A Study Of The Dual Cameras Collaborative Video Surveillance System

In today’s world in which the level of people’s life is increasing day by day, the demand on video surveillance system is becoming higher and higher. All sorts of industrial areas have enhanced the consciousness of safety and protection. Undoubtedly, the dual cameras video surveillance is an effective way of security. According to the situation without human’s intervention that moving targets could be self-analyzed by images taken by the camera, the paper proposes a design of dual cameras collaborative tracking system based on DE-2 development board. Mainly using the way combining panoramic video cameras and intelligent high-speed ball machine (PTZ), it can make real-time monitoring to the scene effectively. The research can be divided into the following contents.In terms of moving target detection, the paper compares the difference among adjacent frames method, three frame differential method, adaptive background difference and gaussian mixture background, and implements their simulation results. By using the method of experimental contrast, the author judges advantages and disadvantages of the four algorithm respectively to improve the operation efficiency of the entire system. In target tracking, the paper deducts and discusses the target tracking process of the Mean shift algorithm and kalman filtering algorithm in detail, gives the experimental results and analysis of the Mean shift algorithm in target under the slow and rapid movement, improves the kalman algorithm to solve the problem of failed tracking under the rapid movement condition of the target and shows the superiority of the algorithm.In terms of dual cameras collaborative tracking, the construction of the system has the following characteristics. The panoramic camera and intelligent high-speed ball machine connect to the audio and video acquisition card through video line, then transmit and storage the collected image and the recorded video; Intelligent high-speed ball machine connects to the video tracking server by 485 communication bus and realizes the comprehensive control of the ball machine pan-tilt; It displays monitoring image in real time taken by two cameras connecting the monitor and video server through VGA line; At the same time, it shares the image in the system by using the switch equipment setting several PCs and video tracking server on the same network segment. This system has good effect in realizing the dynamic target detection and tracking in target area. Drawing lessons from a number of successful cases, it innovates in different levels in the design of hardware and writing of software.Finally,by achieving verification and analysis of the system, the dual cameras collaborative tracking system is a kind of optional intelligent monitoring and tracking system in real production and life, which has certain practicability.