Transactions of the ASAE · 1990 · 80 citations · 0 references
AgricultureComputer ScienceFood ScienceComputer VisionFruit ScienceMachine VisionAgricultural Machine VisionColor AnalysisColor GradingColor QualityFruit Quality EvaluationGradingAutomated GradingFruit GradingFruit Image AnalysisHueImage-based GradingLinear Discriminant AnalysisMachine Vision-based SortingQuality Grading
ABSTRACT A machine vision algorithm for grading of fresh market produce according to color and damage was developed. Red-green-blue pixel intensity values were mapped to one of eight possible hues. Treating the relative hue distribution of pixels in six orthogonal views as quantitative variables, discriminant analysis was used to classify observations. When applied to the task of grading bell peppers, accuracies of up to 96% and 63% were found for grading by color and damage, respectively..