Occlusion Handling Through Multi Agent System

Heindwick, George (2019) Occlusion Handling Through Multi Agent System. [Final Year Project Report] (Unpublished)

PDF (Please get the password by email to repository@unimas.my , or call ext: 082-583914/3973/3933) Heindwick George.pdf Restricted to Registered users only Download (2MB)

Abstract

Object occlusion is a phenomenon that often occur when tracking moving object and caused by the scene object or other targets which affect the perfect visibility of the target seen from the camera point of view. Tracking object under occlusion sense is a challenging task especially in multi camera environment which yield urgent, high dynamism and large scale problem. The uses of Multi-Agent system can solve this problem efficiently due to its nature in performing better with respect to hardware resources and software limitation when the three problem exist. The hierarchical structure of Multi-Agent system provide feature for information exchange and negotiation between agents within the same layer asynchronously lead to lower processing time because task is divided into several agents as compare to single agent. In occlusion handling, Kalman Filtering uses the prediction and update where the motion of tracked object is predicted and updated each frame. Background Subtraction is one of the method used in object tracking. Image processing may not be accurate all the time but the evaluation of tracking with human and tracking with multi agent has proved that the system has higher precision on either object tracking or occlusion handling as compare to human

Actions (For repository members only: login required)

View Item