Homography in computer vision explained

Finding Homography Matrix using Singular-value Decomposition and RANSAC in OpenCV and Matlab.

[yiliangxie]

This helped a lot at 1 day before my CV exam. Thanks!

[پوریاشریفی-هر]

بهنام جان دمت گرم فقط عکس اول با کیفیت محتوا کمی تفاوت داشت. مطالبت بسیار مفید و آموزنده بود

[DavidGoodman]

This is a great video and helped me get unstuck on a problem I was working on. Thank you!

[b.bakhtiari]

Thanks Behnam, very informative like always 🙏

[mariamgarba]

13:16 A little mistake here... -x'2H11 should be -x1H11

[tymothylim]

Thank you very much for this video! It must have taken quite a lot of effort to make this video! Thank you for your effort :))

[dewikharismawati]

Thank you for creating this video, it helps me understand homography better.

[meghdadhasheminnasab]

I believe the projection matrix has to be 3 by 4, not 4 by 4. then considering Z = 0, third column can be eliminated and we are left by a 3 by 3 Homography matrx

[rusencelik]

so how can you recover the tvec and rvec from this homography matrix if we know the intrinsic parameters

[klbm]

at 5:53 it should be a 3x4 matrix, 4x4 by 4x1 does not give 3x1. Camera intrinsic matrix is 3x3, extrinsic is 3x4 giving us a 3x4 matrix, not a 4x4. Then we take z=0, and ignore the 3rd column, giving us a 3x3 matrix. Please rectify or correct me if i am wrong.

[turtleden]

Awesome explanation thank you a lot with such a difficult topic!

[cindycindy]

17:12 Should the top left cell in the 9x9 matrix be -x2 instead of -x1

[engineersingh]

Thanks for very informative video on homography in CV.
Please clarify when I am taking coordinate from image, they will be in x, y like 0, 0 for top-left corner and some other point say 115, 269. Now while calculating homographic matrix, shall we need to convert them into meters like meter/pixel into pixel coordinate or can we use image coordinates directly in the equations for calculating Homographic matrix.

[sahartabrizi]

does this model work for CCD cameras as well? so we can calculate coordinates of a 3D point in the world using its image coordinates and Homography matrix of camera? it would be great if you explain how we can generate homography matrix for a CCD camera. Many thanks.

[zhuoerlyu]

Hi, it is really an excellent tutorial, Thank you so much.
But I still have a question about its application.
I am gonna use this to remove perspective distortion which requires me to get a frontal image from an arbitrary position of the same image.But i only have one picture, therefore, the question is that how can I get the corresponding points to calculate the H matrix.
Looking forward your reply!Thank you
Best