Research on an Image Mosaic Algorithm Based on Improved ORB Feature Combined with SURF
Yuping Feng, Shuguang Li · 2018
Aiming at the high requirements of speed and accuracy of image mosaic algorithm, an image mosaic algorithm is proposed by improving the Oriented FAST and Rotated BRIEF(ORB) features combined with the Speeded Up Robust Features(SURF). The algorithm uses the multi-scale space of the SURF algorithm to extract the feature points, which makes up the problem that the ORB algorithm does not have the invariance of scale transformation, and then uses Random Sample Consensus(RANSAC) to remove the mismatched points after rough matching by ORB, and finally uses the Laplacian pyramid to complete seamless integration. Experimental results show that the improved algorithm combines the advantages of the two algorithms and improves the deficiencies of the traditional ORB algorithm. In the case of scale transformation, the improved algorithm has little difference with the original speed, the number of correct matching points is improved by 3 times, and the precision of image mosaic is improved. The proposed algorithm can accurately and quickly realize the seamless mosaic of images with scale.