您可以在使用 HSV 范围的图像中分割棕色对象。由于棕色在某种程度上是一种较深的红色,因此您需要稍微调整一下参数。如果您发布参考图片,我们可以找到更准确的范围。
一旦你有了对象蒙版(你通常应用一些形态来清理蒙版),你可以很容易地用findContours得到轮廓。
下面的例子解释了这一点:
#include <iostream>
#include <vector>
#include "opencv2/opencv.hpp"
using namespace std;
using namespace cv;
int main()
{
Mat3b img = imread("path_to_image");
Mat3b hsv;
cvtColor(img, hsv, COLOR_BGR2HSV);
Mat1b mask1, mask2;
inRange(hsv, Scalar(0, 100, 20), Scalar(10, 255, 255), mask1);
inRange(hsv, Scalar(170, 100, 20), Scalar(180, 255, 255), mask2);
Mat1b mask = mask1 | mask2;
Mat1b kernel = getStructuringElement(MORPH_ELLIPSE, Size(7,7));
morphologyEx(mask, mask, MORPH_OPEN, kernel);
vector<vector<Point>> contours;
findContours(mask.clone(), contours, CV_RETR_LIST, CV_CHAIN_APPROX_SIMPLE);
Mat3b res = img.clone();
for(int i=0; i<contours.size(); ++i)
{
drawContours(res, contours, i, Scalar(0,255,0));
RotatedRect r = minAreaRect(contours[i]);
Point2f pts[4];
r.points(pts);
for (int j = 0; j < 4; ++j)
{
line(res, pts[j], pts[(j + 1) % 4], Scalar(0,0,255));
}
Rect box = boundingRect(contours[i]);
rectangle(res, box, Scalar(255,0,0));
}
imshow("Original", img);
imshow("Segmented", res);
waitKey();
return 0;
}
初始图片
分段的棕色物体(美式足球)
更新为实际图像
由于您发布的图片比我之前示例中的图片更难(因为您的瞳孔外有很多几乎是棕色的颜色),您还需要:
- 正确的范围值
- 找到最大的 blob
这段代码显示了这个:
#include <iostream>
#include <vector>
#include "opencv2/opencv.hpp"
using namespace std;
using namespace cv;
int main()
{
Mat3b img = imread("D:\\SO\\img\\eye.jpg");
Mat3b hsv;
cvtColor(img, hsv, COLOR_BGR2HSV);
Mat1b mask;
inRange(hsv, Scalar(2, 100, 65), Scalar(12, 170, 100), mask);
Mat1b kernel = getStructuringElement(MORPH_ELLIPSE, Size(3, 3));
morphologyEx(mask, mask, MORPH_OPEN, kernel);
vector<vector<Point>> contours;
findContours(mask.clone(), contours, CV_RETR_LIST, CV_CHAIN_APPROX_NONE);
if (contours.empty()) {return -1;}
int idx_largest_blob = 0;
int size_largest_blob = contours[0].size();
if (contours.size() > 1)
{
for (int i = 0; i < contours.size(); ++i)
{
if (size_largest_blob < contours[i].size())
{
size_largest_blob = contours[i].size();
idx_largest_blob = i;
}
}
}
Mat3b res = img.clone();
drawContours(res, contours, idx_largest_blob, Scalar(0, 255, 0));
RotatedRect r = minAreaRect(contours[idx_largest_blob]);
Point2f pts[4];
r.points(pts);
for (int j = 0; j < 4; ++j)
{
line(res, pts[j], pts[(j + 1) % 4], Scalar(0, 0, 255));
}
Rect box = boundingRect(contours[idx_largest_blob]);
rectangle(res, box, Scalar(255, 0, 0));
imshow("Original", img);
imshow("Segmented", res);
waitKey();
return 0;
}
结果:
注意:如果您需要更准确的信息,您应该发布一个专门询问瞳孔检测的新问题。
我会删除一些有用的链接,以防万一:
http://answers.opencv.org/question/12034/face-eyes-and-iris-detection/
https://github.com/trishume/eyeLike
https://github.com/laoyang
http://cmp.felk.cvut.cz/~uricamic/flandmark/
http://opencv-code.com/tutorials/pupil-detection-from-an-eye-image/
http://thume.ca/projects/2012/11/04/simple-accurate-eye-center-tracking-in-opencv/
http://opencv-code.com/tutorials/eye-detection-and-tracking/