Research And Implementation Of Infrared Camera Based Surveillance Method For Night Scenes

In recent years, people pay more and more attention to intelligent monitoring systems. Ordinary surveillance systems typically use color cameras, but the quality of video in night is poor,resulting in lower recognition accuracy. The infrared camera can not only achieve the normal monitoring during the day, but also launch infrared at night to implement monitoring. So the infrared camera is graduallyapplied in the monitoring system.This paper studies the night scene monitoring methods, including infrared image pre-processing technology, the adaptive improvements of infrared image vehicle detection method and vehicle tracking.Firstly, we used gray processing, denoising and contrast enhancement to process the infrared image.Secondly, we usedthe Part-based method combined SVM to achieve the detection of the target, and on this basis proposedthe adaptive detection algorithm in order to adapt to dramatic changes of the background illumination. We select the Kalman algorithm used for vehicle tracking. Finally, we designed and implemented a the nightdetection system based on infrared cameras, the experimental results show that this system can effectively identify and track vehicles at night.