The detection of multi-objective billiards in snooker game video

Abstract

In this paper, we proposed a kind of multi-objective detection technique based on watershed in snooker game video. It can automatically extracts target markers without any man-machine interactions. As to target that has similar color with background, we make full use of its reflective region to detect it indirectly which reduces burden of detecting directly. Meanwhile, it can effectively separate gathered targets especially like gathered crimson balls. At first, for the single color of billiard table, we use color segmentation, Hough transform algorithm to extract pool table so as to get rid of the interference of other objects outside the pool table; Second, combine color segmentation with watershed algorithm to extract all non-green objectives; Then according to the relationship of reflective point and the target itself, we can pick up the green one. At last, we use the difference between color segmentation label and number of reflex points of target area to segment all gathered goals. Experiments show that the proposed algorithm can effectively detect multi-objects in snooker game video.