OpenCV：如何使用掩码参数进行特征点检测（SURF）

Question

我想将 SurfFeatureDetector 限制为一组区域（掩码）。对于测试，我只定义了一个掩码：

Mat srcImage; //RGB source image
Mat mask = Mat::zeros(srcImage.size(), srcImage.type());
Mat roi(mask, cv::Rect(10,10,100,100));
roi = Scalar(255, 255, 255);
SurfFeatureDetector detector();
std::vector<KeyPoint> keypoints;
detector.detect(srcImage, keypoints, roi); // crash
//detector.detect(srcImage, keypoints); // does not crash

当我将“roi”作为掩码传递时，我收到此错误：

OpenCV Error: Assertion failed (mask.empty() || (mask.type() == CV_8UC1 && mask.size() == image.size())) in detect, file /Users/ux/Downloads/OpenCV-iOS/OpenCV-iOS/../opencv-svn/modules/features2d/src/detectors.cpp, line 63

这有什么问题？如何正确地将掩码传递给 SurfFeatureDetector 的“检测”方法？

问候，

Answer 1

关于面具的两件事。

• 掩码应该是一个 8 位无符号字符的 1 通道矩阵，转换为 opencv 类型 CV_8U。在您的情况下，掩码的类型为 srcImage.type()，它是一个 3 通道矩阵
• 您正在将roi 传递给检测器，但您应该传递mask。当您对roi 进行更改时，您也在更改mask。

以下应该可以工作

Mat srcImage; //RGB source image
Mat mask = Mat::zeros(srcImage.size(), CV_8U);  // type of mask is CV_8U
// roi is a sub-image of mask specified by cv::Rect object
Mat roi(mask, cv::Rect(10,10,100,100));
// we set elements in roi region of the mask to 255 
roi = Scalar(255);  
SurfFeatureDetector detector();
std::vector<KeyPoint> keypoints;
detector.detect(srcImage, keypoints, mask);     // passing `mask` as a parameter

Answer 2

我将您的 ROI 代码添加到我正在处理的一些现有代码上，并进行了以下对我有用的更改

cv::Mat mask = cv::Mat::zeros(frame.size(), CV_8UC1);  //NOTE: using the type explicitly
cv::Mat roi(mask, cv::Rect(10,10,100,100));
roi = cv::Scalar(255, 255, 255);

//SURF feature detection
const int minHessian = 400;
cv::SurfFeatureDetector detector(minHessian);
std::vector<cv::KeyPoint> keypoints;
detector.detect(frame, keypoints, mask);              //NOTE: using mask here, NOT roi
cv::Mat img_keypoints; 
drawKeypoints(frame, keypoints, img_keypoints, cv::Scalar::all(-1), cv::DrawMatchesFlags::DEFAULT);
cv::imshow("input image + Keypoints", img_keypoints);
cv::waitKey(0);

如果不更改类型并使用 mask 而不是 roi 作为掩码，我也会收到运行时错误。这是有道理的，因为检测方法需要一个掩码——它应该与原始图像的大小相同，而 roi 不是（它是一个 100x100 的矩形）。要直观地看到这一点，请尝试显示蒙版和 roi

cv::imshow("Mask", mask);
cv::waitKey(0);

cv::imshow("ROI", roi);
cv::waitKey(0);

类型也必须匹配；掩码应该是单通道，而您的图像类型可能是类型 16，它映射到 CV_8UC3，三通道图像

Answer 3

如果您希望对不规则蒙版应用相同的方法，那么：

Mat& obtainIregularROI(Mat& origImag, Point2f topLeft, Point2f topRight, Point2f botLeft, Point2f botRight){

        static Mat black(origImag.rows, origImag.cols, origImag.type(), cv::Scalar::all(0));
        Mat mask(origImag.rows, origImag.cols, CV_8UC1, cv::Scalar(0));
        vector< vector<Point> >  co_ordinates;
        co_ordinates.push_back(vector<Point>());
        co_ordinates[0].push_back(topLeft);
        co_ordinates[0].push_back(botLeft);
        co_ordinates[0].push_back(botRight);
        co_ordinates[0].push_back(topRight);
        drawContours( mask,co_ordinates,0, Scalar(255),CV_FILLED, 8 );

       // origImag.copyTo(black,mask);
        //BasicAlgo::getInstance()->writeImage(black);
        return mask;  // returning the mask only
    }

然后像往常一样，生成SIFT/SURF/...指针

// 为 SIFT 特征检测器创建智能指针。

Ptr<FeatureDetector> SIFT_FeatureDetector = FeatureDetector::create("SIFT");
vector<KeyPoint> SIFT_Keypoints;
vector<KeyPoint> SIFT_KeypointsRotated; 
Mat maskedImg = ImageDeformationOperations::getInstance()->obtainIregularROI( rotatedImg,rotTopLeft,rotTopRight,rotBotLeft,rotBotRight);
SIFT_FeatureDetector->detect(rotatedImg, SIFT_KeypointsRotated, maskedImg);
Mat outputSIFTKeyPt;
drawKeypoints(rotatedImg, SIFT_KeypointsRotated, outputSIFTKeyPt, keypointColor, DrawMatchesFlags::DEFAULT);