SIFT更改关键点比例（有意义的关键点邻域）

Question

我是opencv的新手，而且正在开发中。我正在使用SIFT关键点匹配两个图像。为了做到这一点，我使用了以下代码。但是它提供了一些错误匹配。

我想如果我改变了关键点的大小。这将有助于获得更正确的关键点。当前大小的关键点是不够的。请帮助我改进size of the key points。

Ptr<FeatureDetector> featureDetector = FeatureDetector::create("SIFT");
    featureDetector->detect(input_right, keypoints_right);// is pointer so i used ->
    featureDetector->detect(input_left, keypoints_left);



    //-- Step 2: Calculate descriptors (feature vectors)
    Ptr<DescriptorExtractor> featureExtractor = DescriptorExtractor::create("SIFT");


    featureExtractor->compute(input_right, keypoints_right, descriptor_right);
    // which is pointer so i used ->
    featureExtractor->compute(input_left, keypoints_left, descriptor_left);
    // show features

    // check detected keypoints right
    Mat outputImageright;
    Scalar keypointColor = Scalar(255, 0, 0);     // Blue keypoints.
    drawKeypoints(input_right, keypoints_right, outputImageright, keypointColor, DrawMatchesFlags::DEFAULT);

    namedWindow("Right View");
    imshow("Right View", outputImageright);


    Mat outputImageleft;
    Scalar keypointColorred = Scalar(0, 0, 255);     // Red keypoints.
    drawKeypoints(input_left, keypoints_left, outputImageleft, keypointColorred, DrawMatchesFlags::DEFAULT);

    namedWindow("Left View");
    imshow("Left View", outputImageleft);