Hi, I would like to ask a help at the forum to obtain a Homography Matrix. I am studying the subject at the paper and computer (MatLab), I am not implementing the things in C Language with OpenCV ... The goal is to learn the algorithm theory and after to use the OpenCV.
Thus, who works with Computer Vision, Camera Calibration and MatLab, could, please, help me ?
The situation:
I got two views and 4 points at each view of a Chessboard with the help of 3DStudioMax. The points are below:
p_src  p_dst (pixels) 
(meters)  Image_01  Image_02 

(0.1;0.1)  (296;199)  (319;208) 
(0.2;0.1)  (367;210)  (383;193) 
(0.1;0.2)  (273;267)  (333;286) 
(0.2;0.2)  (346;281)  (395;266) 
Now, I have to mount the equation to find the Homography Matrix (H).
DOUBT: How do I mount those equations ? Is some thing like below?
p_dst = H * p_src
[296 ; 199 ; 1] = H1 * [0.1 ; 0.1 ; 1]
[367 ; 210 ; 1] = H1 * [0.2 ; 0.1 ; 1]
[273 ; 267 ; 1] = H1 * [0.1 ; 0.2 ; 1]
[346 ; 281 ; 1] = H1 * [0.2 ; 0.2 ; 1]
[319 ; 208 ; 1] = H2 * [0.1 ; 0.1 ; 1]
[383 ; 193 ; 1] = H2 * [0.2 ; 0.1 ; 1]
[333 ; 286 ; 1] = H2 * [0.1 ; 0.2 ; 1]
[395 ; 266 ; 1] = H2 * [0.2 ; 0.2 ; 1]
where ,
H1 = [ h11_1 h12_1 h13_1 ; h21_1 h22_1 h23_1 ; h31_1 h32_1 h33_1 ]
H2 = [ h11_2 h12_2 h13_2 ; h21_2 h22_2 h23_2 ; h31_2 h32_2 h33_2 ]
My objective is to find H1 and H2 ... Is it possible?
The OpenCV official book "says" that to find Homography Matrix, 2 images and four points at each one would be enough!
Very Thanks by the attention and help,
Carlos
carlosjunior@gmail.com
