Research of Image Matching Based on Improved SURF Algorithm

Abstract

In the study, it presents an efficient algorithm based on SURF (Speeded Up Robust Features). The method applies the SURF algorithm in the detection and description for image features, first it applies the SURF feature detector in extracting reference images and matching feature points in the image, respectively, then it uses DAISY algorithm instead of the SURF algorithm to calculate the character description of each feature point vector. In the process of feature points matching, the false matching points are eliminated through RANSAC algorithm. Finally, according to the rest of the match point, it can estimate the space geometric transformation parameters between two images and thus matching process is completed.

DOI : http://dx.doi.org/10.11591/telkomnika.v12i2.3951