A computer vision system was used to evaluate external physical damage, mold contamination, and flouryto-vitreous endosperm ratio in corn and mold contamination in soybeans. For each of these quality factors, optimal conditions for acquiring video images and processing algorithms were developed. White light in front-lighting mode with a black background for the sample was suitable for all defects except for mold contamination which required use of red light (610 nm). The image processing algorithms were suitable for defect detection in samples both individually and in groups. The average success rates for detecting broken, chipped, starch-cracked and moldy corn kernels were 100%, 83%, 88% and 84%, respectively. The success rate for detecting moldy soybeans was 80%. Physical Damage Initial damage to corn kernel pericarp is caused by combine harvesting (Roberts, 1972). Subsequent handling during various conditioning and processing steps significantly contributes to the physical damage
A N image processing algorithm was developed for detecting stress cracks in corn kernels using a commercial vision system. White light in back-lighting mode with black-coated background having a small aperture for the light provided the best viewing conditions. The kernel images, when processed using the algorithm developed, produced white streaks corresponding to the stress cracks. Double stress cracks were the easiest to detect. Careful positioning of the kernel over the lighting aperture was necessary for satisfactory detection of single and multiple stress cracks.
Color sorting was performed to upgrade seed quality by removal of fluorescent coated seeds. The fluorescent coating was attributed to sinapine leakage from nonviable seeds. Nine seedlots, three seedlots each of cabbage (Brassica oleracea L. Capitata group), broccoli, and cauliflower (B. oleracea L. Botrytis group) were custom coated. Seed samples were pretreated before coating with or without 1.0% NaOCl for 10 minutes to enhance leakage. All samples revealed a percentage of seeds with fluorescence. The light emission from selected fluorescent and nonfluorescent coated seeds was quantified by fiber-optic spectrophotometry. Fluorescence was expressed from 400 to 560 nm, with peak emission being from 430 to 450 nm. These data confirmed our visual interpretation of blue-green fluorescence. The ratio of light emission from fluorescent compared to nonfluorescent coated seeds ranged from 4.5 to 7.0 for all samples and averaged 5.7. An ultraviolet (UV) color sorter was employed to separate fluorescent (reject) from nonfluorescent (accept) coated seeds. The percentage of nonfluorescent coated seeds (averaged over seedlot and NaOCl pretreatment) before and after sorting was 89.5% and 95.9%, respectively. Therefore, color sorting was able to remove a high percentage of fluorescent coated seeds with an average loss (rejection of nonfluorescent coated seeds) of 6%. An increase in the percent germination was recorded in eight of the nine seedlots following color sorting, and the greatest improvement was obtained with seedlots of medium quality. Germination of three medium quality lots was increased by 10 to 15 percentage points. The average increase in germination with or without NaOCl pretreatment was 8.2 and 5.5 percentage points, respectively. In conclusion, the germination of Brassica seedlots could be improved by separating (removing) fluorescent from nonfluorescent coated seeds. UV color sorting technology was employed to demonstrate that seed conditioning could be conducted on a commercial basis to upgrade seed quality.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.