SIFT matching method based on K nearest neighbor support feature points

In order to reduce the image matching errors caused by many factors, the method based on k nearest neighbors support feature points was proposed. When ratio is little, get the support feature set by Scale Invariant Feature Transform and choose the k nearest neighbor (KNN) points from the set as the supporters. According to the similarity of supporters around the matching point, determining the matching is correct or not. If the matching is correct, add the point to the support set to expand it dynamically, thereby ensure the support points roundly and accurately. Experimental results show that the method can retain the right matching, at the same time eliminating more than 90% of the mismatches and increasing the correct matching rate greatly.