OpenCV_描述和匹配兴趣点

为了进行基于兴趣点的图像分析，我们需要构建能够为一地描述关键点地展现方式，即从兴趣点提取描述子。这些描述子通常是 二值类型、整数型或浮点型组成地一维或二维向量，描述了一个关键点和它的邻域。好的描述子要具有足够地独特性，能唯一地表示图像中地每个关键点。它还要有足够地鲁棒性，在照度变化或视角变动时仍能较好地体现同一批点集。

图像匹配是关键点地常用功能之一。它的作用包括关联同一场景地两个图像、检测图像中事物发生地点等。

*局部模板匹配

最常见地图像块是边长为奇数地正方形，关键点位置就是正方形中心。可以通过比较块内像素地强度值，来衡量两个正方形图像块地相似度。

常见的方案是采用简单的差的平方和（Sum of Squared Differences,SSD）算法。

效果：

代码：

int main()
{
	cv::Mat mouse1 = cv::imread("mouse1.jpg");
	cv::resize(mouse1, mouse1, cv::Size(480, 320));
	cv::Mat mouse2 = cv::imread("mouse2.jpg");
	cv::resize(mouse2, mouse2, cv::Size(480, 320));
	std::vector<cv::KeyPoint> keypoints1;
	std::vector<cv::KeyPoint> keypoints2;
	cv::Ptr<cv::FastFeatureDetector> fastD = cv::FastFeatureDetector::create(80);
	fastD->detect(mouse1, keypoints1);
	fastD->detect(mouse2, keypoints2);

	//定义正方形领域
	const int nsize(11);
	cv::Rect neighborhood(0, 0, nsize, nsize);
	cv::Mat patch1, patch2;
	
	//针对第一幅图像中的每个关键点，在第二幅图像中找出最佳匹配
	cv::Mat result;
	std::vector<cv::DMatch> matches;
	for (int i = 0; i < keypoints1.size(); i++)
	{
		neighborhood.x = keypoints1[i].pt.x - nsize / 2;
		neighborhood.y = keypoints1[i].pt.y - nsize / 2;
		//如果邻域超出图像范围，就继续处理下一个点
		if (neighborhood.x < 0 || neighborhood.y < 0 || neighborhood.x + nsize >= mouse1.cols || neighborhood.y + nsize >= mouse1.rows)
		{
			continue;
		}
		patch1 = mouse1(neighborhood);
		//复位最佳匹配的值
		cv::DMatch bestMatch;
		//针对图像二的全部关键点
		for (int j = 0; j < keypoints2.size(); j++)
		{
			neighborhood.x = keypoints2[j].pt.x - nsize / 2;
			neighborhood.y = keypoints2[j].pt.y - nsize / 2;
			if (neighborhood.x < 0 || neighborhood.y < 0 || neighborhood.x + nsize >= mouse2.cols || neighborhood.y + nsize >= mouse2.rows)
				continue;
			patch2 = mouse2(neighborhood);
			//匹配两个图像块
			cv::matchTemplate(patch1, patch2, result, CV_TM_SQDIFF_NORMED);
			//检查是否最佳匹配
			if (result.at<float>(0.0) < bestMatch.distance)
			{
				bestMatch.distance = result.at<float>(0, 0);
				bestMatch.queryId