Video surveillance system

• Main Author: Tay, Edwin Si Hao.
• Other Authors: Chong Yong Kim
• Format: Final Year Project
• Language: English
• Published: 2009
• Subjects: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision
• Online Access: http://hdl.handle.net/10356/17207
• Institution: Nanyang Technological University

Surveillance systems are an integral part of many safety/security plans today, and the tracking of moving objects plays a significant role to reduce the amount of effort required by the security personnel. Although the costs of storing these critical information have decreased with the advancement of technology, the systems can be streamlined by adopting more advanced surveillance systems. These systems can store only the important information in higher resolution, while using a lower bit rate for the background image. The object detection algorithms could do with some code optimization, to improve the speed of the processing, so as to increase the frame rate. This project implements a video surveillance system using an ADSP-BF561 dual core processor, which is able to process both NTSC and PAL video format, and display the output through a TV screen with analog input. The proposed surveillance introduces a minimum bounding box to estimate the object detection region by taking in the location, height and width parameters. This will reduce the amount of processing required as compared to a wrap-around curve to show the object detection. The accuracy of the system is also improved by allowing the user to refresh the reference frame, so as to prevent errors that arise due to changes in the environment of the scene.