An Improved Automatic Zone Detection Algorithm based on Mechanic Flow for Complex Scene Event
Kannikar Intawong, Patcharaporn Prompantakorn, Waraporn Boonchieng · 2021
Due to the objects detection and an complex background in the scene, the accuracy and stability of the tracking algorithm are often interrupt. Therefore, The accurately objects tracking and zone detection in the scene are difficult problem in object tracking applications. A robust zone detection method is proposed to solve the problem of occlusion. It adopt weight ideas in mechanism design. In order to modulate the problems of background issue, object detection is firstly divide into blocks and the different weights are related with motion direction for each blocks. In addition, we proposed the weight mechanism for keeping the object and tracking when object disappear in the scene. Additional, The experiments represent that our algorithm can be achieved the performance in handling various complex videos surveillance both vehicles and persons detection.