Ore classification based on color and texture feature fusion

For most of ore sorting is done by manpower. There are some problems such as low sorting efficiency and unstable quality of products and so on. In order to improve the sorting efficiency and quality of ore, this paper propose a method which based on color and texture feature fusion to sort ore. The features of RGB, HSV, Ycbcr, LAB, NTSC and Gabor are extracted from two samples of good and inferior ore, and a total of 510 samples of good and inferior ore were sorted as samples of the experiment. The six features and their combination are respectively trained by support vector machine (SVM) classifier. Our preliminary analysis over evaluating indicator, Euclidean metric and scatter diagram demonstrate that Combination feature provides the best performance with effect.