Automatic Detection Approach of Basketball Track in Sports Video

Video image track of basketball flight path can precisely analyze flight trajectory curve in basketball so as to direct basketball shooting training. Thus, this paper puts forward image video sequence analysis and adaptive objective model-based flight trajectory tracking method of basketball. This algorithm fully utilizes correlation of time and space between adjacent blocks. On one hand, in order to overcome effect of basketball rotary motion, matching criterion which is not related to regional location is set up; on the other hand, diamond movement search path is improved based on sport correlation of basketball and center-biased characteristics in order to reduce search times and improve tracking speed. The experiment results show that methods in this paper can effectively track basketball under complex background and have effective robustness of basketball rotary movement and partial occlusion.