Extract Characteristic of High Reflect Rate Surface Texture Using Wavelet Transformation

Abstract

Against the problem of extracting the color aberration of high-reflect-rate ceramic tile surface texture, author came up a way of using wavelet transformation to extract the image texture features. Basically, the way is suggesting using two dimension wavelet decomposition on each passage of the image and extract energy characteristic from each detailed sub-graph. This energy signal merges the message of color and feature. Calculate samples' color histogram statistics as the critical data for doing classification. In the end, using Minimum Distance Classifier to sort the samples. According to the experiment, compare the existing method which mostly using the color histogram distribution, this method can show better message of color distribution in space. Experiment shows that this method can get a better classify result.