基于多单应性矩阵的图像拼接

# AANAP An implementation of AANAP in CVPR 2015 paper. by YaqiLYU https://www.zhihu.com/question/34535199/answer/135169187 _________________________________________________________________________ Lin C C, Pankanti S U, Ramamurthy K N, et al. Adaptive as-natural-as-possible image stitching [C] 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, 2015: 1155-1163. _________________________________________________________________________ HOW TO START 1. Down APAP code from http://cs.adelaide.edu.au/~tjchin/apap/ PLEASE DOWNLOAD Source codes [MDLT code] 2. Upzip mdlt.zip, run main.m, Make sure APAP is working correctly 3. Download Dateset1, Dateset2, Dateset3 from http://cs.adelaide.edu.au/~tjchin/apap/, unzip three datasetS to ./mdlt/images/ 3. Copy all *.m to ./mdlt/, convert apap to aanap 4. run AANAP.m, set fast_stitch=1 for speed, set fast_stitch=0 for quality! Just for fun! For commercial purposes, please contact the author. BTW, Please forgive my poor English! _________________________________________________________________________ 这是CVPR 2015的图像拼接论文AANAP的MATLAB实现，基于APAP代码，论文： Lin C C, Pankanti S U, Ramamurthy K N, et al. Adaptive as-natural-as-possible image stitching [C]// 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, 2015: 1155-1163. 仅供学习和参考，商业用途请联系原论文作者！ by YaqiLYU 我在这里：https://www.zhihu.com/question/34535199/answer/135169187 _________________________________________________________________________ HOW TO START: 1. 下载APAP: 在https://cs.adelaide.edu.au/~tjchin/apap/ 下载APAP的MATLAB代码，解压后得到./mdlt.    我用的是Source codes [MDLT code]，也就是说，仅支持两图拼接，多图拼接请自行修改，非常简单！ 2. 请先运行main.m，确保APAP可以正常工作。 3. 下载测试图像库：继续在https://cs.adelaide.edu.au/~tjchin/apap/ 下载三个图像数据库，解压到./mdlt/images目录。 4. 拷贝AANAP的所有.m代码到APAP根目录 ./mdlt，注意，这里的所有matlab文件都拷贝过去。 5. 运行AANAP.m就能得到拼接结果了，我提供了两个选项:    设置fast_stitch=1，快速拼接模式，效果较差。    设置fast_stitch=0，高质量拼接模式，速度较慢。 玩的开心！ 版权归源论文作者所有，请不要用于商业用途。