2020-10-05 01:27:32 -0600 | received badge | ● Student (source) |
2013-10-18 21:05:54 -0600 | asked a question | grabCut bad_alloc problem The input image size is: 4080*3072. The problem is: First-chance exception at 0x76424b32 in qt2.exe: Microsoft C++ exception: std::bad_alloc at memory location 0x003db614. The grabCut method can not process an image with this size? |
2013-10-09 03:39:17 -0600 | commented answer | bounding ellipse needed I know you point. hope opencv will have a bounding ellipse api in the future. |
2013-10-01 08:43:35 -0600 | commented answer | bounding ellipse needed Thank you very much~~I have a problem that need to describe the degree of ellipse of a contour. and I want to extract the bounding ellipse, then compare the contour's area and its bounding ellipse area. minAreaRect is not suitable. And I think I can judge by each pixel inside the contour, But I can not get all the points inside the contour. |
2013-09-30 23:16:18 -0600 | asked a question | bounding ellipse needed Is there a method to get the bounding ellipse of a contour. I mean all the points should be inside the ellipse. fitEllipse get the best fitting ellipse,not the bounding ellipse. |
2013-09-27 22:22:16 -0600 | asked a question | approximate ellipse detection I have an image with some approximate ellipses inside, and I want to segment them out. fitEllipse function can only detect accurate ellilps and does not have any error tolerance. So is there any API or suggestive methods? Thanks in advance~~ |
2013-09-24 07:08:01 -0600 | commented answer | findContours and Mat(contours[i]) problems I changed the code as yours, but the problems remains. area=24752, W2=68, H2=1. |
2013-09-24 04:14:24 -0600 | asked a question | findContours and Mat(contours[i]) problems The value of H2 is always 1, what's the problem? |
2013-09-24 04:09:15 -0600 | asked a question | findContours and Mat(contours[i]) problem std::vector<std::vector<cv:oint> > contours;
cv::findContours(gray_img->clone(), contours, CV_RETR_LIST, CV_CHAIN_APPROX_SIMPLE);
int contour_num=contours.size();
for (int i = 0; i < contour_num; i++){
double area = cv::contourArea(contours[i]); //value of "area" is correct, such as 6000 |