Opencv learning notes (9)-self-implemented watershed algorithm and tested with opencv

In order to verify the algorithm in the previous blog, you can make a slight modification by using the image structure in opencv. For simplicity, using mat. at <uchar> (I, j) for image operations is certainly inefficient. For more information, see http://blog.csdn.net/caiye917015406/article/details/7791815133. But here is only the test result...

In the algorithm, manually select the threshold value for binarization of the image, which has a great impact on the effect of the watershed algorithm. We have also used cvadaptivethreshold in opencv to automatically select and seek, but the effect is not as good as it is and we look forward to improvement...

The following is the test code.

# Include <iostream> # include <queue> # include <vector> # include <cv. h> # include