Check out the OpenCV API . This is what he says on page 344, where FindContours described:
The function extracts the contours from the binary image using the algorithm [19]. Contours are a useful tool for analyzing the shape and object of detection and recognition.
If you look in the bibliography at the end of the document, this is link 19:
S. Suzuki and K. Abe. topological structural analysis of digitized binary images along the border. CVGIP, 30 (1): 32-46, April 1985.
Sounds like a good starting point. In addition, there is always a source if you want detailed information about the implementation.
misha source share