opencvsharp findhomography怎么使用

Question

            var imagePoint1 = new List<Point2f>();
            var imagePoint2 = new List<Point2f>();
            for (int i = 0; i < matchePoints.Length; i++)
            {
                imagePoint1.Add(keyPoints1[matchePoints[i].QueryIdx].Pt.ToPoint());
                imagePoint2.Add(keyPoints2[matchePoints[i].TrainIdx].Pt.ToPoint()); 
                
            }

            Mat homo = Cv2.FindHomography(imagePoint2, imagePoint1, HomographyMethods.Ransac);

参考的是这篇文章：https://www.jb51.net/article/169211.htm

Answer 1

题主代码明显没看仔细，少了转换

          // 算法RANSAC对匹配的结果做过滤
          var pSrc = pointsSrc.ConvertAll(Point2fToPoint2d);
          var pDst = pointsDst.ConvertAll(Point2fToPoint2d);

根据文章代码应该大概如下，要转换为Point2fToPoint2d的集合

            var imagePoint1 = new List<Point2f>();
            var imagePoint2 = new List<Point2f>();
            for (int i = 0; i < matchePoints.Length; i++)
            {
                imagePoint1.Add(keyPoints1[matchePoints[i].QueryIdx].Pt.ToPoint());
                imagePoint2.Add(keyPoints2[matchePoints[i].TrainIdx].Pt.ToPoint());

            }

///////转换
            var pSrc = imagePoint1 .ConvertAll(Point2fToPoint2d);
            var pDst = imagePoint2 .ConvertAll(Point2fToPoint2d);

            Mat homo = Cv2.FindHomography(pSrc, pDst, HomographyMethods.Ransac);