机械与电子2017，Vol.35Issue(6)：18-21,25,5.
基于SIFT特征提取和最佳缝合线的图像拼接技术
Image Mosaic Technology Based on SIFT Feature Extraction and Optimal Stitching Line
摘要
Abstract
Based on the feature extraction algorithm of SIFT and the idea of the optimal stitching line, a robust image mosaic algorithm was proposed to splice images with different angles of view and different light intensity.Firstly, the SIFT algorithm was used to describe the key points of the image, and the nearest neighbor method was adopted to complete the first matching of the key points in the image, based on which, the RANSAC algorithm was taken for the second matching and the projection transformation model parameters were calculated.Then, the Max-Flow / Min-Cut was calculated to obtain the optimal stitching line, along which the image was stitched.Finally, in order to eliminate the visual cleavage of the suture line, the gradual fade method was used to make the gray images on both sides a uniform transition so as to complete the image mosaic.Experiments point out that this method, inheriting the robustness of the SIFT algorithm, can effectively remove the ghosting problem caused by traditional image fusion so that the result image is natural and realistic.关键词
图像拼接/SIFT算法/RANSAC/最佳缝合线/图像融合Key words
image mosaic/SIFT algorithm/RANSAC/seam-line/image fusion分类
信息技术与安全科学引用本文复制引用
刘雨航,金一,竺长安..基于SIFT特征提取和最佳缝合线的图像拼接技术[J].机械与电子,2017,35(6):18-21,25,5.基金项目
国家重点基础研究发展计划(973)(2014CB049503) （973）
