A statistical features-based color difference classification method

Abstract

This paper presents a method for color difference based on Histogram statistical values. First, the color image in RGB color space is converted to HSI color space, median filter was selected for image preprocessing.Then, the ceramic samples HSI Histogram statistical of each channel was calculated respectively. Take the Histogram statistical as the color difference classification character value. For real time requirement, using minimum distance classifier as classification basis. Compared with the S, I channel, the results showed that adopted the H channel Histogram statistical value as feature vector has higher accuracy for ceramic tile color difference classification.