The Implementation of a Real-Time Image Surveillance System with Abnormal Sound Detection

• Main Authors: Chao-Sheng Chao, 趙昭勝
• Other Authors: Chau-Ren Tsai
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/13256629165082229888

碩士 === 國立臺灣科技大學 === 電機工程系 === 97 === In traditional surveillance systems, which use a wide-angle lens with field camera and store image all day. It can’t grab target’s face clearly for identification and storing image all day, because of it will waste of hard-disk’s resources. In the system, using high-resolution PTZ camera to be in conjunction with field camera and can track targets. And abnormal sound detection, whether abnormal event happen or not and store event-image, it will save many hard-disk resources. This research will combine TI TMS320DM642 Evaluation Module to development surveillance system in Server-side, the system has stand-alone characteristic and developing the graphical user interface based on PC in Client-side, Server-side will transmit the alarm information and image to the Client-side automatically. When the abnormal events happen in Server-side and it will transmit the alarm information and store the event-image in time, if necessary it will able to browse the event-image. Thereby, the system develops a real-time multi-object tracking and abnormal sound detection system with online warning function.