Image Stitching Method Based on Genetic Algorithm
Yang Xu, Sun Changbo · 2017
In allusion to the inaccuracy of homography matrix using RANSAC from image stitching based on image features, a new image stitching algorithm based on genetic algorithm was proposed.First, locate and describe the feature points of images using feature extraction algorithm, and and the coarse matching was performed based on the ratio between nearest distance and next nearest distance.Then get optimal transform model using genetic algorithm.At last, use the Laplacian transform pyramid reconstruction algorithm to reconstruct and obtain the final fused image.The experiment show that compared with common methods, the method proposed in this paper performs better in terms of accuracy of homography matrix.Subjectively, it improves the quality and effect of fusion image.