Aluminum Alloy X-ray Image Classification Using Texture Analysis

This paper presents an automatic classification approach to the X-ray image classification issue of aluminum alloy by image texture analysis methods. Different from the common processing methods, the texture-based approach (XTexture) treats the X-ray image as a special texture image for further processing. By extracting self-correlation moment and wavelet-coefficient moments as the basic classification features based on image texture analysis, XTexture selects nearest neighbor method based on weighted Euclidean distance to classify the images. The experiments demonstrate that XTexture represents an initially better performance