使用opencv拼接图像

拼接两幅图像

Step #1: Detect keypoints (DoG, Harris, etc.) and extract local invariant descriptors (SIFT, SURF, etc.) from the two input images.

Step #2: Match the descriptors between the two images.

Step #3: Use the RANSAC algorithm to estimate a homography matrix using our matched feature vectors.

Step #4: Apply a warping transformation using the homography matrix obtained from Step #3.

参考资料

https://www.pyimagesearch.com/2016/01/11/opencv-panorama-stitching/

http://blog.csdn.net/lhanchao/article/details/52974129

http://blog.csdn.net/lhanchao/article/details/52849446