I read double type data and I want to normalize it.
But before that I need to round off , I mean I want to remove digit under the point.
in C++, there is floor and ceil , but it doesn't work in a matrix.
Please tell me is there ceil or floor function for a matrix.
Thnx
Example 1)
Mat srcMatrix(Size(256, 256), CV_64FC1, matrix);
Mat dstMatrix(srcMatrix.size(),CV_8UC1);
double scale=1.0;
double offset=0.0;
//dstMatrix = srcMatrix * scale + offset
srcMatrix.convertTo(dstMatrix, dstMatrix.type(),scale,offset);
Example 2) you could avoid to specify size and type for dstMatrix:
Mat dstMatrix;
srcMatrix.convertTo(dstMatrix, CV_8UC1,scale,offset);
in both cases convertTo performs round (not trunc/ceil). See also this answer.
Converting a 64bit floating point matrix into a integer matrix is quite straightforward.
Mat matrix1(Size(256,256), CV_64F, matrix);
Mat matix_reduced_precision;
matrix1.convertTo(matrix_reduced_precision, CV_8U);
Asked: 2016-06-30 12:47:25 -0600
Seen: 3,969 times
Last updated: Jul 01 '16
Ow just convert your double matrix into a int8 or int 32 matrix and it will round for you :)
i'm looking for how to define it.. in http://docs.opencv.org/2.4/modules/co...
Mat matrix1(Size(256, 256), CV_64F, matrix); //change array to matrix
I made matrix which is defined like that, matrix is double type number.
If it is ok could u tell me how to write to int8 or 32..?
thnx a lot :)