如有问题请联系:clr_mv@163.com
更多文章请关注微信公众号:机器视觉专业论坛
机器视觉中Blob是非常重要的特征,对于Blob的选择可以依据其几何特征,OpenCv中具有专门的
类来解决这个问题:
class CV_EXPORTS_W SimpleBlobDetector : public Feature2D
{
public:
struct CV_EXPORTS_W_SIMPLE Params
{
CV_WRAP Params();
CV_PROP_RW float thresholdStep;//阈值步长
CV_PROP_RW float minThreshold;//最小阈值
CV_PROP_RW float maxThreshold;//最大阈值
CV_PROP_RW size_t minRepeatability;
CV_PROP_RW float minDistBetweenBlobs;//两个Blob的最小距离
CV_PROP_RW bool filterByColor//采用颜色分割;
CV_PROP_RW uchar blobColor;//主要针对色彩的强度,0选择暗的Blob,255选择亮的Blob
CV_PROP_RW bool filterByArea;//采用最大最小面积分割
CV_PROP_RW float minArea, maxArea;
CV_PROP_RW bool filterByCircularity;//采用圆度分割0~1
CV_PROP_RW float minCircularity, maxCircularity;
CV_PROP_RW bool filterByInertia;//采用惯性率分割0~1
CV_PROP_RW float minInertiaRatio, maxInertiaRatio;
CV_PROP_RW bool filterByConvexity;//采用凹凸度分割0~1
CV_PROP_RW float minConvexity, maxConvexity;
void read( const FileNode& fn );
void write( FileStorage& fs ) const;
};
CV_WRAP static Ptr<SimpleBlobDetector>
create(const SimpleBlobDetector::Params ¶meters = SimpleBlobDetector::Params());
};
该类的使用方法是首先设置参数Params,然后提取满足Param的Blob的重心。SimpleBlobDetector类虽然可以识别出满足条件的Blob,并识别出Blob的重心,但是无法将其作为ROI区域分割开。
为了分割出Blob区域作为ROI区域,本例的主要步骤为:
1.SimpleBlobDetector 选出满足条件的Blob的中心。
2.findContours 查找所有Blob的轮廓。
3.floodFill 漫水填充满足条件的Blob区域。
4.在原图中分割出满足条件的Blob区域作为ROI区域
本例需要注意的有以下几点:
1.SimpleBlobDetector::Param具有默认值,注意初始化。
2.由vector转换为vector 的方法:points.push_back(keypoints[i].pt);
#include <iostream>
#include <opencv2/opencv.hpp>
#include <opencv2/highgui/highgui.hpp>
#include <opencv2/core/core.hpp>
#include <opencv2/imgproc/imgproc.hpp>
#include "opencv2/features2d.hpp"
using namespace cv;
using namespace std;
int main(int argc, char** argv)
{
cv::Mat imagecolor, Blob, Mask;
imagecolor = imread("../data/Blob.png");
//为SimpleBlobDetector设置参数
SimpleBlobDetector::Params params;
params.filterByCircularity = true;
params.filterByConvexity = false;
params.filterByInertia = true;
params.filterByArea = true;
params.filterByColor = false;
params.minThreshold = 0;
params.maxThreshold = 255;
params.blobColor = 255;
params.minArea = 200;
params.maxArea = 10000;
params.minInertiaRatio = 0;
params.maxInertiaRatio = 0.5;
params.minCircularity = 0.5;
params.maxCircularity = 1;
//提取满足条件的Blob的关键点(中心)
Ptr<SimpleBlobDetector> detetor = SimpleBlobDetector::create(params);
vector<KeyPoint> keypoints;
detetor->detect(imagecolor, keypoints);
drawKeypoints(imagecolor, keypoints, Mask, Scalar(255, 0, 255), DrawMatchesFlags::DRAW_RICH_KEYPOINTS);
//查找所有Blob轮廓
Mat imagegray;
cvtColor(imagecolor, imagegray, CV_RGB2GRAY);
vector<vector<Point> > contours;
vector<Vec4i> hierarchy;
findContours(imagegray, contours, hierarchy, CV_RETR_TREE, CV_CHAIN_APPROX_SIMPLE, Point(0, 0));
Mat drawing = Mat::zeros(imagecolor.size(), CV_8UC3);
for (int i = 0; i < contours.size(); i++)
{
drawContours(drawing, contours, i, Scalar(255,255,255), 1, 8, hierarchy, 0, Point());
}
//以满足条件的Blob的重心为起始点进行填充(该方案只适合B重心在Blob上的区域)
vector<Point2f> points;
Mat flood;
for (int i = 0; i < keypoints.size(); i++)
{
points.push_back(keypoints[i].pt);
floodFill(drawing, points[i], Scalar(255,255,255));//注意输入图像为三通道,因此对三个通道都要填充
}
cv::Mat element5(2, 2, CV_8U, cv::Scalar(255, 255, 255));
cv::Mat open;
cv::morphologyEx(drawing, open, cv::MORPH_OPEN, element5);//开操作去除未被选中的blob的contour
imagecolor.copyTo(Blob, open);//分割出满足条件的Blob区域
imshow("Blob", Blob);
imshow("原图", imagecolor);
waitKey();
return 0;
}