Video Processing Algorithm Development Of Intelligent Elevator Surveillance Camera Based On DSP

High-rise buildings with elevators have become an important component of modern society.The frequency of elevator accidents makes the safety of taking an elevator face a severe challenge.The traditional way of manual surveillance can not guarantee the safety of taking an elevator.Therefore,this thesis develops a video processing algorithm of intelligent elevator surveillance camera that realizes the real-time surveillance of the elevator and is able to give an alarm response to abnormal conditions in time.This algorithm is very meaningful to the safety of taking an elevator.This thesis develops a video processing algorithm of intelligent elevator surveillance camera based on DSP which uses TMS320DM8127 as the processor.First of all,this thesis realizes such functions of the camera as video encoding of different resolutions and image capture.Furthermore,this thesis also develops an elevator door state monitoring algorithm which is able to detect the current state of elevator door correctly and give an alarm response to abnormal conditions in time.This algorithm uses ViBe and KCF to locate the current position of the door,and then gets the correct state according to the displacement and the running characteristic of the elevator door.A series of rules are set to detect abnormal conditions.Eventually,the thesis implements the elevator door state monitoring algorithm on the DSP core of TMS320DM8127 and optimizes it by such means as C code and compiler to improve the time performance.After testing,the video processing algorithm of intelligent elevator surveillance camera has realized the functions of image capture and providing HD videos of 1080P25 and D1/25 by H.264 encoding.The algorithm is also able to get the state of elevator door correctly and give an alarm response to abnormal conditions in time.Meanwhile,the algorithm meets the requirement of real-time intelligent surveillance system.