Introduction - If you have any usage issues, please Google them yourself
There is provided an image recognition method based on color and texture features, comprising the steps of: dividing the detected image A and the matching image B into n subblocks calculating the similarity of the color characteristics of the detected image A and the matching image B in the subblock k (A, B), which is similar to the color feature similarity distance C (A, B) of the matching image A and the matching image B, and the texture features are similar (A, B) the similarity distance S (A, B) and the texture feature similarity distance T (A, B) of the detected image A and the matching image B are judged to have the similarity S of the detected image A and the matching image B (A, B) .Compared to the prior art, the present invention calculates the feature similarity of the image based on the matching of color and texture features, improves the accuracy and objectivity