Software Design Of Sound Detection And Alarm Function For Video Surveillance Terminal System

With the continuous development of society and the increasing demand for security, video surveillance system has been widely used. After the digitalization and networking, video surveillance system is moving in the direction of intelligence. At present, the intelligent analysis of video image sequence is the main means in intelligent surveillance system, while audio signal has not been fully utilized. The abnormal sound signal in monitoring sites can reveal and characterize the abnormal state and sudden accidents, so it has high engineering application value to achieve sound detection and alarm functions for video surveillance terminal.This thesis designed and implemented the software of sound detection and alarm function for video surveillance terminal system basing on TMS320DM368processor platform. Based on module design theory, the whole software was divided into several relatively independent functional modules, and a communication mechanism was designed to exchange messages between different modules. Aiming at the high requirements of real-time and accuracy in terminal system, this thesis designed a fast and effective sound detection scheme. This scheme acquires real-time sound signal with audio A/D converter, detects endpoints of sounds based on short-time magnitude and short-time threshold-crossing rate, extracts mixed characteristic parameters including short-time magnitude and MFCC (Mel Frequency Cestrum Coefficient), and recognizes sounds based on improved DTW (Dynamic Time Warping) algorithm. Experimental results show that the recognition time is less than1second and average recognition rate is89.3%. This Thesis also implemented alarm module, with the functions of real-time alarm detection, report generation, plan management, alarm event reporting, alarm linkage actions and so on. The entire software has been tested, and the result showed that this software met the design requirements.