OpenCV函数cvHoughCircles根据Hough变换检测圆

#include<cv.h>
#include<highgui.h>

int main()
{
	IplImage* src = NULL;
	IplImage* dst = NULL;
	IplImage* color = NULL;

	src = cvLoadImage ("circle.jpg", 1);
	dst = cvCreateImage (cvGetSize(src), IPL_DEPTH_8U, 1);
	if (src->nChannels == 1)
	{
		dst = cvCloneImage (src);
	}
	else
	{
		cvCvtColor (src, dst, CV_RGB2GRAY);
	}

	CvMemStorage* storage = cvCreateMemStorage (0);
	cvSmooth (dst, dst, CV_GAUSSIAN, 5, 5);
	
	CvSeq* circles = cvHoughCircles (dst, storage, CV_HOUGH_GRADIENT, 2, dst->width / 3, 300, 100, 0, 200);
	color = cvCreateImage (cvGetSize(src), IPL_DEPTH_8U, 3);
	cvCvtColor (dst, color, CV_GRAY2RGB);
	for (int i = 0; i < circles->total; i++)
	{
		float* p = (float*)cvGetSeqElem (circles, i);
		CvPoint pt = cvPoint (cvRound(p[0]), cvRound(p[1]));
		cvCircle (color, pt, cvRound(p[2]), CV_RGB(255, 0, 0), 3, 8, 0);
	}

	cvNamedWindow ("src", 1);
	cvShowImage ("src", src);
	cvNamedWindow ("circle", 1);
	cvShowImage ("circle", color);

	cvWaitKey (0);

	cvReleaseMemStorage (&storage);
	cvReleaseImage (&src);
	cvReleaseImage (&dst);
	cvReleaseImage (&color);

	return 0;
}

这是我的源代码，这个我测试了很久，调cvHoughCircles的参数也调了很久，但始终没有调到理想的结果，下面是我

调试中相对来说效果好点的结果：

再分享一下我老师大神的人工智能教程吧。零基础！通俗易懂！风趣幽默！还带黄段子！希望你也加入到我们人工智能的队伍中来！https://www.cnblogs.com/captainbed