Transactions of the ASAE · 1998 · 30 citations · 0 references
AgricultureAgricultural EngineeringCrop SciencesCereal ScienceComputer VisionMachine VisionAgricultural Machine VisionCrop Stress DetectionKernel MethodsPlant Stress DetectionAccuracyCrop Damage DetectionImage-based Damage DetectionImage-based Damage QuantificationMachine Vision SystemsMechanical DamageMechanical Damage AssessmentMold Damage Assessment
Machine vision algorithms were developed for measuring corn (Zea mays) kernel mechanical damage andmold damage. Mechanical damage was determined using both single-kernel and batch analysis by extracting from kernel images the damaged area stained by green dye and by calculating the percentage of total projected kernel surface area that was stained green. Mold damage was determined using single-kernel analysis by isolating the moldy area on kernel images and by calculating the percentage of total projected kernel surface area covered by mold. The vision system demonstrated high accuracy and consistency for both mechanical and mold damage measurement. The standard deviation for machine vision system measurements was less than 5% of the mean value, which is substantially smaller than for other damage measurement methods.