Opencv STL向量转Mat进行矩阵乘法
矩阵乘法是图像处理中非常基本的任务，OpenCV 使用重载的 * 运算符来处理。点的 STL 向量可以通过转换转换为 Mat。
vector<Point2f> p1(2);
p1[0].x=1;p1[0].y=0;
p1[1].x=0;p1[1].y=1;
Mat p1M=Mat(p1);
正如 OpenCV 文档中提到的，这将创建具有单列（有 2 个元素）的矩阵，其行数等于向量数量：
[1 0;0 1]------>p1M.rows=2;p1M.cols=1
当您想要矩阵乘法 (p1M*p1M)...[2x1]*[ 时，这会产生问题2x1]???...本质上我相信所有到 Matrix 的转换向量所做的就是按原样合并向量......
但是，命令 p1M.at 和 p1M.at 分别返回 0 和 1。这让我认为 p1M*p1M 是可能的，但不幸的是它只编译并生成运行时错误：
OpenCV 错误：gemm 中断言失败 (a_size.width == len)，文件 /home/james/OpenCV-2.3.1/modules/core/src/matmul.cpp，第 708 行 抛出“cv::Exception”实例后调用终止 What(): /home/james/OpenCV-2.3.1/modules/core/src/matmul.cpp:708: 错误: (-215) a_size.width == len 函数 gemm
我正在考虑编写一个函数就这么做吧！向量到 Mat，反之亦然......我错过了什么吗？
Matrix multiplication is a very basic task in image processing and OpenCV takes care of with a overloaded * operator. A STL vector of points can be converted into Mat by casting.
vector<Point2f> p1(2);
p1[0].x=1;p1[0].y=0;
p1[1].x=0;p1[1].y=1;
Mat p1M=Mat(p1);
As mentioned in OpenCV documentation, this will create matrix with a single column(with 2 elements) with rows equal to no of vectors:
[1 0;0 1]------>p1M.rows=2;p1M.cols=1
This creates a problem when you want to matrix multiply (p1M*p1M)...[2x1]*[2x1]???...Essentially i believe all the casting vector to Matrix does is to merge the vectors as it is....
However, the command p1M.at<float> (0,1) and p1M.at<float> (1,0) returns 0 and 1 resp. this made me think p1M*p1M is possible but unfortunately it only compiles and generates a run time error:
OpenCV Error: Assertion failed (a_size.width == len) in gemm, file /home/james/OpenCV-2.3.1/modules/core/src/matmul.cpp, line 708
terminate called after throwing an instance of 'cv::Exception'
what(): /home/james/OpenCV-2.3.1/modules/core/src/matmul.cpp:708: error: (-215) a_size.width == len in function gemm
I am thinking of writing a function to just do that! vector to Mat and vice versa...Am i missing something?
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（1）
您可能想研究一下 cv::Mat::reshape
http://docs.opencv.org/modules/core/doc /basic_structs.html#mat-reshape
当您从一系列点创建 Mat 时，它会为点的每个组件创建一个通道。因此，如果您使用 Point3f，它会生成具有 3 个通道的单列 Mat。
您可以尝试通过调用将 p1M 转换为您期望的矩阵
You might want to look into cv::Mat::reshape
http://docs.opencv.org/modules/core/doc/basic_structures.html#mat-reshape
When you create a Mat from a list of points it creates one channel for each component of point. So if you use a Point3f it makes a single column Mat with 3 channels.
You could try converting p1M to the matrix you are expecting by calling