Improved segmentation algorithm of pixel texture used in complex image

When Image feature extracted from the sub-block in image is used in object segmentation, it can cause object boundary inaccurate. Texture feature extracting algorithm based on every pixel is purposed in the paper. After the pixel texture feature is standardized and reduced dimension by PCA, it is clustered by hierarchical method. Improved segmentation algorithm segments the complex image into different objects. Compared with the traditional methods such as threshold method, Sobel method and so on, the segmentation boundary is accurate, and the phenomenon of under-segmentation and over-segmentation also reduce significantly.