The Research on Detection Method of Cotton Contamination Based on Improved RBF Neutral Network

For the low automation level and accuracy in detection of cotton contaminations, this paper adopts GMDH clustering algorithm to define the number of hidden nodes and the center of primary function in RBF neutral network adaptively. The improved RBF neutral network is applied to detect the cotton contaminations. The result shows that the RBF neutral network based on GMDH clustering algorithm could locate contaminations in cotton precisely, and has high location accuracy. That has significance in improving the quality of fabrics and reducing production costs.