the scale factor is applied to the (pixel) values, not to the size of the Mat.
an example:
Mat m1(3,3,CV_32F, 2.0f);
cerr << m1 << endl;
[2, 2, 2;
2, 2, 2;
2, 2, 2]
Mat m2;
m1.convertTo(m2, CV_8U, 0.5);
cerr << m2 << endl;
[1, 1, 1;
1, 1, 1;
1, 1, 1]
How does it handle overflows then like what if it was Mat m1 with type CV_32F and values [2, 2, 2; 2, 2, 2; 2, 2, 2] then we do m1.convertTo(m1, CV_8UC1, 255); I believe this would lead to [510, 510, 510; 510, 510, 510; 510, 510, 510] If im not mistaken. But unsigned ints only have a max value of 255? so is it set to 255 if it exceeds the maximum possible value?
Asked: 2016-09-17 23:39:42 -0600
Seen: 2,704 times
Last updated: Sep 18 '16
