Abstract: Video analysis gives the detailed information about moving objects from the video. It is of great importance in the fields of sports, collision detection on road and rail routes, speed and acceleration calculation and many other fields. In this paper, we have compared four systems based on different algorithms and applied for different domains. First technique is for basketball video analysis, where semantic event extraction, slow motion replay detection methods have been introduced. Second, fire and smoke detection algorithm has been introduced. Third, the system uses surveillance cameras to detect the traffic state. Fourth, obstacle detection based railroad collision warning system. All these systems are diverse in nature with different application areas. The comparison showed that the system for sports have gained highest accuracy of all the other systems.

Keywords: Sports video analysis, fire and smoke detection, railroad obstacle detection, traffic state detection.