The Application Study of Apple Color Grading by Particle Swarm Optimization Neural Networks

Color is an important index of fruit's external quality, and is studying object of fruit sorting. The apple surface images were acquired by computer vision technology, image's RGB format was converted to HIS format, and hue of every pixel was calculated to get the hue histogram of whole apple. Hue histogram was simplified as appearance times average at seven sub-ranges, and this seven appearance times average was regarded as characteristic parameters of apple color grading. The neural networks were trained by particle swarm optimization (PSO) algorithm, and the apple color was graded with the trained networks. For 16 apples, the grading correctness rate was 94%. The method achieves very high grading speed, gets high precision and has practical value.