前言
本文将使用OpenCV C++ 进行二维码检测。
一、二维码检测
首先我们要先将图像进行预处理，通过灰度、滤波、二值化等操作提取出图像轮廓。在这里我还添加了形态学操作，消除噪点，有效将矩形区域连接起来。
Mat gray;
cvtColor(src, gray, COLOR_BGR2GRAY);
Mat blur;
GaussianBlur(gray, blur, Size(3, 3), 0);
Mat bin;
threshold(blur, bin, 0, 255, THRESH_BINARY_INV | THRESH_OTSU);
//通过Size（5，1）开运算消除边缘毛刺
Mat kernel = getStructuringElement(MORPH_RECT, Size(5, 1));
Mat open;
morphologyEx(bin, open, MORPH_OPEN, kernel);
//通过Size（21，1）闭运算能够有效地将矩形区域连接 便于提取矩形区域
Mat kernel1 = getStructuringElement(MORPH_RECT, Size(21, 1));
Mat close;
morphologyEx(open, close, MORPH_CLOSE, kernel1);
如图为经过一系列图像处理之后得到的效果。之后我们需要对该图进行轮廓提取，找到二维码所在的矩形区域。
//使用RETR_EXTERNAL找到最外轮廓
vector<vector<Point>>MaxContours;
findContours(close, MaxContours, RETR_EXTERNAL, CHAIN_APPROX_SIMPLE);
for (int i = 0; i < MaxContours.size(); i++)
{
Mat mask = Mat::zeros(src.size(), CV_8UC3);
mask = Scalar::all(255);
double area = contourArea(MaxContours[i]);
//通过面积阈值找到二维码所在矩形区域
if (area > 6000 && area < 100000)
{
//计算最小外接矩形
RotatedRect MaxRect = minAreaRect(MaxContours[i]);
//计算最小外接矩形宽高比
double ratio = MaxRect.size.width / MaxRect.size.height;
if (ratio > 0.8 && ratio < 1.2)
{
Rect MaxBox = MaxRect.boundingRect();
//rectangle(src, Rect(MaxBox.tl(), MaxBox.br()), Scalar(255, 0, 255), 2);
//将矩形区域从原图抠出来
Mat ROI = src(Rect(MaxBox.tl(), MaxBox.br()));
ROI.copyTo(mask(MaxBox));
ROI_Rect.push_back(mask);
}
}
}
由以下代码段我们就可以很好的找出二维码所在的矩形区域，并将这些区域抠出来保存以便进行下面的识别工作。
//找到二维码所在的矩形区域
void Find_QR_Rect(Mat src, vector<Mat>&ROI_Rect)
{
Mat gray;
cvtColor(src, gray, COLOR_BGR2GRAY);
Mat blur;
GaussianBlur(gray, blur, Size(3, 3), 0);
Mat bin;
threshold(blur, bin, 0, 255, THRESH_BINARY_INV | THRESH_OTSU);
//通过Size（5，1）开运算消除边缘毛刺
Mat kernel = getStructuringElement(MORPH_RECT, Size(5, 1));
Mat open;
morphologyEx(bin, open, MORPH_OPEN, kernel);
//通过Size（21，1）闭运算能够有效地将矩形区域连接 便于提取矩形区域
Mat kernel1 = getStructuringElement(MORPH_RECT, Size(21, 1));
Mat close;
morphologyEx(open, close, MORPH_CLOSE, kernel1);
//使用RETR_EXTERNAL找到最外轮廓
vector<vector<Point>>MaxContours;
findContours(close, MaxContours, RETR_EXTERNAL, CHAIN_APPROX_SIMPLE);
for (int i = 0; i < MaxContours.size(); i++)
{
Mat mask = Mat::zeros(src.size(), CV_8UC3);
mask = Scalar::all(255);
double area = contourArea(MaxContours[i]);
//通过面积阈值找到二维码所在矩形区域
if (area > 6000 && area < 100000)
{
//计算最小外接矩形
RotatedRect MaxRect = minAreaRect(MaxContours[i]);
//计算最小外接矩形宽高比
double ratio = MaxRect.size.width / MaxRect.size.height;
if (ratio > 0.8 && ratio < 1.2)
{
Rect MaxBox = MaxRect.boundingRect();
//rectangle(src, Rect(MaxBox.tl(), MaxBox.br()), Scalar(255, 0, 255), 2);
//将矩形区域从原图抠出来
Mat ROI = src(Rect(MaxBox.tl(), MaxBox.br()));
ROI.copyTo(mask(MaxBox));
ROI_Rect.push_back(mask);
}
}
}
}
如图所示，这是找到的二维码矩形。这里只展示其中之一。
二、二维码识别
1.通过findContours找到轮廓层级关系
//用于存储检测到的二维码
vector<vector<Point>>QR_Rect;
//遍历所有找到的矩形区域
for (int i = 0; i < ROI_Rect.size(); i++)
{
Mat gray;
cvtColor(ROI_Rect[i], gray, COLOR_BGR2GRAY);
Mat bin;
threshold(gray, bin, 0, 255, THRESH_BINARY_INV|THRESH_OTSU);
//通过hierarchy、RETR_TREE找到轮廓之间的层级关系
vector<vector<Point>>contours;
vector<Vec4i>hierarchy;
findContours(bin, contours, hierarchy, RETR_TREE, CHAIN_APPROX_NONE);
//父轮廓索引
int ParentIndex = -1;
int cn = 0;
//用于存储二维码矩形的三个“回”
vector<Point>rect_points;
for (int i = 0; i < contours.size(); i++)
{
//hierarchy[i][2] != -1 表示该轮廓有子轮廓 cn用于计数“回”中第几个轮廓
if (hierarchy[i][2] != -1 && cn == 0)
{
ParentIndex = i;
cn++;
}
else if (hierarchy[i][2] != -1 && cn == 1)
{
cn++;
}
else if (hierarchy[i][2] == -1)
{
//初始化
ParentIndex = -1;
cn = 0;
}
//如果该轮廓存在子轮廓，且有2级子轮廓则认定找到‘回’
if (hierarchy[i][2] != -1 && cn == 2)
{
drawContours(canvas, contours, ParentIndex, Scalar::all(255), -1);
RotatedRect rect;
rect = minAreaRect(contours[ParentIndex]);
rect_points.push_back(rect.center);
}
}
}
以上代码段的整体思路为：首先经过图像预处理进行轮廓检测，
通过hierarchy、RETR_TREE找到轮廓之间的层级关系。根据hierarchy[i][2]是否为-1判断该轮廓是否有子轮廓。若该轮廓存在子轮廓，则统计有几个子轮廓。如果该轮廓存在子轮廓，且有2级子轮廓则认定找到‘回’ 。关于轮廓的层级关系，大家可以自行百度查找资料，理解一下其中原理。
//对找到的矩形区域进行识别是否为二维码
int Dectect_QR_Rect(Mat src,Mat &canvas,vector<Mat>&ROI_Rect)
{
//用于存储检测到的二维码
vector<vector<Point>>QR_Rect;
//遍历所有找到的矩形区域
for (int i = 0; i < ROI_Rect.size(); i++)
{
Mat gray;
cvtColor(ROI_Rect[i], gray, COLOR_BGR2GRAY);
Mat bin;
threshold(gray, bin, 0, 255, THRESH_BINARY_INV|THRESH_OTSU);
//通过hierarchy、RETR_TREE找到轮廓之间的层级关系
vector<vector<Point>>contours;
vector<Vec4i>hierarchy;
findContours(bin, contours, hierarchy, RETR_TREE, CHAIN_APPROX_NONE);
//父轮廓索引
int ParentIndex = -1;
int cn = 0;
//用于存储二维码矩形的三个“回”
vector<Point>rect_points;
for (int i = 0; i < contours.size(); i++)
{
//hierarchy[i][2] != -1 表示该轮廓有子轮廓 cn用于计数“回”中第几个轮廓
if (hierarchy[i][2] != -1 && cn == 0)
{
ParentIndex = i;
cn++;
}
else if (hierarchy[i][2] != -1 && cn == 1)
{
cn++;
}
else if (hierarchy[i][2] == -1)
{
//初始化
ParentIndex = -1;
cn = 0;
}
//如果该轮廓存在子轮廓，且有2级子轮廓则认定找到‘回’
if (hierarchy[i][2] != -1 && cn == 2)
{
drawContours(canvas, contours, ParentIndex, Scalar::all(255), -1);
RotatedRect rect;
rect = minAreaRect(contours[ParentIndex]);
rect_points.push_back(rect.center);
}
}
//将找到地‘回’连接起来
for (int i = 0; i < rect_points.size(); i++)
{
line(canvas, rect_points[i], rect_points[(i + 1) % rect_points.size()], Scalar::all(255), 5);
}
QR_Rect.push_back(rect_points);
}
return QR_Rect.size();
}
由以上代码段，我们就可以识别出二维码。效果如图所示。
三、二维码绘制
//框出二维码所在位置
Mat gray;
cvtColor(canvas, gray, COLOR_BGR2GRAY);
vector<vector<Point>>contours;
findContours(gray, contours, RETR_EXTERNAL, CHAIN_APPROX_SIMPLE);
Point2f points[4];
for (int i = 0; i < contours.size(); i++)
{
RotatedRect rect = minAreaRect(contours[i]);
rect.points(points);
for (int j = 0; j < 4; j++)
{
line(src, points[j], points[(j + 1) % 4], Scalar(0, 255, 0), 2);
}
}
最终效果如图所示。
四、源码
#include<iostream>
#include<opencv2/core.hpp>
#include<opencv2/imgproc.hpp>
#include<opencv2/highgui.hpp>
using namespace std;
using namespace cv;
//找到二维码所在的矩形区域
void Find_QR_Rect(Mat src, vector<Mat>&ROI_Rect)
{
Mat gray;
cvtColor(src, gray, COLOR_BGR2GRAY);
Mat blur;
GaussianBlur(gray, blur, Size(3, 3), 0);
Mat bin;
threshold(blur, bin, 0, 255, THRESH_BINARY_INV | THRESH_OTSU);
//通过Size（5，1）开运算消除边缘毛刺
Mat kernel = getStructuringElement(MORPH_RECT, Size(5, 1));
Mat open;
morphologyEx(bin, open, MORPH_OPEN, kernel);
//通过Size（21，1）闭运算能够有效地将矩形区域连接 便于提取矩形区域
Mat kernel1 = getStructuringElement(MORPH_RECT, Size(21, 1));
Mat close;
morphologyEx(open, close, MORPH_CLOSE, kernel1);
//使用RETR_EXTERNAL找到最外轮廓
vector<vector<Point>>MaxContours;
findContours(close, MaxContours, RETR_EXTERNAL, CHAIN_APPROX_SIMPLE);
for (int i = 0; i < MaxContours.size(); i++)
{
Mat mask = Mat::zeros(src.size(), CV_8UC3);
mask = Scalar::all(255);
double area = contourArea(MaxContours[i]);
//通过面积阈值找到二维码所在矩形区域
if (area > 6000 && area < 100000)
{
//计算最小外接矩形
RotatedRect MaxRect = minAreaRect(MaxContours[i]);
//计算最小外接矩形宽高比
double ratio = MaxRect.size.width / MaxRect.size.height;
if (ratio > 0.8 && ratio < 1.2)
{
Rect MaxBox = MaxRect.boundingRect();
//rectangle(src, Rect(MaxBox.tl(), MaxBox.br()), Scalar(255, 0, 255), 2);
//将矩形区域从原图抠出来
Mat ROI = src(Rect(MaxBox.tl(), MaxBox.br()));
ROI.copyTo(mask(MaxBox));
ROI_Rect.push_back(mask);
}
}
}
}
//对找到的矩形区域进行识别是否为二维码
int Dectect_QR_Rect(Mat src,Mat &canvas,vector<Mat>&ROI_Rect)
{
//用于存储检测到的二维码
vector<vector<Point>>QR_Rect;
//遍历所有找到的矩形区域
for (int i = 0; i < ROI_Rect.size(); i++)
{
Mat gray;
cvtColor(ROI_Rect[i], gray, COLOR_BGR2GRAY);
Mat bin;
threshold(gray, bin, 0, 255, THRESH_BINARY_INV|THRESH_OTSU);
//通过hierarchy、RETR_TREE找到轮廓之间的层级关系
vector<vector<Point>>contours;
vector<Vec4i>hierarchy;
findContours(bin, contours, hierarchy, RETR_TREE, CHAIN_APPROX_NONE);
//父轮廓索引
int ParentIndex = -1;
int cn = 0;
//用于存储二维码矩形的三个“回”
vector<Point>rect_points;
for (int i = 0; i < contours.size(); i++)
{
//hierarchy[i][2] != -1 表示该轮廓有子轮廓 cn用于计数“回”中第几个轮廓
if (hierarchy[i][2] != -1 && cn == 0)
{
ParentIndex = i;
cn++;
}
else if (hierarchy[i][2] != -1 && cn == 1)
{
cn++;
}
else if (hierarchy[i][2] == -1)
{
//初始化
ParentIndex = -1;
cn = 0;
}
//如果该轮廓存在子轮廓，且有2级子轮廓则认定找到‘回’
if (hierarchy[i][2] != -1 && cn == 2)
{
drawContours(canvas, contours, ParentIndex, Scalar::all(255), -1);
RotatedRect rect;
rect = minAreaRect(contours[ParentIndex]);
rect_points.push_back(rect.center);
}
}
//将找到地‘回’连接起来
for (int i = 0; i < rect_points.size(); i++)
{
line(canvas, rect_points[i], rect_points[(i + 1) % rect_points.size()], Scalar::all(255), 5);
}
QR_Rect.push_back(rect_points);
}
return QR_Rect.size();
}
int main()
{
Mat src = imread("6.png");
if (src.empty())
{
cout << "No image data!" << endl;
system("pause");
return 0;
}
vector<Mat>ROI_Rect;
Find_QR_Rect(src, ROI_Rect);
Mat canvas = Mat::zeros(src.size(), src.type());
int flag = Dectect_QR_Rect(src, canvas, ROI_Rect);
//imshow("canvas", canvas);
if (flag <= 0)
{
cout << "Can not detect QR code!" << endl;
system("pause");
return 0;
}
cout << "检测到" << flag << "个二维码。" << endl;
//框出二维码所在位置
Mat gray;
cvtColor(canvas, gray, COLOR_BGR2GRAY);
vector<vector<Point>>contours;
findContours(gray, contours, RETR_EXTERNAL, CHAIN_APPROX_SIMPLE);
Point2f points[4];
for (int i = 0; i < contours.size(); i++)
{
RotatedRect rect = minAreaRect(contours[i]);
rect.points(points);
for (int j = 0; j < 4; j++)
{
line(src, points[j], points[(j + 1) % 4], Scalar(0, 255, 0), 2);
}
}
imshow("source", src);
waitKey(0);
destroyAllWindows();
system("pause");
return 0;
}
总结
本文使用OpenCV C++进行二维码检测，关键步骤有以下几点。
1、图像预处理，筛选出二维码所在的矩形区域，并将该区域抠出来进行后续的识别工作。
2、对筛选出的矩形区域进行轮廓检测，判断它们之前的层级关系，以此来识别二维码。
3、最后根据检测到的二维码“回”字，将其绘制出来就可以了。
以上就是我对二维码检测识别的整体思路，欢迎大家一起交流学习。若存在不足的地方欢迎大家指正。