Image hashing method based on simple linear iterative clustering and salient region

The invention discloses an image hashing method based on simple linear iterative clustering and a salient region, which relates to the technical field of image processing, and comprises the following steps: acquiring an image for preprocessing, performing superpixel segmentation on the image by utilizing improved linear iterative clustering SLIC, and calculating a quaternion digital-analog mean value as a color feature sequence; extracting a saliency map of the image by using a saliency detection algorithm, performing dimension reduction on the saliency map, and obtaining a salient feature sequence through a vector distance; and obtaining a final image Hash sequence through pseudo-random arrangement, calculating a Hamming distance, and verifying image similarity. The method has good tolerance to noise and deformation in the image and good adaptability to color change, the image corresponds to a unique hash code, and hash conflicts are avoided; according to the method, the images are converted into the Hash se