Loading...
| Friend's email | |
| Your name | |
| Your email | |
| enter code | |
This page was sent successfuly
- Type of Document: M.Sc. Thesis
- Language: Farsi
- Document No: 46229 (19)
- University: Sharif University of Technology
- Department: Computer Engineering
- Advisor(s): Kasaei, Shohreh
- Abstract:
- Today, advances in technology caused the emergence of a need for intelligent systems which can operate without human supervision, and have variety of applications in security, industry, transportation and sport events. Most of these systems require faster and more accurate object tracking. Lots of surveillance and computer vision applications need a process to detect and track objects. For example, the exact location of objects or people is part of the input for applications like movement analysis, counting the number of items, and recognition of people’s behavior. Occlusion is one of the major obstacles in performance of auto-tracking systems. There are several methods to address this issue. One of the most conventional and at the same time powerful approaches is based on multi-view interactive cameras.However, placing more cameras will add to the complexity. This emphasize the demand for auto-tracking algorithms, since tracking becomes an impossible task for the user in such complicated systems.
In this report, a novel and effective method is proposed for accurate person detection in multi-view systems with stationary observing cameras and partially overlapping views. As the first step, background subtraction is implemented by approximate median method.It follows by projecting the voxels to images of previous step to create the visual hull.Then occupancy volume is built with respect to quantities of the voxels in vicinity. Finally occupancy map from top view will be formed. This approximates the initial location of the target, and it is detected based on a local density mass. Knowing the number of targets, we assign a particle filter for each one of them, and do the tracking by the data gathered using background separation of each frame. This online approach is independent of visual features and homography. Top view tracking greatly handled occlusion. Furthermore, we have considerable improvement in accuracy and precision in this proposed method - Keywords:
- Occlusion ; Object Detection ; Object Tracking ; Dynamic Target Tracking ; Multiview Cameras
Digital Object List
-
محتواي کتاب
- view
Bookmark