Based On Embedded Vxworks Under High Dynamic Range Radiation Image Processing

With the develpoment of embedded system technology,applications based on embedded system are more and more and bring great conveniences to the intelligent earth. Recently,the intelligent video analysis based on embedded system has experienced a rapid development,has became a hot spot in the field of video research and has been widly used in intelligent transportation.VxWorks system has a great performance of stability and real time and can robustly work in adverse circumstances.At the same time, high dynamic range images are more realistic than low dynamic range images and have great application value.Now high dynamic range radiance images are obtaining a wide attention.In the article we focus on the processing of high dynamic range radiance image in the embedded VxWorks and give a new method that is more suitable to the processing of high dynamic range radiance image and gets a better effect of image enhancement.Attention were concentrated on:Research the processing of high dynamic range radiance image.First,we analyse the reason which causes the high dynamic range radiance image,the existing problems and the ways that completly record the high dynamic range radiance image of the real scene.Second,we analyse and summarize the current classic peocessing methods,At last,we present a new Retinex image enhancement method based on curvature bilater filter.Experimental results demonstrate that our method can efficiently reduce halo artifacts and has a better peocessing effect among the dynamic range compression,the contrat enhancement and the radiation intensity weakening around the light source. The enhancement image is closer to the real scene.Analyse the system property of embedded VxWorks and the principle of USB device and drive the Logitech’s USB camera in embedded VxWorks.Realize the video capture of USB camera in embedded VxWorks.This part mainly transplants the open source program Gspca from Linux and researchs the communication among the multiple processes in embedded VxWorks.