Learning OpenCV 3

Chapter 4. Images and Large Array Types

Dynamic and Variable Storage

The next stop on our journey brings us to the large array types. Chief among these is cv::Mat, which could be considered the epicenter of the entire C++ implementation of the OpenCV library. The overwhelming majority of functions in the OpenCV library are members of the cv::Mat class, take a cv::Mat as an argument, or return cv::Mat as a return value; quite a few are or do all three.

The cv::Mat class is used to represent dense arrays of any number of dimensions. In this context, dense means that for every entry in the array, there is a data value stored in memory corresponding to that entry, even if that entry is zero. Most images, for example, are stored as dense arrays. The alternative would be a sparse array.  In the case of a sparse array, only nonzero entries are typically stored. This can result in a great savings of storage space if many of the entries are in fact zero, but can be very wasteful if the array is relatively dense. A common case for using a sparse array rather than a dense array would be a histogram. For many histograms, most of the entries are zero, and storing all those zeros is not necessary. For the case of sparse arrays, OpenCV has the alternative data structure, cv::SparseMat.

Note

If you are familiar with the C interface (pre–version 2.1 implementation) of the OpenCV library, you will remember the data types IplImage and CvMat. You might also recall CvArr. In the C++ implementation, ...