Super-pixel spatial spectrum multi-kernel hyperspectral image classification method based on LBP features

The invention provides a super-pixel spatial spectrum multi-kernel hyperspectral image classification method based on LBP features. The method comprises the following steps: firstly, performing superpixel segmentation on an image subjected to dimensionality reduction by a principal component analysis method to generate a hyperspectral image with a superpixel index; then, using weighted average filtering and an LBP algorithm to extract spatial features between superpixels and in the superpixels to obtain spatial kernels between the superpixels and spatial kernels in the superpixels, and combining the extracted spectrum kernels for fusion; and finally, inputting the combined kernel into a support vector machine classifier to generate a classification result graph. According to the invention, the LBP algorithm is combined with the superpixels, and the edge feature information in the superpixels is extracted by using the LBP algorithm, so that the problems of pixel edge information loss and inaccurate edge pixel