Realtime flicker removal for fast video streaming and detection of moving objects

Abstract High-speed cameras are used in computer vision systems to track balls, shuttlecocks, or players many different sports. Collected information is for statistics, as coaches’ and players’ aids improve technique tactics referees’ verify their decisions enrich television broadcasts. Sports arenas which games played often equipped with lights generating flickering effects captured movies. A fast yet effective enough algorithm necessary remove so movement detection object tracking algorithms could be used. In this paper, we propose a flicker removal working an online filter on frame streams at speeds exceeding 200 frames per second. Most of the solutions found literature concentrate effectiveness accuracy not speed operation. contrast, our original solution designed mind sufficient before calculating differential detect streams. Our adaptive works when lighting conditions changing (new light sources) performs well various sources that causing flickering. The results experiments carried out show high method implemented CPU GPU, allowing objects interest preliminary applications commercially offered instant review system badminton.