Feature Point Matching Algorithm for Fusion of Color Invariants and SURB Detection

Aiming at the low operating efficiency and low matching accuracy problems in image matching, we propose an optimized matching method based on the color invariants and SURE [speeded up robust features-oriented FAST and rotated BRIEF (SURF-ORB)] detection on the basis of ORB matching algorithm. First of all, we obtain the color invariance of two images in the same complex scene through the color space transformation. Then, we use SURE algorithm to extract the feature points with the scale invariance and build an ORE arithmetic descriptor in chromatic information. At last, on the basis of classification and disposal of K nearest neighbor algorithm, we use the cost function and random sample consensus algorithm to remove the mismatching pairs. Results of standard graphics library ALOI and many other real images indicate that the proposed algorithm improves the image matching accuracy and shortens the image matching time in the complex environment.