Detection of Watermelon Seeds Exterior Quality based on Machine Vision

To investigate the detection of watermelon seeds exterior quality, a machine vision system based on least square support vector machine was developed. Appearance characteristics of watermelon seeds included area, perimeter, roughness, minimum enclosing rectangle and solidity were calculated by image analysis after image preprocess. The broken seeds, normal seeds and high-quality seeds were distinguished by least square support vector machine optimized by genetic algorithm. Compared to the grid search algorithm, the classification results of watermelon seeds exterior quality achieved by genetic algorithm were analyzed in detail. Meanwhile machine vision grid laser was applied to detect the surface irregularities defects of watermelon seeds. This study demonstrated the feasible of detecting the watermelon seeds exterior quality by machine vision.