Motion JPEG based surveillance system

• Main Author: Koh, Wan Ling.
• Other Authors: Chong Yong Kim
• Format: Final Year Project
• Language: English
• Published: 2009
• Subjects: DRNTU::Engineering::Electrical and electronic engineering::Computer hardware, software and systems
• Online Access: http://hdl.handle.net/10356/18025
• Institution: Nanyang Technological University

Technical and experimental information pertaining to the development of an intelligent video surveillance system is detailed in this report. This smart system is built on the Texas Instruments TMS320DM642 Video/Imaging Fixed-Point Digital Processor. Important concepts relevant to Motion JPEG, such as Discrete Cosine Transform (DCT) and color space Models, are entailed in this report. Various thresholding approaches are also investigated and presented to enhance understanding of the workings of motion detection. Morphological operations and edge detection techniques are also looked into. This project aims to amplify the video surveillance system to a more intelligent one that is capable of capitalizing on the large amount of video surveillance data, especially in the security industry. Various methods to spot intruders and separating them from the background model are explored to conclude an appropriate one for implementation on the Texas Instrument Processor. The output display is sub-divided into 4 different frames, exhibiting the original video data, the difference image calculated based on the sub-band DCT image, binary threshold image and lastly, the segmentation image after integrating morphological operations and minimum bounding rectangle algorithms.