cvReshape用法详解

cvReshape用法详解

修改矩阵的形状

下面的代码是将单行的矩阵转变成二维矩阵

/* Changes a shape of multi-dimensional array.
   new_cn == 0 means that number of channels remains unchanged.
   new_dims == 0 means that number and sizes of dimensions remain the same
   (unless they need to be changed to set the new number of channels)
   if new_dims == 1, there is no need to specify new dimension sizes
   The resultant configuration should be achievable w/o data copying.
   If the resultant array is sparse, CvSparseMat header should be passed
   to the function else if the result is 1 or 2 dimensional,
   CvMat header should be passed to the function
   else CvMatND header should be passed */

CvMat* mat_a = cvCreateMat(1, HEIGHT*WIDTH, CV_8U); //单行矩阵便于赋值操作  
CvMat* mat_b = cvCreateMat(1, HEIGHT*WIDTH, CV_8U); //同上  
cvReshape(mat_a, mat_a, 0, HEIGHT); //把单行矩阵整形为二维矩阵  
cvReshape(mat_b, mat_b, 0, HEIGHT); 

cvReshape只生成一个新的矩阵头，而数据都指向原来的地址，所以是两个矩阵共有一组数据，这一点在使用中要注意，原来的数据撤消是否会影响后生成的矩阵的使用

cvReshape是按行形成向量，如果想按列形成向量，就先调用cvTranspose对矩阵进行转置,再调用cvReshape

• 代码

CVAPI(CvArr*) cvReshapeMatND( const CvArr* arr,
                             int sizeof_header, CvArr* header,
                             int new_cn, int new_dims, int* new_sizes );

#define cvReshapeND( arr, header, new_cn, new_dims, new_sizes )   \
      cvReshapeMatND( (arr), sizeof(*(header)), (header),         \
                      (new_cn), (new_dims), (new_sizes))

CVAPI(CvMat*) cvReshape( const CvArr* arr, CvMat* header,
                        int new_cn, int new_rows CV_DEFAULT(0) );