the working surface, both of which reduces the efficiency of the packing operation. Varieties Studied The major emphasis of this study was the lighting of the red varieties of apples since they are the main varieties grownin Michigan. The yellow varieties were given some consideration, but the green varieties were not considered as they are used mainly for processing, _ and the grading operation is not as critical as i is for the fresh fruit market. - The major Michigan varieties, Jonathan, Spy, McIntosh, Common Delicious, andDouble Red- Delicious were used in this investigation. The GoldenzDelicious was the yellow variety studied. All samples used were selected by Dr. D. H- Dewey of the Horticulture Department, of Michi an. State Universit , as bein re resentative of the variet . g Y 8 P Y . Defects Studied _Defects studied fall into two general categories; surface defects and off-colored fruit. . Internal defects such as water core-and internal breakdown cannot be detected by surface illuminationand therefore are not included in this study. There were several surface defects studied: (a) insect damage, which can be anything from a small surface scar-to an Open worm. hole; (b) scab, a fungus growth on the surface; (c) breakdown (commonly called rot); (d) limb bruise, a darkening of the surface caused by rubbing by a tree limb; (e) bruise, a mechanical indentation of the surface that occurs at or after harvest and causes no change in surface color except as breakdownsetsin. There are other defects but they are similar to the studied defects and the results obtained will apply to all defects. In general the defects are darker than the apple surface. One exception is apple scab. Off-color apples are those apples that have not developed sufficient characteristic color to make the required grade requirements. State and federal grade laws set the tolerances of the percentage of defectiveapples in-a given grade. These laws also set the maximum . size of defects. and the minimum amount of characteristic colored area for the grade. Michigan grade requirements are shown inuTable I (25). The Problem The problem that this investigation confronted was the determination of the light source or sources that would providethe greatest difference invisibility between the normal apple surface and the defect. The proper 'light sources will make it'possible to obtain.the- maximum. possible grade with a minimum of labor. TABLEI MICHIGAN APPLE GRADE REQUIREMENTS (25) Limits of Defects in U. 5.. Apple Grades Color Requirements Grade 'DefeCt U233?” I683 21ml U311: Decay None None ‘ None Jonathan spot None None None Scald None None None Scab NOne 1/4" 3/4" Broken skin,bruises None, except slight ones caused by normal handling Limb bruises max. dia 1/8" 1/2” 1/10 of surface Ruxifax. dia. 1/8" 1/4" 3/4" Healed insect stings max. dia. 1/8“ 3/16" 1/4H Open worm holes None None None A , percent normal surface, U.S. Apple Grades Variety Grade ‘Extra Fancy - Fancy U.S. No. l Delicious, Golden 75% 75% 75% Delicious, Red 50 25 15 Jonathan 66 33 25 ‘Mclntosh 66 33 25 ’ Spy 50 25 15 10 OBJECTIVE In the preceding section it has been established that apple pro- duction is of importance to the economy of the state of Michigan and the nation, and that there is a need for the improvement of the lighting of the apple grading areas of the packing plant. The purpose of this study is to determine the prOper illumination for use in the apple grad- ing area of the packing plant, therefore the following objectives have been established: 1.. To determine the proper available light source, or sources, that will permit the removal of the defective apples with the maximum Speed and accuracy and with a minimum of effort. .. To determine the prOper level of illumination (footcandles) using commercially available luminaires that will permit the removal of the defective apple with maximum speed and accara-‘zy and with a minimum of efforts. 11 . REVIEW or LITERATURE Illumination for‘ Fruit Grading and. Sorting .Ditchman (1958) said the main purposein grading lighting is to light the defect to be detected in such a way as to make it stand out (maximum contrast) and at the same time light the things in the general field of'lvision' so as to avoid harsh adjustment for the workers' eyes. The general illumination should be at least 10 percent of the total illuminationif the workers' eyes are to be allowed to rest when he glances upward. If the background is dark the eyes tire quickly because they do not adjust easily to such marked contrasts. Peterson (1951.) reported on limited work on cherries: There'is no artifical light found that does everfihing in culling work. Some lights enhance the color of the product, others show up certain defects better, and it is found also that the color of belts influence the color perceived, and they are of primary impo r'tanc e . He also reported that the efficiency of the workers was influenced by the type of light, the background color, and the rotation of the fruit. . Additional work was done on cherries, as well as limited work on tomatoes, by Parker (1954), who obtained spectral reflectance curves of the mainrMichigan cherry varieties and their defects. He conducted tests on the effects of rotation of the fruit on the sorting efficiency. Parker, further reported that for. maximum perceptibility of color differences, the illuminant should radiate throughout the entire Spectrum-and be saturated in the area of the maximum difference between the reflectance curves of the fruit and those of the defects. The color of the background should be theaverage color of the products being . Separated. He did considerable investigation in the area of Special filters, mounted over standard fluorescent lamps for use in cherry sorting. lZ .Carrying this work further was-AlSOp and others (1958) working with the cherry processors to determine an illumination system using commercially available luminaires that would give the maximum efficiency of defect removal. Through a process of installing the various colors of "white" fluorescent lamps in luminaires, and varying the-level of illuminationand making personal observations, the team of observers came tothe conclusion that the pr0per fluorescent color tovuse was the deluxe cool white for the red varieties. A level of illumination of 150-200 footcandles was recommended by this team. Tungsten filament lighting twas- satisfactory with the same level of illumination as recommended for fluorescent lighting. Fortthe sorting of the dark cherries Alsop says, "For dark. sweet cherries, the white, frosted incandescent lamps were the only ones giving satisfactory light. The difference in ease of sorta ing with these lamps and any fluorescent was astonishing. " Numerous publications by various state colleges and the USDA on the subject of packing plant layout and equipment were reviewed and the only one that mentioned lighting was the Michigan State University Special Bulletin 417,. Equipment and Layout for FruitPacking Houses by H. P. Gaston-and J.. H. Levin, which recommends 100 footcandles at - all critical areas and 5-10 footcandles of general illumination. . Either incandescent or fluorescent lamps may be used. If fluorescent lamps are used the lamp color should be either white or cool white. - Accord- ing. to Levin. (1958), their recommendations were taken from a previous USDA bulletin and supplemented by common practice. Illuminationfor Grading Other Agricultural Products Probably most of the work on illumination for grading agricultural products hasbeen done on cotton. Nickerson (1946) says, ”Good lighting is so important for cotton classification that the cotton. industry goes to fc ll. 13 great lengths to provide'it in classification rooms. " It is further reported that the amount of light is not the only factor in a good cotton classification room, but good quality also is needed, and may be even more important than quantity. For cotton classificationworkit has been found that the best artificial light is one that. matches natural daylight. Nickerson further reported that fluorescent daylight lamps are widely used for general illumination, but that special care‘must be given when they are used for color grading. Fluorescent illumination provides a. large source of low brightness light with little heat and good diffusion, making it an ideal light source providing its color character- istics would make it a satisfactory substitute for daylight. , While Nickerson has done mo st of her work on. cotton, She reported (1956) that: . . . This relates to the importance of the Spectral energy distribution of lighting to be used in color work. . For cotton we have found that the combination of fluorescent and incandescent lighting described is quite satisfactory for classing the colors of cottons, most cottons being near-white. It does not follow that it will be equally successful for all agricultural products, nor that it will serve all purposes. A It may be entirely satis- factory for inspecting or sorting purposes, but unsatisfactory for initial grading of aproduct. Gould (1952) reported that color is not how you seeit, but how you light it. Illumination in Other Industries Thexllluminating Engineering Society in itsrecommendations for canneries (1950) goes into detail on the illuminationof the inspection and work areas of the cannery, and discusses the reasons for its recommendations. They report that high contrast between the defect and the immediate surround is necessary for accurate inspection. 14 If the contrast is poor, there are two ways to remedy the situation. One is to paint the field of vision to a reflectance equal to the product being inSpected, and the other is to increase the brightness of the product by increasing the illumination. The recommendation quotes -Crouch (1945) who says 90 percent of the maximum ability to see contrast is obtained when the brightness of the product and the surround is at 20 footlamberts. The recommendation goes on to say that there is a need for high visual acuity in. inspection tasks and an increase can be obtained by control of brightness. Crouch says that 80 percent of the maximum visual acuity occurs at 25 footlamberts of product brightness. ~ Speed of vision is also affected by the brightness level and, according to Crouch, the Speed of vision increases withan increase of product brightness. The size of the defect and the reflectance of the defect were varied in tests and in all cases as the brightness increased the speed of vision increased. Intactual illumination of the cannery, the recommendation calls for glarefree general illumination with the use of high intensity supple- mentary illumination at all critical seeing locations. The general illumination should be at least 10 percent of the level of the supplementary illumination. Luminaires should be located so that they provide shadow- free illumination at the working areas. Taylor (1942.) reported’that when radiant energy in the visible spectrum strikes an object, part of the energy is reflected at the surface with little or no change in its spectral character. The rest enters the product to a varying depth and is transmitted, absorbed, or diffused, causing the resultant color as seen. .According. to Taylor, the two dominant factors in producing small color differences are the thickness of the color layer and the background of the material. 15 In order to make a small color difference more easily detected, the ~ illumination for inspection should be such that it accentuates the color difference‘as much as possible. For the best results, the luminant ~ should radiate throughout the entire spectrum, but be rich in the - spectral areas where the colored object has a maximum absorption, as small changes-in these areas will often appreciably influence the color of the object and increase the contrast between the object and the defect. The opposite of this is found in the color printing industry. Linsday (1949) says: A When a colored material such as yellow printing ink which reflects everything but blue, is examined under its comple- mentary color (blue) it absorbs this blue light and, therefore, looks dark or black. When the yellow proof, then, is examined under saturated blue light, the solid areas of yellow appear black against a bright background because the white background reflects most of the blue light which strikes'it. The eye assumes that this higho-brightness blue is essentially white, which gives a visual effect of black on white. Area on the proof where the yellow is not solid will appear as a gray value whose density depends upon the amount of pigment present. . . . By varying this process through the three primary colored-inks used in printing, the quality of the print can be checked. With the use of white illumination, the appearance of the final colored print can fully be checked. Color Perception Color perception-is what we see apart from variations in time and space. It does not refer to size, shape, texture, gloss, transparency, of flicker, though it is influenced by all of these. Color is what is immediately responsible for color perception. It is one aspect of the light entering our eyes (18). It is a known fact that the eye adapts to a considerable extent to counteract changes in quality and quantity of different sources of l6 illumination but its powers of adaptation are not unlimited according to Helson (1952). The state of adaptation along with the spectral dis- tribution of energies reaching the eyes result in the color of the objects. . Evans (1949) reported that as a light source falls on a series of colored objects, each in succession, the observer sees each color as affected by the light source. It is impossible for the observer to predict the results of the next step from the color that he sees because he is dealing with a series of successive absorptions of energy. At each step the observer evaluates the resulting energy distribution and sees the resulting color, making it impossible to predict the color of the next step from the color seen. The only way that it can be done is to know the exact energy distribution of the source and the absorption properties of each colored object. In this respect Helson (1955) said that the Spectral distribution needs to be considered only for special problems or situations. It must be noted at this point that the main concern of the author is the effect of the various light sources on color perception and the color rendering pr0perties of the light sources. The principles by which the eye sees and the theory of color vision are well established and many texts on the subject are available (See Reference Nos. 2, 7, 18, and 2.9). Illumination and Color Perception The Spectral distribution of the energy radiated by a light source has a marked effect upon the color as seen by the eye. Helson (1955) said, "That the limits of color contrast are reachedin everyday experience is shown by the fact that the color of objects do not remain constant in different illumination and when various conditions of view- ing are changed. . . . " It is reported that as the illumination changes 17 from daylight to a strongly chromatic illumination, objects and their color losetheir individuality to a great extent and a single hue and - mood prevail. The hue of any sample is a function of the relation of its reflectanceto the adaption reflectance, which depends largely on the background. By its preponderating influence on the adaption level, the background can determine whether the object will be tinged with the illuminant hue, the contrasting hue, or be lost in saturation to the point of achromaticity. Work with tungsten-light by Hunt (1950) led to the conclusion thatmarked changes of saturation. occur on altering the adaptation conditions and leads to an explanation in terms of Spilling over of re3ponses of one. color into the responses of another and fits into other visual and physiological phenomena. Fluorescent lamps because of their wide range of colors present ‘ a slightly different problem in color rendition. Jerome (1953) reported that for the color of the test object to be properly rendered the fluorescent - luminant must emit energy throughout the visible Spectrum in order that the selectivity of the test object may properly alter the incident energy 1 and cause'proper color rendition in every wave band. He also said that the Spectral energy distribution of a light is the only complete - specification of its color rendering properties. . Effects of Illumination on Man Illumination can affect manr 3 performance and rate of fatigue. - Spectral quality of light cannot be neglected according to Simonson (1948) for it is asimportant as intensity and light distribution» as a .factor affecting visual performance and fatigue. It has been found that color effects. ontvisual. acuity is affected as the level of illumination changes. 18 One of the 'reasons for fatigue is reported by Weston (1954), who said that whenlighting conditions are subnormal for the» visual task, fatigue is. accelerated mainly because of the undue mental exertion necessary for interpretation and discrimination. "Deteriation of visual functions in the-course of prolonged work is an important and objective criterion for visual fatigue . . . , " according to Simonson(l952). He also reported that visual fatigue is not an-isolated process but affects the entire body. Blackwell (1958) reported that in a detailed study on illumination for seeing, he determined that the eye sees, or ”assimilates visual information" by bits, and that the level of illumination determines the number of as similations that the average eye can make in one second. He stated that with sufficient illumination the eye can make ten assimilations per second, which he considers a conservative estimate of the maximum visual capacity with today' 3 knowledge. Another-limiting factor is what Blackwell calls the ”response limitation concept, " which is made up of some sequence of input information, assimilation, and re3ponse. On this he said: Now, I believe this re3ponse limitations idea is very important because it does vary depending upon the task you are required to do. In ordinary reading for the fast reader there is no response limitation in my opinion. Instead you just let the information pour in and you get meaning from this impouring, highly complex informational input. Now, in certain assembly line jobs there is an obvious response limitation. You must do something about the information you extract and do it right then and there before you can extract more information. This puts a re3ponse limitation in the system. . However, response limitations are not introduced as often as we might think. I use to work on an assembly line as .a boy, picking peaches off a belt, and the way this was done, if you were‘any good at it, was that you didn't look at one peach, examine it, reach out and pick it off the belt if it wasn't any good. Instead, you let your eyes roam dynamically over the 19 whole belt and when you Spotted one that was bad you went on Spotting others while you grabbed the bad peach while viewing it out of the corner of your eye. Thus, even in the case, which is somewhat response-limited, the situation is not as reSponse limited as though one element of information were followed by one reSponse. Instead information continues to come in and the reSponses are limiting to an extent, that is you can't get too far behind or you will miss the bad peaches altogether, but there is not an item-by~item reSponse limitation. According to Blackwell, the main contrast with which to be con- cerned is brightness contrast, as it is more important than color con- strast. Usually all the color contrast available is worth less than 10 percent of brightness contrast. Another factor affecting the level of illumination is the location of the center of the eye in relation to the target. Blackwell goes on to say that if. the target is off center of the eye it is hard to detect but if it is in the center it is easy to detect. The center of the visual field of the eye is more sensitive than the other parts of the eye's visual field. One way to counteract this is to raise the illumination level of the entire field so that the outer regions of the eye's visual field have the required illumination for detection of the target. Other Factors Affecting Grading Efficiency Grading of apples and other fruits is a visual-manual operation in which the eye must detect the defective fruit and inform the brain, which in turn must activate the arm muscles into action to remove the defective fruit. Other factors that must be considered, in addition to illumination, are Speed of approach of the fruit, the direction of approach of the fruit, and the number of defects if the eye is to detect the defects and cause their removal. 20 ‘Intests conducted by Malcolm and DeGarmo (1953) it was found that the best sorting efficiency was obtained when four rows of Speci- mens, rotating at a rate of three times per foot of translation, and at a speed of three minutes per one thousand Specimens, passed the worker. They found further that the side approach resulted in slower sorting speeds than did the direct approach. The difference in the percent of defective Specimens to be removed had no effect on the sorting efficiency when the range of defects varied from 10 percent to 30 percent. When more than one type of defects were to be removed they found that slower sorting Speeds resulted. 21 - STUDY OF EXISTING APPLE PACKING PLANT LIGHTING A study of apple packing plants in Western Michigan revealed a widevariation in lighting. There was no definite trend toward fluorm ‘ escent lighting over tungsten filament lighting. Fifty percent of the ,twelvepackingplants-used fluorescent lamps over the grader, while 33-percent used incandescent filament lamps, and 17 percent used both. The reverse was true for general lighting of the packing room; 83. 5 percent used filament lamps, 8. 25 percent used fluorescent lamps, and 8. 25 percent used none. For detailed breakdown of this information and other survey data see Appendix II. Grader Lighting (The Size of lamp-over the grader varied from 20 to 96 watts for fluorescent lamps and from 75 to 200 for filament lamps. As for color of fluorescent lamps, six graders used cool white, four white, one daylight, and one a non- standard color. Nine of the installations surveyed used reflectors on all or part of the-luminaires. . Six fluorescent installations used industrial type fixtures which-include reflectors. Of the filament lampinstallations two used Shallow dome reflectors and one used PAR- 38 lamps. The level of illuminationat the critical grading areas varied widely. The level of illumination was not measured at each packing plant, but was estimated by the author to have ranged from a low of two orthree footcandlesto approximately 100 footcandles. While some plants had Spots of higher intensities in the critical areas, none averaged higher than 100 footcandles. 22 General Lighting The incandescent filament lamp size used for general lighting in the surveyed apple packing plants ranged from 40 to 200 watts. The fluorescent installation used a 96 watt lamp. One packing plant had no general lighting. Four of the plants surveyed used no reflectors on the luminaires, three used shallow dome reflectors, three used RLM reflectors, and one plant with low ceilings used recessed fixtures. The level of general illumination in the packing plants surveyed was estimated by the author to range from a low of no light to a maxi- mum of five footcandles. Color of Grader Belts The color and reflectance of the background upon which the fruit is viewed affects the visibility of the defects. In the plants surveyed five used black belts, four used white belts, two used tan belts, and one used green belts. Nine of the graders surveyed used rubberized canvas belts; two used a plastic material; two used metal chains and one grader used both rubberized canvas belts and chain. Packing Plant Construction The reflectance of the surrounds, (walls, ceilings, and floors) affect the illumination in an apple packing plant because it determines the amount of general illumination that is reflected from the surrounds toward the object that it is desired to see. Also, the reflectance of the surrounds affects the rate of fatigue. A high or low reflectance can accelerate the rate of fatigue, while a medium reflectance retards the rate of fatigue. 23 Seven of the packing plants surveyed had cement or cinder block walls and five had glazed tile. Of the plants using blocks, four were left natural and the rest were painted, one each of white, silver, and green. Four packing plants with glazed tile walls were yellow and one was brown. Six of the ceilings were constructed of wood, four of galvanized steel, one of plaster board, and one of building board. Four of the wooden ceilings were left natural and one was painted white. The ceilings constructed of other materials were all left natural. 24 PERCEPTIBILITY OF APPLES AND THEIR DEFECTS This section is concerned with the perceptibility of red apples as distinguished from their surface defects. For the analysis of the perceptibility difference it is necessary to analyze the Spectral reflectance curves of the apple surfaces and their surface defects. Spectral Reflectance Curves The Spectral reflectance curves were obtained by measurement of the percent reflectance from the apple or defects surfaces by means of a SpectrOphotometer. Readings were obtained in 10 Mu. (Millimicron) increments from 400 to 720 Mu. , the accepted range of vismn for the human eye. Ten spectral reflectance curves were obtained of the red sur- face, two each from McIntosh, Spy, Jonathan, Common Delicious, and Double Red Delicious. Seven curves were obtained of the green surface of the red apple, two each from McIntosh, Jonathan, and Common Delicious and one each from Double Red Delicious and Spy. Six curves were obtained of the defects on red apples. For the yellow variety, Golden Delicious, two curves were obtained of the yellow surface and one of the green surface. Data for all Spectral re- flectance curves is found in Appendix I. For investigational purposes, averages of the normal and green surfaces were obtained by averaging the readings at each wavelength. The resultant averages were used as the average normal and green surface curves that are used in the determinations that follow. The defective surfaces do not follow a pattern as does the normal and green surfaces. Furthermore, it is desired to investigate the effects of all surface defects and averaging would eliminate the 25 high reflecting surfaces. Therefore the maximum reflectance of the six defect curves at each wavelength reading was used to plot the defective surface curve. This curve is called the maximum defective surface curve. Everything below the maximum defective surface curve is considered a defect. Figure 2 shows the spectral reflectance curves for the average normal surfaces, average green surfaces, and the maximum defective surface. Before analyzing the spectral reflectance curves, the Spectral definition of the colors should be made. According to the General Electric Company (9) the accepted range of each color is: Violet 400-430 Mu. Blue 430-490 Green 490-560 Yellow 560-590 Orange 590-630 Red 630-720 The mature red apple contains very little color in the violet, blue, and green range. In the yellow range the amount of color starts increasing and continues the increase throughout the orange and red range to a maximum at the long wavelength end of the scale before passing beyond the visible range. The rise through this range is interrupted with a dip that bottoms at about 675 Mu. Norris (1958) explains that this dip is caused by a chlorophyll absorption band. The spectral reflectance curves for the average yellow sur- face, the average green SurfaCeS for the red and yellow varieties are similar to'that of the average red surface, except that there is more color in the green and yellow areas. The maximum defect Spectral reflectance curve Show that ‘most of the defects are dark and contain little color. Reflectance Pe rc ent 70 60 50 40 30 20 10‘ 26 Aver ge yellow surface, ‘ \ no mal Average yellow surf Average re é—h- surface, Avera ge red surf ce, nornfial‘? M ximum dJIect 450 500 550 600 650 700 Wavel ength- Millimic ron Figure 2. Average Spectral reflectance curves for average normal and green apple surfaces and their maxi- mum defects. 27 The waif-2-3 of the Spectral reflectance curves reveals an obcer 121111310191 more impo. tant than the color of the apple, the fact that the human eye does not see as does the Spectrophotometer. The maximum sensitivity of the human eye occurs at 555 Mu. (Figure 3), and the SpectrOphotometer has no maximum sensitivity but sees all wavelengths with equal sensitivity. Visibility Function Since tne nu man eye does not see as does the Spectrophotometer, t 19. next log-'11. v— 1 S1 ep was to obtain the Spectral reflectance curves as seen by the human eye. This was obtained by multiplying, wave- length by wavelength, the eye sensitivity (luminosity function) by the Spect r211 reflectanc of the desired curve. The resulting curves, {It called visibility f1ntc1-, for the three red apple surfaces are shown Ex...,rr11..a:;1«':1n of the vi. sisi" it Ly function curves Shows that instead of lookingi llu-e 12.1-th r.:_gi ln .1"l .- . ctral reflectance apple surface curves, the resulting curves look more like. the luminosity curve with the 6V1“Ppt‘.’;n of the v; ”but, function curve of the red apple surface. 413 K 1’1 . -; ' ,. I : 3 L\ L: . is \ ::é " (E r '1 d) m fl, / \ - - \ mo 45.13 560 5510 600 650 700 Wavel en gth- Millimic ron Figure 3. Relative Spectral sensitivity for daylight adapted human eye. ‘ T . . __...‘..Cf.r.# 53,: —.u..-mw«; ....> Visibility Function 29 50 / \4—— Green surface 10 Maximum def¥t ——-—-—> \\ / r //\\ / / 2a su Haze-9k. o / . 4-00 450 500’ 550 600 650 700 Wavelength Millimicron Figure 4; Average red apple and its defects as seen by the eye, visibility function. 30 DETERMENATION OF THE APPLE PACKING PLANT ILLUMINATION The basic reqzzirements of the desired luminant have been set forth .-. in t} e preceding section. The objectives of this section are the determination of the proper luminant or' luminants and the level of illumination for use inapple grading. Contrast and email Illumination Tine works 3'. on the grading line, must see and. remove the defective aWJw, With. a minimum of effort and time. Thus the be so. :h‘ .'-;at it creates the maximum difference, or c ntrast, bet my een $2.1“. «Le surface brightness of the desired apple urfact. and its detects. .To brews the Ir asimum contrast the luminant used in apple grading must be on“ that its predominant wavelengths occurs in the area of maximum contrast. . A study of luminant relative Spectral energy distribution curves reveah. =' t}. at the various light sources have (a) different peak waveiengths, (is) different saturation, and (c) different amounts of energy radiated at the various wavelengths. . For therelative energy distribution data of some typical luminants, see AppendixIII. In the determination of the proper luminant for use in apple grading, the relative values of the surfaces and contrasts, related to a. common value, must be known. This makes it possible to properly weigh the relative merits of the three surfaces and the two contrasts under consideration. .The maximum value of the ' average green surface, occurring at 560 Mu. , was used as the value to which all of the curves are related. - Figure 5 shows the 31 relationships of the three apple surfaces under consideration. The relationship of. the two contrast curves is shown in Figure 6. The contrast data is the mathematical difference between the visibility function. curves of the two surfaces under consideration. It must be remembered that all of the values are for surfaces as viewed under an a achromatic luminant, that is, one that has the same relative values throughout the entire visible Spectrum. Since the least contrast occurs between the red surface and the maximum defect surface, the luminant used should favor this area more than the contrast between the red and green surfaces. Selection. of the Luminant The selected luminant must emit energy throughout the entire visible spectrum yet be rich in the area of maximum contrast. Maximum contrast in the area of the red surface-maximum defect contrast is essential. A luminant that radiates a major portion of its energy in this region will give the maximum luminous flux (reflected spectral energy) in the desired area. The maximum luminous flux results when the coefficient of correlation between the visibility function of the contrast curve and the visibility function of the light source energy curve is a maximum. The coefficient of correlation was determined by the equation for two variable correlation: NZXY - (2X) (EY) r7?— N:>:x"‘—(2X)a NEYz-(EY)Z where r is the coefficient of correlation, N is the number of pairs of variables, X is the contrast between the apple surface and defect Visibility functions Y is the luminant energy visibility function. in: ility Function "n r ‘ "‘ PC, .L’f. 8TH; q: DLL. R'eiative ‘V 32 CG ; ’ 9C; /l\ / \ / \ verage green ".79 - m— 60 l 50 / 40V ]/ \ _ Maxim fact \ _ .50 / \ 20 ‘ Aver e red 10 surfac 0 _ , - 40'!) 450 500 550 600 650 700 Wavelength- Millimicron Figure 5. Relative visibility functionrof the red apple and its ,defects as related to the average green. surface of the red apple. . .Av— Function ‘1 Re lat i ve Vi 2-, .T. .'-.. ,f; 5,- Pu L «and 3“!- 33 1‘50 8 \4—Fed-green contrast v 70 / \ ”:6 I \ 4C“: / \ 37.- / \ 2.91:"; / _ V 10 / . 4- Red surface - fhixi‘un: defect I contrast 0 400 450 500 550 600 650 700 Wavelength- Millimicron Figure 6. Relative visibility function of the contrast between - the red apple and its defects‘as related to the visibility function of the average green surface. 34 The resulting coefficient of correlations are shown in Table II for the various luminants. The nonwhite fluorescent luminants cause undesirable psychological effects on the workers and can be eliminated. The main lighting problem in grading of apples is the lighting of the least contrast, or that between the red surface and the maxi- mum defect. The best coefficient of correlation was 0. 96 and was obtained by using the deluxe cool white fluorescent lamp. Perfect correlation is 1. 0. The deluxe cool white fluorescent lamp does not have the highest coefficient of correlation with the red~green contrast, but since the red-green contrast is relatively greater than the red surface-maximum defect contrast, the deluxe cool white fluorescent lamp is the one to use in grading apples. The ISO-watt incandescent filament lamp also is recommended for use in grading red apples because of high coefficient of corre- lation for the two red contrasts. Since the yellow apple is relatively unimportant in Michigan as compared to the red varieties, it is recommended that the Michigan apple packing plants equip their graders with lighting as recommended for the red apple. Laboratory Determination of the Luminant .In the previous section the selection of the pr0per luminant was discussed from the theoretical point of view. Observational laboratory investigations were conducted to test the theoretical conclusions before making trial installations in apple packing plants. Under the illumination of the deluxe cool white fluorescent lamp apple samples appeared slightly duller in color than when viewed in natural daylight. The dark defects, such as limb bruise TABLE II COEFFICIENT OF CORRELATION FOR THE APPLE SURFACES AND THE VARIOUS LUMINANTS Luminant Fluo re sc ent I“: :nde S .. Contrast cen -Day-. Deluxe Deluxe White. Cool ~ Warm 150- light cool warm white white watt white white Redumaxi- mum defect 0°90 0.96 0.83 0.91 0.94 0.90 0.93 ' Redn green 0.86 0.85 0.82 0.88 0.87 0.80 0.88 Yellow~ max..defect 0.90 0.83 0.65 0.76 0.84 0.75 0.65 Yellow- green 0.41 0.37 0.17 0.29 0.53 0.32 0.58 and wormv damage appeared distinctly. The green surfaces appeared . slightly yellow instead of green, but were readily visible. - The daylight fluorescent luminant caused the apple to appear dark. Green surfaces appeared very light, almost devoid of color. -Dark defects were difficult to detect because of the dark appearance of the entire surface. .The deluxe warm white fluorescent luminant made the apple look more as it does in daylight, and the green surfaces were readily visible. However, the dark defects were harder to see as they assumed a red hue. ,When the apple samples were viewed under the nonwhite fluorescent luminants, the results varied with the luminant. Under 36 the pink fluorescent the dark defects stood out boldly on the red surface but the green surface all but disappeared. When viewed under the green luminant, the green surface stood out very dis- tinctly on a dark background which the red surface assumed, and the dark defects were not visible. Under the blue fluorescent luminant, the red surface assumed a blue color. The green surface absorbed little of the blue color and was still visible. The dark defects were completely invisible. .When Viewed under a 150uwatt reflector type tungsten lamp the entire surface of the apple appeared much brighter than it did under fluorescent illumination, even though an attempt was made to provide the same level of illumination. The red surface appeared very bright and red, the green surface was readily visible, and the dark defects assumed a reddish hue. . ter this series of laboratory investigations, there was little doubt but that the deluxe cool white fluorescent lamp was the pr0per fluorescent luminant for apple grading. The incandescent filament lurnimnt was a good second source. Determination of the Preper Illumination Level Now that the desired luminants have been established, there remained the determination of the proper level of illumination. The level of illumination must be such that workers on the apple grading operation can easily and readily detect the defective fruit and remove it as it passes. For lighting canneries the Illuminating Engineering Society (15) recommends a) that for 90 percent of the ability to see contrast, the human eye needs 20 footlamberts of product bright- ness, or adds also b) that to get 80 percent of the maximum visual acuity a product brightness of 25 footlamberts is needed. 37 The product brightness, or surface reflectance, of the apple surfaces was determined by the following equation (16): EUXKXR)‘ ZUxK)‘ R: where R is the desired product brightness, U)‘ is the relative energy of the luminant for wavelength x. K)‘ is the eye sensitivity for wavelength x. Rx is the reflectance as determined by the SpectrOphotometer for wavelength X. The resulting surface brightness of the various surfaces is shown in Table III. . From Table I it is found that the minimum amount of normal red surface is 50 percent for the maximum grade for the red varieties so the average of the red and green. surfaces can be used to determine the surface brightness for use in determing the illumination require-- ments for grading red apples. This means that the average surface brightness is 23. 4 percent. The minimum amount of yellow surface is 75 percent for the yellow varieties but since the surface bright- ness is higher than the average red surfaces and the recommendation is to equip the Michigan apple graders for red apples, theillumination level will not be determined as it is lower than for red apples. To obtain 80 percent of the maximum visual acuity, or 25 foot- 1amberts, the luminant must supply 107 footcandles of illumination, 25 673:;— = 107. Due to the operating characteristics of lamps and the possi- bility of accumulating dirt on the lamp the determined light output must be increased by a service factor. The service factor varies because of the amounts of dirt in the area of the lamp, room 38 TAB LE III SURFACE BRIGHTNESS OF THE APPLE SURFACES UNDER DELUXE COOL WHITE FLUORESCENT AND ISO-WATT INCANDESCENT FILAMENT ILLUMINATION Luminant Surface Deluxe Cool White lSO-Watt Fluorescent Incandescent (percent) (percent) Red normal 7.6 9.1 green 39. 2 38. 9 . normal 51.0 55.8 Yellow {green 47. 7 47. 5 Maximum defect l 1. 2 ll . 5 Avera e red 23. 4 24. 0 g yellow 49.4 51.6 conditions in which the lamp Operates, and the operating conditions of the lamp. A medium duty service factor of 60 percent was selected. This means that the luminant must supply 178 footcandles to fulfill the stated requirements necessary for proper vision in grad- , 52g— 2: 178. Therefore 178 footcandles of illumination on the apple surface is recommended when grading ing of red apples red apples . 39 Gene ral Packing. Plant Illumination The recommended 178 footcandles of illumination on the red apple surface in the critical apple grading areas is to be a combi— nation of general and supplemental illumination. To minimize the eye strain of the workers the general illumination should be at least 10 percent of that of the supplemental illumination (15). This means that the general illumination should be at least 18 footcandles. Therefore 160 footcandles of supplemental illumination on the sur» face of the red apple is recommended for the grading Operation. A Practical Method of Determination of Grading Illumination As the investigation into the required illumination for apple grading progressed, it became evident that there was a simplified, practical method of determining the required illumination. In fact, this simplified practical method will work for all grading and inspection work where detection of a different colored defect on a surface is required. The first step in this method is to observethe desired surface and its defective surfaces luminated with the various luminants and determine which luminant gives the greatest contrast. This is the procedure that was followed in the laboratory investigations dis- cussed earlier. Once the desired luminant is determined, this luminant is used to determine the reflectance of the product. With the desired luminant radiating its energy onto the surface of the object, a light meter, placed about one foot from the object, measures the incident light and the product brightness (reflected light) (Figure 7). The ratio of the product brightness and the incident light gives the reflectance of the surface: 40 Light source Light meter Fruit surface Figure 7. Light meter method of reflectance determination. 41 Product B rightne s s 100 2 . t fl *t' . Light onto Surface x Percen Re ek’ ance To obtain the desired level of illumination necessary to obtain 25 footlamberts of product brightness it is necessary to divide the 25 footlamberts by the reflectance and the service factor: 25 (Reflectance) (Service Factor.) = Required Level of Illumination. One advantage of this practical method is that all that is re- quired is an assortment of luminants and a light Ineter to determine the luminant and the required level of illumination for any grading task° no 5 ectro hotometer is needed. ’ . Effects of the Background on Seeing The effects of background on seeing was investigated in the laboratory using the selected luminants at the recommended level of illumination. These tests revealed that for maximum ease of detection of the defects, the background surrounding the apple should have a color and reflectance about equal to that of the apple. When this condition exists the apple appears to blend into the background, making the defect the main point of interest for the eye. . When the background was darker than the apple the main point of interest was the entire apple and the dark defects appeared to blend into the apple. At the same time the apple appeared darker than when viewed under a background similar in color to that of the - apple. The green surface was readily visible under the dark back- ground. 42 Viewing the apple on a background, lighter and with a higher reflectance than that of the apple, caused the apple to become the main. point of interest and the defects, both dark and light, seemed to be less visible. 43 TEST LIGHTING INSTALLATIONS The ideal commercially available luminant for use in apple sorting has been determined by theoretical and laboratory means. .To prove the selection, actual lighting installations were made in apple packing plants. During the tests red apples were being sorted. Tungsten Filament Lamps Early in the investigation a packing plant manager with a new plant requested help in designing the lighting system. Since there was little available information at the time on the prOper fluorescent color, it was decide to install 150-watt PAR-38 filament lamps be- cause of their builtuin reflectors. The lamps were mounted four feet apar’r and four feet above the working surface. This furnished approximately 250 footcandles of light on the working surface di» rectly under the lamp and about 50 footcandles halfway between the lamps. The workers commented on how much easier it was to see than before. The temporary system consisted of two bare ZOO—watt filament lamps mounted about 10 feet above the working surface. Talks with the workers revealed that there was no adverse heat problem and no reflected glare. . The plant manager was so well satisfied that when deluxe cool white fluorescent lamps were sug- gested for part of the grading operation, he was not interested. Fluore 5 cent Lamps Laboratory work with the fluorescent lamps revealed that the standard lamp would not produce sufficient light for apple grading. High output fluorescent lamps were selected as they produced the 44 greatest amount of light of the desired color from among the various types of fluorescent lamps commercially available. In ad- dition to, the deluxe cool white lamps, daylight and deluxe warm white were obtained. . In some packing plants the fluorescent fixtures could not be installed at the pr0per height to obtain the desired level of illumi- nation. These plants were packing the apples in cell packs. In order to provide the most convenient location for the cell dividers it was necessary to locate them over the grader. -This meant that the luminaires had to be located higher then the desired three feet ‘ above the grader working surface. It was necessary to locate them about five feet above the working surface and this resulted in obtaining 130 footcandles of illumination instead of the desired 160 footcandles of supplementary illumination. The owners and managers of the plants in which the fluorescent lights were installed were enthusiastic about the lamps and were unanimous in their selection of the deluxe cool white lamp as the one that gives the greatest contrast and ease of removal of defects. One owner said that now he was able to see and remove the scald on apples leaving controlled atmOSphere storage, something he could not do when using other luminants. The workers reported less shadow and the defects were easier to see. . Results on Apple Quality The packing plant owners and managers who cooperated were more concerned about quality of the apple pack than improved labor efficiency. .In all cases the manager or owner reported that the recom- mended lighting resulted in a better pack. The workers could see the defective fruit and remove it. This was readily noted by the increase in the amount of defective fruit removed during the day. 45 46 ‘ DESIGN OF THE ILLUMINATION SYSTEM The illumination requirements have been determined for use in apple packing plants. The next step is to design an illumination system that will give the desired 178 footcandles of illumination on the working surface of the grader. ~At least 10 percent of this is to be provided in general illumination. The luminant to be used - will be either the deluxe cool white fluorescent or the incandescent filament lamp. .The following installation details are based on the assumption that the walls of the packing plant are painted a neutral pastel color, having a reflectance of at least 30 percent, .and that the ceiling is white or a neutral pastel. It must be remembered that the recommendations in this report are minimums and additional illumination improves the visual efficiency. The formula used, as well as much of the other technical information used in the deve10pment of the design procedure, was obtained from the Westinghouse Electric Corporation's: Lighting Handbo ok. General Illumination The first step in the design of the general illumination system is to select the type of luminaires to be used, either industrial fluorescent luminaire or filament lamp equipped with RLM dome reflector. VA reflector is a must to obtain the maximum efficiency from the lamp. Next, the number of luminaires needed in the packing room is determined. . This is determined by the following steps: 1. Lo ¢ 6. 47 Determine the mounting height of the luminaire above the floor (Figure 8‘». Allow sufficient room for the fork lift truck to Operate. Divide the width f the packing room by the mounting height of the luminaire to determine the number of rows of luminaires needed. If the remainder is greater than oneuthird of the height, add another row. Divide the length of the packing room by the mounting height of the luminaires to determine the number of luminaires in each row. If the remainder is over one— q 3’ U "if of the mounting height, add another luminaire. Multiply the number of rows by the number of luminaires in each row to determine the total number of luminaires. Determ..’.ne l';;.;n‘.in;;‘=.ir.=: location as shown in Figure 9. De.'.':e;:'.nir.e the lamp size by the following equation: Fcctsandles x Area of Room ._ .. I'- Number of Lamps x O. 36 '2 Lumens per Lamp ere {1) Foctcandles equals 1.8 in apple packing plants. (2.) Area of room is its length times its width. (3) Number of lamps equals the number of luminaires times the number of lamps per luminaire. (4) O. 36 is a constant based upon the room size, lamp mounting height, and the reflectance of the walls, ceilings and floor as determined from an average of the packing plants surveyed and the service factor of the luminaires. 48 1 O A h The same measurement is used for incandescent filament . .. . . . 1am and RLM reflector. h :1: Mounting height p - F1001? V Figure 8. Mounting height of luminaire in packing room. 3 r YD I. Fluorescent luminaires to be located the same as incandescent filament luminaires Figure 9. Luminaire location in packing room. 49 (5) Lumens per Lamp is used to determine the lamp wattage. This is obtained from the latest manu- facturer's lamp catalog. If the lumens are more than 10 percent greaterthan the nearest size, select the next la rger size. . Supplemental Illumination The supplemental illumination must provide 160 footcandles of shacowfrezz, giz-irefree illumination on the critical working sur— faces T.--e critical. surfac s are the dumping section, the pickout section and the packing section (Figure l). .The incandescent filament luminaire over the grader should be mounted four feet above the working surface, with a spacing of three feet between the luminaires (Figure 10). Only lSO-watt re— flector type flood lamps are recommended. The type A incandescent iarnp equipped with RLM dome reflector is not recommended because of its low efficiency as compared with the reflector type lamp. The fluorescent luminaire should. be of the industrial type, .Inounted in a continous'row over the grader three feet above the working surface (Figure 11). The luminaires should be equipped with high output deluxe cool white fluorescent lamps. In all cases the luminaires should be mounted over the center line of the grader. If obstacles prevent this, such as a defect removal belt, two rows of luminaires must be installed (Figure 12). These luminaires should be installed two feet from the center line on either side of the grader. 50 47 "w }——~+ ---1.-—--— —n~__§—- Critical Areas No Scale Figure 10. Location cf incandescent filament luminaires over the apple grader. m J ——J ,1“ . A a " l ! ”it ,,___.__ ____/_._‘___‘_____} U Critical Areas No Scale Figure 11. Location of fluorescent luminaires over the apple grader. 51 Inc ande sc ent Fluo r e s c ent I“ H‘rl Obstacle No Scale Figure 12. Location of luminaires when they cannot be located over the center of the grader. 52 SUMMARY AND CONCLUSIONS Two of the problems the Michigan apple packerfaces in the marketing of his fruit are increasingpackaging costs and increas- irg demand for a better quality pack. . A survey revealed that many Michigan apple packing plants are hampered with poor lighting. . Poor lighting makes it difficult for the worker to see the defective fruit and thereby affects aive rse'iy both the packaging costs and the quality of the pack. Good lighting will help correct these adverse conditions. Gccd Lighting means a more efficient worker and a better quality pack. Good lighting in the packing plant is lighting of the pr0per quality and quantity. Light has proper quality when it accentuates the marirnurn ccrstrast between the defect and the normal surface 3f the apple. Proper quantity of light is the level of illumination that will permit the worker to detect defective applesin the minimum of time. . The desired light source to use in grading apples must emit energy throughout the visible spectrum, and be rich inthe area of maximum contrast between the defective and normal apple surfaces. ~ To select the proper light source, it is necessary to study first, the reflectance of the apple surfaces, andsecond, the relative energy distribution curves of the various light sources made avail- able by the lampmanufacturers. To determine the reflective properties of the apple surface, ten spectral reflectance curves of normal apple surfaces and seven of green surfaces were obtained with a Beckman spectro- photometer. .Six curves of defects were obtained. 53 Averages of the normal and green surfac cse sfor the red and yellow varieties were used in the determinations. Since it is de :51 red to see all the defe set S, the maximum readings of the six defects at each wavelength were used. . This curve is called the maximum defect curve .. An interest- 'r‘" ng and important observation was noted from these reflectance curves. The yellow apple is redder according to the spectrophotometer than is the red apple. The spectropho-- nmetei has equafi. sensitivity throughout the entire visible spectrum, 3 While the human eye“ sensitivity rea: es a maximum at 555 Mu. and. approaches zero at either end of the visible Spectrum. The. visibility f'a;:::.:.:'t:'.on, or the spectral reflectance curve as the eye sees it un 'er ant:rornza.i::'.c light, was arrived at by multiply-- ing, wavelength by wavelengt '11, t're spectral reflectance curve by the l'uLmin'psitjzr func'rtien (eye .f'en.:esi';;i"v: :, curve), The centres t between the normal surface and the green sur- face or the arm 5: ve surface is the mathematical difference between the "is; bility f notion cur ves of the two surfaces. The least contrast was the u’)‘it'."‘..‘ between the _-ormal red surface and the maximum defect as determined by relating the contrast curves to a common factor. . A light source that radiates the major portion of .its energy in the region of maximum contrast will give the maxi- mum luminous flux (reflected Spectral energy) in the desired areas. The maximum luminous flux results when the coefficient of correlation for a two--variable correlation, between the contrast curve and the visibility function of the light source energy curve is a maximum. 54 Of the "white" fluorescent lamps, the deluxe cool white lamp had the best correlation (0. 96) for the correlation between the least contrast and the light sources. Perfect correlation is l. O. The 150~watt incandescent filament lamp with a correlation of 0. 93 was acceptable. Observational tests were performed in the laboratory to verify the results of the theoretical conclusions. The nonwhite fluorescent lamps cause undesirable psychological effects on the worker making these lar ups of little value in the grading of apples. Of the more desirable "white" fluorescent lamps the deluxe cool white lamp incremwa the visibility of the defect because the defects do not take on the red hue they do under the deluxe warm wi‘ufte lamp. The daylight lamp darkens the entire red surface and causes the dark: defects to be difficult to detect. The green surface was readily visible under these lamps. The incandescent filament lamp causes dark defects to take or. a reddish hue .. The green surface is readily visible. The deluxe cool white fluorescent lamp was considered the best for apple sorting. The l.50---watt incandescent filament lamp is acceptable. The apple, when viewed on. a similar colored background, tends to blend into the background and the defect becomes the main point of interest for the eye. When viewed on a background darker or lighter than the apple, the apple itself became the main point of interest of the eye and the defect becomes harder to see. A simplified practical method was developed to determine the lighting requirements for the apple grading areas, as well as for any other type of grading or sorting where it is desired to detect a defect on a normal surface. The equipment consists of a 55 light meter, an assortment of light sources, and lighting fixtures. The procedure consists of viewzng the apple and its defects under the various light sources and determining the one that gives the maximum contrast between the defect and the normal surface. Then using the sele “ ed li gl- t source, measure the incident light and the product brightness of the apple surface with a light meter The light meter should be held as close as possible to the normal surface, preferably no further away than one foot. The ratio of fit. etwo re...u..r.gs g- ves the re le ectan UCE‘. of the surface. A prod-nut l2: -shre: s of 3.5 footlamberts is needed to obtain 80 percent of ma: {Ernurru visual a3t‘ it“; a: .:.d W-. ll give more than 90 per rent sf maximum ability to see com;- .322; t. The averag. prod itct origin .ss of the red apple under the delu; its c. ool wh: at -Icnresuent lamp was 23. 4. percent and 24 percent “under T”? watt 5.: .. a2‘..2.d.=‘.scer.t filament lamp. To obtain 25 footlam- " .3537." ts of Dr'sd..2'* "":.31.Ll12'l(:':82:‘ 1782.“ it ‘ ctcandles of lamp output on the surface of the apple is needed. when a medium duty service factor of 6%} percent: .:5 inst oduce .-.. The same ::‘ec:.crnmendation was made for yellow varieties, even though t1:.ey can use less light because of higher reflectance, because of the relative unimportance of the yellow apple in Michigan The 178 footcandles of light is to be provided by a combination of general packing plant lighting and supplemental lighting over the critical areas. To provide light that will cause a minimum of eye strain, the general. illumination should be at least 10 percent of the total requirement. For the red apples the general lighting should be at least 18 foot candles and the supplementary lighting 160 footcandles. 56 The design procedure for designing the general lighting system was developed. The supplemental lighting for apple grading should provide a shadowfree, glarefree light on the working surface, and be of sufficient intensity to provide the required product bright— ness. For apple grading, fluorescent luminaires should be mounted in a continuous row over the critical areas and at a height of three feet above the working surface. The fluorescent luminaires are to be equipped with high output deluxe cool white lamps to obtain the required footcandle level. Incandescent filament lamps should be mounted four feet above the working surface and Spaced three feet apart. ..n all cases, the luminaires should be mounted over the center line of the working surface if possible. If obstacles prevent this, two rows of luminaires must be installed. These luminaires should be installed two feet from either side. of center line of the grader. The. deluxe cool white fluorescent lamp was the unanimous choice of the west Michigan apple packing plant owners and managers in whose plants test installations were made. After comparing the results obtained with several fluorescent lamp colors they all selected the deluxe cool white fluorescent lamp. The manager of the plant in which the lSO—watt incandescent filament reflector type lamp was installed was well satisfied with the results. APPENDICES 57 APPENDIX I SPECTRAL REFLECTANCE CURVES The Spectral reflectance curves used in this investigation were obtained with a Beckman-Model DU Spectr0photometer, which is owned by the Michigan State Highway Department and is located in their research laboratory on Michigan State University campus. This instrument is equipped with a Model 2580 reflectance attach-u ment. Standard Operating procedure was followed (3). For a standard, magnesium carbonate was used and the resetting readings were corrected to a certified magnesium oxide block. The correction factors are shown in Table IV. Since the certified block was calibrated in 20 Mu. increments and the read«- ings were taken every 10 Mu. , the correction factors were inter-- polated for the intermediate figures. The sample consisted of a slab of apple of sufficient size to completely fill the Opening in the sample drawer. At least one— efghth of an inch of £188.11 was left under the epidermal layer so as to detect any color reflectance that might come from the flesh. . An attempt was made to get the sample to lie flat without wrinkling of the surface. The slight curvature that remained was not felt to have affected the results as the instruction book (3) says that samples with diameters as small as one inchcan be used. All samples used had a diameter greater than one inch. The readings were taken every 10 Mu. throughout the entire length of the accepted visible Spectrum, from 400 to 720 Mu. The corrected reflectances for all varieties measured and the defects measure are found in Tables V through VIII. Table IX gives the average readings used in all calculations in the investigation. 58 TABLEIV CORRECTION FACTORS FOR USE IN CONVERTING MAGNESIUM CARBONATE STANDARD BLOCK TO A CERTIFIED MAGNESIUM OXIDE STANDARD Wavelength. Correction Mu. Factor 400 0.911 410 0.917* -433 0.922 43$ 0.928* 440 0.934 45% 0 . 935* 460 0.936 47% 0.942* 480 0.948 493 0.950* 533 0.951 510 0.954* 52s 0.957 530 0.960* 540 0.962 550 ‘ 0.964* 560 0.966 570 0.969* 580 0.973 590 0.973* 600 0.973 610 0.975* 620 0.976 630 0.978* 640 0.980 650 0.980* 660 0.980 670 0.980* 680 0.979 690 0.981* 700 0.983 710 0.983* 720 0.983* a9: Inte rpolated value 3 59 TABLE V CORRECTED SPECTRAL REFLECTANCE READINGS FOR TYPICAL McINTOSH AND DOUBLE RED DELICIOUS SURFACES Variety McIntosh 1 Double Red Delicious Sample No. l 2* 3 4* l 2* 3 Wave- Per Per Per Per Per Per Per length Cent Cent Cent Cent Cent Cent Cent 'Mu. 400 1.79 1.2.7 1.41 6.95 2.04 2.64 2.39 410: 1.81 13.2 1.65 8 20 2.05 2.66 2.40 420 1 23 13.7 1.80 9.18 2.06 2.68 2.60 430 2.55 13.3 2.05 9.28 2.37 2.87 2.70 440 3.00 13.6 2.26 9.44 2.65 2.89 2.72 450 3.12 15.2 2.36 11.0 2.68 2.90 2.64 460 3.20 16.8 2.36 12.2 2.64 2.90 2.62 470 3.09 16.9 2.31 12.4 2.68 2.92 2.54 480 2.84 17. l 2.01 12.8 2.47 2 88 2.46 490 2.78 18.8 1.92 13.7 2.42 2.76 2.38 500 2.58 18.5 1.84 l .7 2.39 2.76 2.28 510 2.43 24.0 1.81 22.4 2.41 2.69 2.20 520 2.29 32.6 1.62 29.6 2.26 2.59 2.20 F30 2.20 48.4- 1.63 33.6 2.19 2.60 2.11 540 2.10 51.0 1.64 34.6 2.20 2.60 2.21 550 2.07 520 1.64 35.6 2.21 2.56 2.22 560 2.12 52.5 1.64 36.7 2.38 2.51 2.22 570 2.41 50.3 1.65 37.3 2.48 2.64 2.42 580 3.30 48.2 1.99 37.4 3.09 3.11 3.02 590 5.84 46.5 3.20 38.0 4.73 4. 16 4.38 600 11.8 46.0 6.80 39.0 8.10 7.30 7.69 610 18.7 44.8 12.3 38.0 13.6 11.3 11.3 620 25.5 42.8 17.8 36.7 19.2 13.6 16.2 630 30.8 42.9 25.2 36.2 25.4 21.4 22.4 640 33.9 41.0 29.4 33.8 29.2 25.4 26.2 650 33.0 35.3 30.4 28.8 30. 1 27.0 28.5 660 29.4 30.4 28.4 24.5 28.8 26.4 28 5 670 24.5 24.3 24.0 18.3 25.5 22.4 25.5 680 25.0 25.4 25.4 18.6 27.0 24.0 26.6 690 39.2 41.0 40.2 32.4 40.6 33.2 40.6 700 57.0 58.1 56.0 52.5 55.5 51.5 55.2 710 64.9 66.2 63.5 62.1 63.6 59.4 62.1 720 67.8 69.6 66.0 67.3 66.7 62.9 65.6 :5: Green Surfaces 60 TABLE VI CORRECTED SPECTRAL REFLECTANCE READINGS FOR TYPICAL COMMON DELlC-IOUS AND SPY SURFACE a -u-.—-'. - .n-u—‘noq -——v m-b._-‘\n—.uc—.——'~um~.—. - - —.- . . an '\ —-—-_ ..,__. —_——- ---~--au-U _‘—‘ .- n-—<-‘--—'m. “.‘-Mp _-——u Vsjirie'w Common Deiicitus Spy sznqne New 1*-.-_ P* j 3 '"T_w§_ 1* 2 3 Was-- Per 5 Per I? Per j P6 ‘ P61" Per Per Ce I _. nt C e nt C e n't C ent 16 mg 5:1”. C ent: ! Cent C ent ! I l 430 11.4 8.92 3 10 2.91 4 20 2.57 3.08 :) 3*.8 9.40 3.21 3.36 4 81 2.39 3.47 4 0 '” 9 'fi.‘“ 3.22 3.8?’ 5 3- 2.77 4.32 l; f 5 17 C 3.14 4-46 5.92 5.'6 5 10 44 ; 1r 3-xs 4.?5 6.22 4.16 5.3 450 1+ 5 I? 4 2 90 4.65 6.77 4.35 5.37 460 16 3 i5 J 2.90 4.77 7.30 4.42 5.39 4”” i’ S -3 v 2.82 4.80 '?.35 4.45 5 18 1~ in 1. -3.4 2 46 4.43 '7.10 4.16 4.74 44) 18 2 “5.4 2 56 4.41 7.35 3.99 4.55 E r- 2; 5 21.0 2 57 4.“: 8.65 3.88 4.46 3-") 30 ".7 /9. J 2 ”i8 4, 6 ll,“ 3.' L‘. 4.39 :26 L7 5 -‘ l 2.58 4.’8 -l4.3 3.54 4.20 k 1 £7 , i: a 1.34 4 90 16-3 3.44 4.22 ~40 43 1 44.3 2.60 ’4 97 L7.3 3.3 4.23 -n 44 42.8 2.82 5.2 18.3 3.3 4.45 t -45 45.2 3.28 7.90 19 3 04 533 46. 44.5 4 35 7.45 ’0 66 b «n.s 43. 41. 42. 41. 39. 38. 36. 32. 27. 21. 22. 54. 51. 59. 62. momm O\ ,p p. C; :50 v-h U0 hprpomOC‘I-JOOrPO‘OO‘O C on; '. 2.\2.‘ .. 'fxlmNOb-‘C‘ON‘JTO'LR‘ N OJ 590 46. 600 47. 610 47. 62C 46. 630 46. 640 44. 650 39. 660 34_ 670 28. 680 28. 690 43. 706 58. 716 65. 720 68. a: 'an<3‘.< Data for this table taken from the Spectral distribution curves prepared by Westinghouse Corporation. 11. 12. 13. 14. 15. 75 REFERENCES AISOp, D. (1958). Personal correspondence. . Bartley, S. H. (1941). Vision. D. VanNorstrand Co. Beckman Instruments, Inc. (1954). Beckrnan Instruction Manual 305-A. lackwell, H. R. (19 58). Development and use of a generalized method for Specification of interior illumination levels on the basis of performance data. Unpublished report at Illuminating Engineering Research Institute Research Symposium. 87 pp. Crouch, C. L. (1945). The relation between illumination and vision. Illuminating Engineering. 40: pp. 747-784. Ditch; .an, J. P. (1958). Personal communication. Evans, R. M. (19 48). An Introduction to Color. John Wiley and Sons, Inc. , New York. ’_ - (1949). Light sources and colored objects. Illumi» nating Engineering. 44: pp. 47-54, . General Electric Company. (1946). Lamp Bulletin LD-l. Gould, W. A. (1952). Artificial light for color evaluation of fruit and vegetables. Food Packer. 33: pp. 33-35. Helson, H. (1955). Color and seeing. Illuminating Engineering. 50: pp. 271-78. ., Jeffers, V. B. (1940). Fundamental problems in color vision, II. Jour. Experimental Phychology. 26: pp. 1-27. ., Judd, D. B., and Warren, M. H. (1952). Object- color changes from daylight to incandescent filament illumination. Illuminating Engineering. 47: pp. 221-33. Hunt, R. W. G. (1950). The effect of daylight and tungsten light-adaption on color perception. Journal of Optical Society of America. 40: pp. 362-371. Illuminating Engineering Society Committee Report. (1950). Lighting for canneries. Illuminating Engineering. 45: pp. 45—65. 16. l7. 18. 19. 20. 25. 26. 27. 2.8. 29. 30. 76 Illuminating Engineering Society Lighting Handbook. '. Malcolzm, D. G. and DeGarIno, E. P. 2nd Edition. (1952). Illuminating Engineering Society. New York. pp. 4~30~4~31. Jerome, C. W., and Judd, D. B. (1953). Specification of color rendering properties of fluorescent lamps. Illuminating Engineering. 48: p. 256. Judd, D. B. (1952). Color in Busines_s, Science, and Industry. John Wiley and Sons, New York. Levin, J. H. (1958). Personal communication. Linsday, E. A. (1949). The use of fluorescent lamps for examining color proof-s. Colored lighting for color printing. General Electric Bulletin LS~-122. (1953). Visual racteristics in grading R inspections of products for surface ch r eport 45. -a operations. USDA Marketing Resea ch . Marketing, Yearbook of Agriculture) 1954. (1954). USDA. pp: 61-70, 121-140. . Michigan Agricultural Statistics. (1957). Michigan Department of Agriculture . . Michigan CrOp Reporting Bulletin. (1958). Michigan Crop Reporting Se rvic e . Motts, G. N. (1.958). Apple grader's manua‘i. Michigan State University Extention Folder F-l99 (Revised). Nickerson, D. (1956). Color measurement and its application to the grading of agricultural products. USDA Misc. Publ. 580. (1956). Achievement of lighting standards for grading cotton. USDA Agricultural Marketing Service-94. Norris, K. H. (1958). Measuring transmittance properties of agricultural commodities. Agricultural Engineering 39: p. 640. Optical Society of America. (1953). The Science of Color. Thomas Y. Crowell, New York. Parker, B. F. (1954). Some effects of chromatic illumination, reflectance and product rotation on sorting efficiency of cherries and tomatoes. Thesis for degree of Ph. D. , Michigan State University. 163 pp. 31. 33. 34. 77 Peterson, G. M. (1951). The effects of color environment on grading efficiency in food processing plants. Thesis for degree of M. S. , Michigan State University. 78 pp. . Simonson, E. and Brozek, J. (1948). The effects of Spectral quality of light on visual performance and fatigue. Journal of Optical Society of America. 38: p. 830. and ___. (1952). Work, vision and illumination. Illuminating Pfri-gineering. 47: pp. 335-344. Taylor, A. H. (1942). The nature and causes of small color differences in industry, lighting for their detection. Journal of Optical Society of America. 32: pp. 651-658. . . Weston, H. C. (1954 . Visual fatigue. Illuminating Engineer-- ing. 50: p. 63. meg-i USE em DEC 5 1960. 1 I A STATE UNIVERSITY L BRARIES 30811004 mlllllllllllnll I III 3 0 1293