Introduction - If you have any usage issues, please Google them yourself
Extract the corresponding feature point of the image, and the feature point streamline with the description, to find 128-dimensional descriptor, and compares with a description of the image sub-image matching