Research On Heterochromatic Materials Recognition Method Based On Color Feature

The detection system for heterochromatic materials, an integration of a set of optic, electricity, gas, machine, could detect and remove heterochromatic impurities mixed in normal materials in order to improve the quality of the products. The system improves the labor efficiency, the degree of automation and the accuracy of grading for products to ensure a good quality of the finished products. Based on the above advantages, the detection system for heterochromatic materials has been widely used in food producing, detection and classification of agricultural products.With the basis of theory of machine vision and pattern recognition, the system processes the image captured by the CCD camera, the resulted command controls the motion of the pneumatic nozzle to remove heterochromatic materials.This paper gives a brief introduction to the detection system for heterochromatic materials, and focuses on features and methods of the system. As the most basic, intuitive and obvious physical features, the color feature is selected due to its strong robustness, insensitivity to noise and the size of the image, it is obtained by the histogram, first-order moment, two-order moment and transformation of the wavelet packet. Three color space are mentioned in the paper, namely, RGB color space, Ohta color space and HIS color space.A threshold segmentation method is chosen for recognizing heterochromatic materials, which treats the feature range of the normal materials as the threshold range since a great variety of heterochromatic materials could not be counted completely.