Research On Intelligent Video Surveillance System Based On Embedded Platform And OpenCV

With the continuous development of maritime transport,the volume of maritime traffic is increasing year by year and the key role of ship transport in the field of transportation is becoming increasingly evident.However,in recent years,ro-ro shipwreck accidents have taken place frequently due to the structural design defects,the particularity of the cargo transported and the bad sea conditions.Therefore,correctly and effectively raising the level of supervision on the safe operation of ships,improving the safety monitoring system and giving timely warning of the danger of disasters have extremely important practical significance for guaranteeing people's lives and property.This project originates from the project of "China Shipping Passenger Ship Safety Monitoring System".It mainly focuses on the weak link of video surveillance of passenger ro-ro ship,designs and develops an embedded intelligent video monitoring system based on computer vision library OpenCV to realize the coastal route passenger service from Dalian to Yantai.The intelligent video surveillance of the rolling-roof vehicle cabin warns the police promptly when the vehicle in the vehicle cabin is shifted to the warning area of the vehicle cabin and notifies the supervisory staff to eliminate the danger.The main work of this paper is as follows:(1)Conducting a field survey on the "Huludao" ro-ro passenger ship of China Shipping Passenger Ship Company,selecting the OK6410 development board and embedded Linux system based on ARM 11 as the system software and hardware development platform,and setting up the software development platform,including embedded Linux system transplantation,transplantation of OpenCV and Qt function library.(2)Three displacement detection algorithms,namely the inter-frame difference method,the standard Gaussian mixture background difference method,the improved Gaussian mixture background difference algorithm are simulated and their advantages and disadvantages are briefly analyzed.It is proposed that the inter-frame difference and the improved Gaussian mixture background difference.The combined algorithm can not only detect the displacement vehicle completely,but also improve the running efficiency compared with the standard Gaussian mixture algorithm.The Kalman filter algorithm is added to the improved algorithm to improve the real-time performance of the vehicle displacement detection algorithm.(3)The software of intelligent video surveillance system is designed and developed.The application software mainly includes five modules,namely vehicle module video acquisition module based on V4L2 driving architecture,vehicle displacement detection module based on open source computer library OpenCV and perspective transformation principle,alarm module,network communication module and video display module.Finally,the developed system is simulated in the laboratory and recorded in the actual parking video test.The test results show that the system design is feasible and meets the corresponding functional requirements.