更新时间:2023-01-15 18:18:40
您应该将第一个图像扭曲到第二个图像上.您可以使用关键点对应关系给出的findHomography
和perspectiveTransform
函数.您可以在此处找到所需的大多数代码... >
更新
顺便说一句,今天我不得不做基本相同的事情.它已在灰色图像(Mat1b
)上进行了测试,但仅需进行较小的更改即可应用于rgb图像(Mat3b
).
这里是代码的相关部分:
Mat1b A = imread("...");
Mat1b B = imread("...");
vector<Point2f> ptsA;
vector<Point2f> ptsB;
// Fill ptsA, ptsB with the points given by the match of your descriptors.
Mat H = findHomography(ptsA, ptsB, CV_RANSAC); // With ransac is more robust to outliers
Mat1b warpedA;
warpPerspective(A, warpedA, H, B.size());
// Now compute diff
Mat1b res;
absdiff(warpedA, B, res);
// res is what you are looking for!
I'm writing a program that find differences between images. For now, I'm finding features with AKAZE, so I've the common point of the 2 images. The problem is that these 2 images have only a part in common. How can I extract the common part from both images? For better explanation: I need to extract the common part from the first image and then from the second, so I can do absdiff
for finding difference.
I'm programming in c++
Thanks to all!
You should warp the first image onto the second. You can use findHomography
and perspectiveTransform
functions given by the correspondence of your keypoints. You can find most of the code you need here.
Update
Incidentally, I had to do basically the same stuff today. It's tested on gray images (Mat1b
), but should require only minor changes to apply to rgb images (Mat3b
).
Here the relevant parts of the code:
Mat1b A = imread("...");
Mat1b B = imread("...");
vector<Point2f> ptsA;
vector<Point2f> ptsB;
// Fill ptsA, ptsB with the points given by the match of your descriptors.
Mat H = findHomography(ptsA, ptsB, CV_RANSAC); // With ransac is more robust to outliers
Mat1b warpedA;
warpPerspective(A, warpedA, H, B.size());
// Now compute diff
Mat1b res;
absdiff(warpedA, B, res);
// res is what you are looking for!