Improved SURF Feature Matching Algorithm Based on FAST Scale Area

In the face of complex scenes in the real environment, the extraction of feature points with scale-invariant features is a hard requirement in feature matching. The calculation of feature point scales will also directly affect the final feature matching effect. In this paper, the feature matching rate of the SURF algorithm (Speeded Up Robust Features) is not high, and the feature point extraction method is complex and most of the extracted feature points are not human visually significant points. It is proposed to use scale areas instead of FAST algorithm (Features from accelerated segment test) to use pixel points for feature point detection. Different from downsampling to achieve scale-invariant form, this paper achieves scale-invariant feature point extraction by implementing scaling improvement of scale area. The scale-invariant feature points extracted by this method are combined with SURF feature point local feature descriptors to form feature vectors. Using the above method and the original SURF algorithm for feature matching comparison experiment. The results show that the algorithm in this paper has better performance in correct matching rate than the original SURF algorithm.