（pod 'OpenCV','~> 2.4.13'）最新的版本，由于有些接口改变太多，我这边导入的事2.4.13版本。
void find_squares(cv::Mat& image, std::vector<std::vector<cv::Point>>&squares) {
// blur will enhance edge detection
cv::Mat blurred(image);
// medianBlur(image, blurred, 9);
GaussianBlur(image, blurred, cvSize(11,11), 0);//change from median blur to gaussian for more accuracy of square detection
cv::Mat gray0(blurred.size(), CV_8U), gray;
std::vector<std::vector<cv::Point> > contours;
// find squares in every color plane of the image
for (int c = 0; c < 3; c++)
{
int ch[] = {c, 0};
mixChannels(&blurred, 1, &gray0, 1, ch, 1);
// try several threshold levels
const int threshold_level = 2;
for (int l = 0; l < threshold_level; l++)
{
// Use Canny instead of zero threshold level!
// Canny helps to catch squares with gradient shading
if (l == 0)
{
Canny(gray0, gray, 10, 20, 3); //
// Canny(gray0, gray, 0, 50, 5);
// Dilate helps to remove potential holes between edge segments
dilate(gray, gray, cv::Mat(), cv::Point(-1,-1));
}
else
{
gray = gray0 >= (l+1) * 255 / threshold_level;
}
// Find contours and store them in a list
findContours(gray, contours, CV_RETR_LIST, CV_CHAIN_APPROX_SIMPLE);
// Test contours
std::vector<cv::Point> approx;
for (size_t i = 0; i < contours.size(); i++)
{
// approximate contour with accuracy proportional
// to the contour perimeter
approxPolyDP(cv::Mat(contours[i]), approx, arcLength(cv::Mat(contours[i]), true)*0.02, true);
// Note: absolute value of an area is used because
// area may be positive or negative - in accordance with the
// contour orientation
if (approx.size() == 4 &&
fabs(contourArea(cv::Mat(approx))) > 1000 &&
isContourConvex(cv::Mat(approx)))
{
double maxCosine = 0;
for (int j = 2; j < 5; j++)
{
double cosine = fabs(angle(approx[j%4], approx[j-2], approx[j-1]));
maxCosine = MAX(maxCosine, cosine);
}
if (maxCosine < 0.3)
squares.push_back(approx);
}
}
}
}
}
#pragma mark -- CvVideoCameraDelegate
- (void)processImage:(cv::Mat&)image{
//image对象就是视频流时时返回的对象，一秒可达30帧；用一个Mat对象（imageClone）对返回的image对象拷贝
cv::Mat imageClone = image.clone();
std::vector<std::vector<cv::Point> > squares;
//图片灰度处理
cv::Mat src_gray;
//第一个参数为转换前的image，第二个参数为，转换后的image，第三个参数为需要转换成什么样的图片，这边为灰度图片。
cv::cvtColor(image, src_gray, cv::COLOR_BGR2GRAY);
// 滤波降噪，第三个参数为阈值，可以搜搜看
cv::Mat filtered;
cv::blur(src_gray, filtered, cv::Size(7,7));
//边缘检测
cv::Mat edges;
int thresh = 50;
cv::Canny(filtered, edges, thresh, thresh * 2, 3);
//膨胀
cv::Mat dilated_edges;
cv::dilate(edges, dilated_edges, cv::Mat(), cv::Point(-1, -1), 2, 1, 1);
// 找到轮廓并存储
std::vector<std::vector<cv::Point> > contours;
cv::findContours(dilated_edges, contours, cv::RETR_LIST, cv::CHAIN_APPROX_SIMPLE);
std::vector<cv::Point> approx;
for (size_t i = 0; i < contours.size(); i++) {
//找出轮廓的凸包
cv::convexHull(contours[i], approx) ;
// 对cv::arcLength(cv::Mat(approx), true)*0.02这个阈值外的线条进行拟合
cv::approxPolyDP(cv::Mat(approx), approx, cv::arcLength(cv::Mat(approx), true)*0.02, true);
if (approx.size() == 4 && std::fabs(contourArea(cv::Mat(approx))) > 1000 &&
cv::isContourConvex(cv::Mat(approx)))
{
double maxCosine = 0;
for (int j = 2; j < 5; j++)
{
double cosine = std::fabs(anglec(approx[j%4], approx[j-2], approx[j-1]));
maxCosine = MAX(maxCosine, cosine);
}
if (maxCosine < 0.3)
squares.push_back(approx);
}
}
// 至此，你可以发现，用了一个白色的框，框住了身份证的大轮廓。
std::vector<cv::Point> largest_square;
find_largest_squarec(squares, largest_square);
const cv::Point* p = &largest_square[0];
int n = (int)largest_square.size();
cv::polylines(image, &p, &n, 1, true, cv::Scalar(0, 0, 0), 3, CV_AA);
_currentMat = image;
_imgView.image = [MMOpenCVHelper UIImageFromCVMat:_currentMat];
}
void find_squaresc(cv::Mat& image, std::vector<std::vector<cv::Point>>&squares) {
// blur will enhance edge detection
cv::Mat blurred(image);
// medianBlur(image, blurred, 9);
GaussianBlur(image, blurred, cvSize(11,11), 0);//change from median blur to gaussian for more accuracy of square detection
cv::Mat gray0(blurred.size(), CV_8U), gray;
std::vector<std::vector<cv::Point> > contours;
// find squares in every color plane of the image
for (int c = 0; c < 3; c++)
{
int ch[] = {c, 0};
mixChannels(&blurred, 1, &gray0, 1, ch, 1);
// try several threshold levels
const int threshold_level = 2;
for (int l = 0; l < threshold_level; l++)
{
// Use Canny instead of zero threshold level!
// Canny helps to catch squares with gradient shading
if (l == 0)
{
Canny(gray0, gray, 10, 20, 3); //
// Canny(gray0, gray, 0, 50, 5);
// Dilate helps to remove potential holes between edge segments
dilate(gray, gray, cv::Mat(), cv::Point(-1,-1));
}
else
{
gray = gray0 >= (l+1) * 255 / threshold_level;
}
// Find contours and store them in a list
findContours(gray, contours, CV_RETR_LIST, CV_CHAIN_APPROX_SIMPLE);
// Test contours
std::vector<cv::Point> approx;
for (size_t i = 0; i < contours.size(); i++)
{
// approximate contour with accuracy proportional
// to the contour perimeter
approxPolyDP(cv::Mat(contours[i]), approx, arcLength(cv::Mat(contours[i]), true)*0.02, true);
// Note: absolute value of an area is used because
// area may be positive or negative - in accordance with the
// contour orientation
if (approx.size() == 4 &&
fabs(contourArea(cv::Mat(approx))) > 1000 &&
isContourConvex(cv::Mat(approx)))
{
double maxCosine = 0;
for (int j = 2; j < 5; j++)
{
double cosine = fabs(anglec(approx[j%4], approx[j-2], approx[j-1]));
maxCosine = MAX(maxCosine, cosine);
}
if (maxCosine < 0.3)
squares.push_back(approx);
}
}
}
}
}