如何在C ++中找出OpenCV .at函数的用途？

Question

Is there a simple reliable way to find out what equivalent type I should use for the .at function for a Mat of a given CV type?

For example, how can I tell that the blanks should be filled with ushort , float and Vec3b ?

Mat mat1(1, 2, CV_16UC1, 12345);
std::cout << mat1.at<___>(0, 1) << "\n";
Mat mat2(1, 2, CV_32FC1, 67.89);
std::cout << mat2.at<___>(0, 1) << "\n";
Mat mat3(1, 2, CV_8UC3, Scalar(65, 66, 67));
std::cout << mat3.at<___>(0, 1)[2] << "\n";

Answer 1

Using data from

• Types: http://ninghang.blogspot.de/2012/11/list-of-mat-type-in-opencv.html
• Vectors: http://docs.opencv.org/modules/core/doc/basic_structures.html#vec

I constructed the following table:

        C1      C2     C3     C4     C6 (not sure whether C6 subtype exists as a macro)
CV_8U   uchar   Vec2b  Vec3b  Vec4b
CV_8S   char    -       
CV_16U  ushort  -
CV_16S  short   Vec2s  Vec3s  Vec4s
CV_32S  int     Vec2i  Vec3i  Vec4i
CV_32F  float   Vec2f  Vec3f  Vec4f  Vec6f
CV_64F  double  Vec2d  Vec3d  Vec4d  Vec6d

with uchar == unsigned char and ushort == unsigned short

for missing types you could create your own typedef if you want to:

typedef Vec<ushort, 2> Vec2us;

or you just access it with am type of same size ( Vec2s ) and convert it afterwards

But in the end I think understanding what the format means (32 bit floating point number with 3 channels = 3 float values per matrix element) if much better than looking at a table...