最近做项目需要摄像机标定和图像转换，OpenCV可以较好的实现这个功能。我的这个例子可以生成两个摄像头的3x3转换矩阵。

但是因为摄像头本身存在成像畸变，尤其是全景摄像机，可能会有更加严重的成像畸变，所有如果试图通过计算两幅完整图像而得到转换单一矩阵，

这个矩阵并不能准确的反应出两幅图像像素之间的对应关系，尤其是靠近边缘区域的像素尤其如此。一个好的建议是将两幅图像分为若干个大小相等

也可以不等的块，分别计算每个块的转换矩阵，这样可以最大程度的降低摄像机成像畸变带来的转换误差。下面是源代码，但是这个代码没有实现分块

<span style="white-space:pre">	</span>//【1】载入原始图片
	Mat srcImage1 = imread("tt1.jpg", 1);
	Mat srcImage2 = imread("tt2.jpg", 1);
	if (!srcImage1.data || !srcImage2.data)
	{
		printf("读取图片错误，请确定目录下是否有imread函数指定的图片存在~！ \n"); return false;
	}

	//【2】使用SURF算子检测关键点
	int minHessian = 600;//SURF算法中的hessian阈值
	SurfFeatureDetector detector(minHessian);//定义一个SurfFeatureDetector（SURF） 特征检测类对象  
	vector<KeyPoint> keypoints_object, keypoints_scene;//vector模板类，存放任意类型的动态数组

	//【3】调用detect函数检测出SURF特征关键点，保存在vector容器中
	detector.detect(srcImage1, keypoints_object);
	detector.detect(srcImage2, keypoints_scene);

	//【4】计算描述符（特征向量）
	SurfDescriptorExtractor extractor;
	Mat descriptors_object, descriptors_scene;
	extractor.compute(srcImage1, keypoints_object, descriptors_object);
	extractor.compute(srcImage2, keypoints_scene, descriptors_scene);

	//【5】使用FLANN匹配算子进行匹配
	FlannBasedMatcher matcher;
	vector< vector< DMatch > > matches;
	//matcher.match(descriptors_object, descriptors_scene, matches);
	matcher.knnMatch(descriptors_object, descriptors_scene, matches,2);
	double max_dist = 0; double min_dist = 100;//最小距离和最大距离
	vector<DMatch> goodMatches;
	for (unsigned int i = 0; i < matches.size(); i++)
	{
		if (matches[i][0].distance < 0.6*matches[i][1].distance)
		{
			goodMatches.push_back(matches[i][0]);
		}
	}
	//【6】计算出关键点之间距离的最大值和最小值
	for (unsigned j = 0; j < goodMatches.size(); j++)
	{
		double dist = goodMatches[j].distance;
		if (dist < min_dist) min_dist = dist;
		if (dist > max_dist) max_dist = dist;
	}

	printf(">Max dist 最大距离 : %f \n", max_dist);
	printf(">Min dist 最小距离 : %f \n", min_dist);

	//【7】存下匹配距离小于3*min_dist的点对
	vector<DMatch>::iterator it;
	for (it = goodMatches.begin(); it != goodMatches.end();)
	{
		if ((*it).distance > 3 * min_dist)
		{
			it=goodMatches.erase(it);
		}
		else
		{
			it++;
		}
	}
	//绘制出匹配到的关键点
	Mat img_matches;
	drawMatches(srcImage1, keypoints_object, srcImage2, keypoints_scene,
		goodMatches, img_matches, Scalar::all(-1), Scalar::all(-1),
		vector<char>(), DrawMatchesFlags::NOT_DRAW_SINGLE_POINTS);

	//定义两个局部变量
	vector<Point2f> obj;
	vector<Point2f> scene;

	//从匹配成功的匹配对中获取关键点
	for (unsigned int i = 0; i < goodMatches.size(); i++)
	{
		obj.push_back(keypoints_object[goodMatches[i].queryIdx].pt);
		scene.push_back(keypoints_scene[goodMatches[i].trainIdx].pt);
	}

	Mat H = findHomography(obj, scene, CV_RANSAC);//计算透视变换 

	//从待测图片中获取四边角点
	vector<Point2f> obj_corners(4);
	obj_corners[0] = cvPoint(0, 0); 
	obj_corners[1] = cvPoint(srcImage1.cols, 0);
	obj_corners[2] = cvPoint(srcImage1.cols, srcImage1.rows); 
	obj_corners[3] = cvPoint(0, srcImage1.rows);
	vector<Point2f> scene_corners(4);

	//进行透视变换
	perspectiveTransform(obj_corners, scene_corners, H);

	//绘制出角点之间的直线
	line(img_matches, scene_corners[0] + Point2f(static_cast<float>(srcImage1.cols), 0), scene_corners[1] + Point2f(static_cast<float>(srcImage1.cols), 0), Scalar(255, 0, 123), 1);
	line(img_matches, scene_corners[1] + Point2f(static_cast<float>(srcImage1.cols), 0), scene_corners[2] + Point2f(static_cast<float>(srcImage1.cols), 0), Scalar(255, 0, 123), 1);
	line(img_matches, scene_corners[2] + Point2f(static_cast<float>(srcImage1.cols), 0), scene_corners[3] + Point2f(static_cast<float>(srcImage1.cols), 0), Scalar(255, 0, 123), 1);
	line(img_matches, scene_corners[3] + Point2f(static_cast<float>(srcImage1.cols), 0), scene_corners[0] + Point2f(static_cast<float>(srcImage1.cols), 0), Scalar(255, 0, 123), 1);

	//显示最终结果
	imshow("效果图", img_matches);

可以看出无论是特征点的选取还是匹配都是相当准确的，紫色的矩形框就是左边图像四个边角点通过转换到右边的效果，也是相当准确。所需时间是在1秒之内