Development Of Intelligent Elevator Surveillance Camera System Based On DSP

As high-rise buildings become increasingly prevalent.The elevator is required to guarantee the highest safety.However,the high frequency of accidents is indicating that the traditional system based on manual monitoring cannot satisfy the current safety requirement.This thesis develops an elevator anomaly surveillance system for real-time monitoring,detecting abnormal situations and storing the videos abnormal situations for future investigation.Both features are of critical significance for elevator safety.This thesis develops an elevator anomaly surveillance camera system based on DSP,which utilizes TMS820DM8127 as the processor.Adopting modular design idea,the system is divided into video capture module,video encoding module,video analysis module,network communication module and alarm video storage and upload module.The video analysis module realizes the real-time alarm reporting function for various abnormal conditions in the elevator.By calling the relevant algorithm library to analyze video data of each frame and optimizing the preliminary result of algorithm through the self-designed strategy,the system can accurately and immediately give response to these abnormal conditions.The combination of capture module,video encoding module and network communication module,achieves the basic functions of surveillance cameras,such as image capture,1080p25 and D1/25 video stream by H.264 encoding and real-time video surveillance.In the alarm video storage and upload module,a caching mechanism is designed to store the alarm videos and pre-alarm videos so that the detailed information of the abnormal situations can be reviewed and further investigated.Through testing,the elevator anomaly surveillance system successfully realizes the real-time alarm reporting function for various abnormal conditions with accuracy maintained about 90%,along with other basic functions.Meanwhile,the algorithm processing time maintains about 79ms,meeting the requirement of real-time surveillance system.