opencv 识别长方形_OpenCV 圆与矩形识别的方法

最近一个项目用到了图像识别,之前从未接触过OpenCV,经过各种找教程,终于是搞懂了一些。

整个具体流程大概是获取图像-->图像二值化,灰度图(cvtColor)-->图像降噪(GaussianBlur)->轮廓识别(cvFindContours)-->形状判断。

大多数教程很专业,各种参数分析看不懂,经过各种搜索终于是搞懂了。

识别圆

在识别圆方面,OpenCV有内置的方法:霍夫圆变化:

HoughCircles(edges, circles, CV_HOUGH_GRADIENT, 1.5, 10, 200, 100, 0, 0);

参数分析:

edges:灰度图像

circles: std::vector circles;数组,用来存储圆的坐标信息

CV_HOUGH_GRADIENT:Hough 变换方式,目前只支持CV_HOUGH_GRADIENT, which is basically 21HT, described in [Yuen03].默认用这个

1.5:累加器图像的分辨率,1的时候是与获取到的图像相同,1.5就是1.5倍

10:圆与圆的最小距离,两个圆心距离如果在范围内则被认定为1个圆

200:100-200两个参数选就够了

100:默认100,数值越低识别圆越不精确(圆的数量识别变多可能有个弧线就被识别是圆)

最后两个参数分别是识别 圆的最小,最大的面积。

矩形识别

矩形识别并没有内置方法,需要自己手写。

最主要的方法是二值化。通过二值化来调节识别的强度。

cvThreshold(tgray, gray, 75, 250, CV_THRESH_BINARY);

参数分析:

src:原始数组 (单通道 , 8-bit of 32-bit 浮点数)。

dst:输出数组,必须与 src 的类型一致,或者为 8-bit。

threshold:阈值

max_value:使用 CV_THRESH_BINARY 和 CV_THRESH_BINARY_INV 的最大值。

threshold_type:阈值类型

threshold_type=CV_THRESH_BINARY:如果 src(x,y)>threshold ,dst(x,y) = max_value; 否则,dst(x,y)=0;

threshold_type=CV_THRESH_BINARY_INV:如果 src(x,y)>threshold,dst(x,y) = 0; 否则,dst(x,y) = max_value.

threshold_type=CV_THRESH_TRUNC:如果 src(x,y)>threshold,dst(x,y) = max_value; 否则dst(x,y) = src(x,y).

threshold_type=CV_THRESH_TOZERO:如果src(x,y)>threshold,dst(x,y) = src(x,y) ; 否则 dst(x,y) = 0。

threshold_type=CV_THRESH_TOZERO_INV:如果 src(x,y)>threshold,dst(x,y) = 0 ; 否则dst(x,y) = src(x,y).

效果图如下:

在矩形识别里面的二值化图:

圆识别:

源码:

#include

#include

#include

#include

#include

#include

#include

#include

#include

#pragma comment(lib,"ws2_32.lib")

#include

using namespace cv;

//

//函数功能:用向量来做COSα=两向量之积/两向量模的乘积求两条线段夹角

//输入: 线段3个点坐标pt1,pt2,pt0,最后一个参数为公共点

//输出: 线段夹角,单位为角度

//

double angle(CvPoint* pt1, CvPoint* pt2, CvPoint* pt0)

{

double dx1 = pt1->x - pt0->x;

double dy1 = pt1->y - pt0->y;

double dx2 = pt2->x - pt0->x;

double dy2 = pt2->y - pt0->y;

double angle_line = (dx1*dx2 + dy1 * dy2) / sqrt((dx1*dx1 + dy1 * dy1)*(dx2*dx2 + dy2 * dy2) + 1e-10);//余弦值

return acos(angle_line) * 180 / 3.141592653;

}

//

//函数功能:采用多边形检测,通过约束条件寻找矩形

//输入: img 原图像

// storage 存储

// minarea,maxarea 检测矩形的最小/最大面积

// minangle,maxangle 检测矩形边夹角范围,单位为角度

//输出: 矩形序列

//

CvSeq* findSquares4(IplImage* img, CvMemStorage* storage, int minarea, int maxarea, int minangle, int maxangle, int(&temp)[30])

{

CvSeq* contours;//边缘

int N = 6; //阈值分级

CvSize sz = cvSize(img->width & -2, img->height & -2);

IplImage* timg = cvCloneImage(img);//拷贝一次img

IplImage* gray = cvCreateImage(sz, 8, 1); //img灰度图

IplImage* pyr = cvCreateImage(cvSize(sz.width / 2, sz.height / 2), 8, 3); //金字塔滤波3通道图像中间变量

IplImage* tgray = cvCreateImage(sz, 8, 1); ;

CvSeq* result;

double s, t;

int sk = 0;

CvSeq* squares = cvCreateSeq(0, sizeof(CvSeq), sizeof(CvPoint), storage);

cvSetImageROI(timg, cvRect(0, 0, sz.width, sz.height));

//金字塔滤波

cvPyrDown(timg, pyr, 7);

cvPyrUp(pyr, timg, 7);

//在3个通道中寻找矩形

for (int c = 0; c < 3; c++) //对3个通道分别进行处理

{

cvSetImageCOI(timg, c + 1);

cvCopy(timg, tgray, 0); //依次将BGR通道送入tgray

for (int l = 0; l < N; l++)

{

//不同阈值下二值化

cvThreshold(tgray, gray, 75, 250, CV_THRESH_BINARY);

cvShowImage("111", gray);

cvFindContours(gray, storage, &contours, sizeof(CvContour), CV_RETR_LIST, CV_CHAIN_APPROX_SIMPLE, cvPoint(0, 0));

while (contours)

{ //多边形逼近

result = cvApproxPoly(contours, sizeof(CvContour), storage, CV_POLY_APPROX_DP, cvContourPerimeter(contours)*0.02, 0);

//如果是凸四边形并且面积在范围内

if (result->total == 4 && fabs(cvContourArea(result, CV_WHOLE_SEQ)) > minarea && fabs(cvContourArea(result, CV_WHOLE_SEQ)) < maxarea && cvCheckContourConvexity(result))

{

s = 0;

//判断每一条边

for (int i = 0; i < 5; i++)

{

if (i >= 2)

{ //角度

t = fabs(angle((CvPoint*)cvGetSeqElem(result, i), (CvPoint*)cvGetSeqElem(result, i - 2), (CvPoint*)cvGetSeqElem(result, i - 1)));

s = s > t ? s : t;

}

}

//这里的S为直角判定条件 单位为角度

if (s > minangle && s < maxangle)

{

for (int i = 0; i < 4; i++)

cvSeqPush(squares, (CvPoint*)cvGetSeqElem(result, i));

CvRect rect = cvBoundingRect(contours, 1); // 获取矩形边界框

CvPoint p1;

p1 = cvPoint(rect.x + rect.width / 2, rect.y + rect.height / 2); //矩形中心坐标

std::cout << "X:" << p1.x << "Y:" << p1.y << std::endl;

}

}

contours = contours->h_next;

}

}

std::cout << "圆的数量是"<

temp[26] = sk;

sk = 0;

}

cvReleaseImage(&gray);

cvReleaseImage(&pyr);

cvReleaseImage(&tgray);

cvReleaseImage(&timg);

return squares;

}

//

//函数功能:画出所有矩形

//输入: img 原图像

// squares 矩形序列

// wndname 窗口名称

//输出: 图像中标记矩形

//

void drawSquares(IplImage* img, CvSeq* squares, const char* wndname)

{

CvSeqReader reader;

IplImage* cpy = cvCloneImage(img);

CvPoint pt[4];

int i;

cvStartReadSeq(squares, &reader, 0);

for (i = 0; i < squares->total; i += 4)

{

CvPoint* rect = pt;

int count = 4;

memcpy(pt, reader.ptr, squares->elem_size);

CV_NEXT_SEQ_ELEM(squares->elem_size, reader);

memcpy(pt + 1, reader.ptr, squares->elem_size);

CV_NEXT_SEQ_ELEM(squares->elem_size, reader);

memcpy(pt + 2, reader.ptr, squares->elem_size);

CV_NEXT_SEQ_ELEM(squares->elem_size, reader);

memcpy(pt + 3, reader.ptr, squares->elem_size);

CV_NEXT_SEQ_ELEM(squares->elem_size, reader);

//cvPolyLine( cpy, &rect, &count, 1, 1, CV_RGB(0,255,0), 3, CV_AA, 0 );

cvPolyLine(cpy, &rect, &count, 1, 1, CV_RGB(rand() & 255, rand() & 255, rand() & 255), 1, CV_AA, 0);//彩色绘制

}

cvShowImage("22", cpy);

cvReleaseImage(&cpy);

}

void SendMessageOne()

{

//开起摄像头

VideoCapture capture;

capture.open(0);

Mat edges; //定义转化的灰度图

if (!capture.isOpened())

namedWindow("【效果图】", CV_WINDOW_NORMAL);

const char* winn = "1111";

if (!capture.isOpened())

//namedWindow(winn, CV_WINDOW_NORMAL);

CvMemStorage* storage = 0;

CvMemStorage* storage = 0;

storage = cvCreateMemStorage(0);

while (1)

{

int Y=0, J=0;

Mat frame;

capture >> frame;

IplImage img0 = frame;

//drawSquares(&img0, findSquares4(&img0, storage, 100, 2000, 80, 100, a), winn);

//cvClearMemStorage(storage); //清空存储

Mat E = frame(Range(1, 320), Range(1, 240));

cvtColor(frame, edges, CV_BGR2GRAY);

//高斯滤波

GaussianBlur(edges, edges, Size(7, 7), 2, 2);

std::vector circles;//存储每个圆的位置信息

//霍夫圆

HoughCircles(edges, circles, CV_HOUGH_GRADIENT, 1.5, 10, 100, 100, 0, 50);

for (size_t i = 0; i < circles.size(); i++)

{

Point center(cvRound(circles[i][0]), cvRound(circles[i][1]));

int radius = cvRound(circles[i][2]);

//std::cout << "圆的X是" << circles[i][0] << "圆的Y是" << circles[i][1] << std:: endl;

//绘制圆轮廓

circle(frame, center, radius, Scalar(155, 50, 255), 3, 8, 0);

int R = frame.at(cvRound(circles[i][1]), cvRound(circles[i][0]))[2];//R

int G = frame.at(cvRound(circles[i][1]), cvRound(circles[i][0]))[1];//G

int B = frame.at(cvRound(circles[i][1]), cvRound(circles[i][0]))[0];//B

int num = R + G + B;

std::cout << "圆心颜色是" << num << std::endl;

}

imshow("【效果图】", frame);

waitKey(30);

}

}

int main()

{

std::thread *a = new std::thread(SendMessageOne);

a->join();

return 0;

}

以上就是本文的全部内容,希望对大家的学习有所帮助,也希望大家多多支持我们。

本文标题: OpenCV 圆与矩形识别的方法

本文地址: http://www.cppcns.com/ruanjian/c/264308.html

你可能感兴趣的:(opencv,识别长方形)