dense-stereo

所属分类:图形图像处理
开发工具:Visual C++
文件大小:545KB
下载次数:34
上传日期:2015-11-10 16:22:57
上 传 者xuyong88
说明:  基于VC++的双目立体视觉检测 采用SIFT特征点计算稀疏匹配 和稠密匹配
(VC++ based binocular stereo vision detection using SIFT feature points to calculate sparse matching and dense matching)

文件列表:
qdmatch (0, 2009-06-21)
qdmatch\headers (0, 2009-06-21)
qdmatch\headers\Correspond.h (1701, 2009-05-31)
qdmatch\headers\GaussConvol.h (588, 2009-05-31)
qdmatch\headers\GlobalPara.h (2019, 2009-05-31)
qdmatch\headers\HeapSort.h (841, 2009-05-31)
qdmatch\headers\Image.h (871, 2009-05-31)
qdmatch\headers\KDTree.h (6391, 2009-05-31)
qdmatch\headers\Keypoint.h (1623, 2009-05-31)
qdmatch\headers\Match.h (2257, 2009-05-31)
qdmatch\headers\NyxMat.h (31810, 2009-06-21)
qdmatch\headers\NyxSvd.h (567, 2009-05-31)
qdmatch\headers\QuasiDense.h (2177, 2009-05-31)
qdmatch\headers\QuasiMatch.h (420, 2009-05-31)
qdmatch\headers\Ransac.h (2821, 2009-05-31)
qdmatch\headers\ScaleSpace.h (4109, 2009-05-31)
qdmatch\headers\Sift.h (202, 2009-05-31)
qdmatch\headers\SiftMatch.h (746, 2009-05-31)
qdmatch\headers\SiftUtil.h (696, 2009-05-31)
qdmatch\headers\SparseMatch.h (229, 2009-05-31)
qdmatch\pic (0, 2009-06-21)
qdmatch\pic\v1.JPG (89018, 2009-05-31)
qdmatch\pic\v2.JPG (88074, 2009-05-31)
qdmatch\pic\v3.JPG (88419, 2009-05-31)
qdmatch\src (0, 2009-06-21)
qdmatch\src\Correspond.cpp (18559, 2009-06-21)
qdmatch\src\GaussConvol.cpp (5141, 2009-06-21)
qdmatch\src\HeapSort.cpp (4613, 2009-06-21)
qdmatch\src\Image.cpp (7435, 2009-06-21)
qdmatch\src\KDTree.cpp (18116, 2009-06-21)
qdmatch\src\Keypoint.cpp (7065, 2009-05-31)
qdmatch\src\main.cpp (5556, 2009-06-21)
qdmatch\src\Match.cpp (7222, 2009-06-21)
qdmatch\src\NyxSvd.cpp (8948, 2009-06-21)
qdmatch\src\QuasiDense.cpp (25313, 2009-06-21)
qdmatch\src\QuasiMatch.cpp (3720, 2009-06-21)
qdmatch\src\Ransac.cpp (30008, 2009-06-21)
qdmatch\src\ScaleSpace.cpp (38305, 2009-06-21)
qdmatch\src\Sift.cpp (6708, 2009-06-21)
... ...

This program is used to find point matches between two images. The procedure can be divided into two parts: 1) Firstly, using SIFT matching algorithm to find sparse point matches between two images. 2) Secondly, using "quasi-dense propagation" algorithm to get "quasi-dense" point matches. "Quasi-dense" means the matches are distributed evenly, and have quantities been improved significantly. This program needs OpenCV to be installed first. 用法: 1. 普通扩散 (与以往同,即求两幅图像的扩散后的公共点) ------------------------------------------------- nyxprog im1.jpg im2.jpg match12.txt ------------------------------------------------- match12.txt格式如下: y1 x1 y2 x2 zncc ... ... ... ... ... 所以矩阵是Nx5的 2. 多幅扩散 假设有1,2,3三幅图像,其中图像1和2已经扩散过,且知道其匹配保存在match12.txt中,要生成1,2和3幅的共同匹配,只需: ---------------------------------------------------------------------------- nyxprog im2.jpg im3.jpg match12.txt match123.txt match23.txt ---------------------------------------------------------------------------- 1) 其中match123.txt是要生成的3幅图像的共同匹配,格式为: y1 x1 y2 x2 y3 x3 y1 x1 y2 x2 y3 x3 ... ... ... ... 相当于是一个Nx6的矩阵 2) match23.txt是要生成的图像2和3的匹配的文件。不过注意,这里的匹配是去除了 1,2,3共有匹配后剩下的不同匹配,所以实际上2和3的全部匹配为 match23.txt中的 加上 match123.txt中的。 match23.txt的格式如以往相同,是Nx5的,即: y1 x1 y2 x2 zncc ... ... ... ...

近期下载者

相关文件


收藏者