Essentially, you need a library that will decode the video file and present you raw data for analysis. There are no libraries (of course, I still don’t know) that will allow you to directly track the high level (i.e. there will be no findBall () method !!!). You will almost certainly have to access the pixel values and analyze them yourself.
For this kind of thing, I would use something like opencv . This is a C ++ library that (1) allows you to access raw pixel data (as well as motion vector data, etc.), (2) provide you with some utility methods for analysis, and (3) allow graphics to be overlaid on top of a video clip .
, .. . , , , .
[edit]:
, , , . Hough Transform - (openCV ). () (.. ..). , , . , , . , , .
( ), - , .., , , : " , (), (, ), / ( )".
[/]