Feature extraction method based on improved linear LBP operator

An improved wood texture feature description algorithm of Local Binary Pattern operator is proposed. Firstly, the linear LBP operator is used to extract the texture features of the image. Then, the similarity between textures is calculated for the sub-region of the wood image. Because of the extremely high texture similarity of the striped wood, this method can be used to quickly and accurately distinguish the two types of images of stripes and patterns. Experimental results show that the algorithm has strong ability to describe wood texture features, is more robust than traditional methods, and has higher recognition accuracy.