Multiple moving object monitoring in robot vision

In this study, the researchers propose and implement an algorithm as an autonomous robot having a stationary camera to monitor many targets (moving objects). The robot uses apportioning to grasp position of all moving objects. Through apportioning, the scene is parted into three major regions while each region has different sub-regions. The status of the robot system can be in three modes of unsafe mode, safe mode and over-safe mode based on the position of all targets. Regarding each mode, the robot accomplishes suitable movement to pursue monitoring of the targets. All movements of the robot are based on two rules. First rule focuses on maintaining the system in the safe mode, while the second rule pays attention to fairness. The robot is implemented and tested in indoor with different factors such as FPS, resolution, minimum number of frames for motion detection, and minimum area of change for motion detection.