| As the fog haze weather frequently occur,there is important significance of the images defogging fast and efficiently in the field of video surveillance.The images obtained by surveillance equipment in haze,having poor contrast and color fidelity in fog weather.It must be difficult for surveillance equipment to estimate the objective using images,and it directly limits the role of the security monitoring system.This dissertation for the images defogging in the field of video surveillance.We analyze the image degradation causes,researched on the basic theories,and built a hardware platform based on star-level camera.This dissertation made some meaningful work for building up the living defogging method and introducing new algorithm.,As follow,introduced the research contents of this paper:1.Based on the surveillance front-end equipment,design a hardware platform based on star-level camera for images defogging.In order to actually accelerate the quality of fog-degraded image,built a reliable hardware platform for image acquisition enhancement,transmission.2.This paper research on the arithmetic based on dark channel prior.In order to reduce the time-consuming,guided-filter are usded instead of soft-matting in this algorithm.We can make the algorithm more concise by useing guided-filter instead of soft-matting.by this way,it greatly reduced the complexity and the consuming time will be drastically reduced.In the optimization with the global atmospheric light estimation,this dissertation also proposes to replace the mean value of the original method of selecting a point in the input image as the atmospheric light value of the whole picture.The results of experiment on haze images demonstrated that the method can fast efficiently and fast raise the parallel and color sharpness of fog degenerative images,and effectively raised the visibility of the video surveillance in haze.3.This paper study for the adjustable parameters of the algorithm,that is used in dark channel prior and guided-filter algorithm.We analysis the influence on the algorithm by the simulation result.4.In order to raise the real-time performance of the algorithm,the image interpolation method are used to reduce the time to calculate the image transmission rate.When we calculate the image transmission rate,we reduce the picture by image interpolation method firstly,and then we calculate the image transmission rate whicth was reduced.We can estimated the image transmission rate by enlarge the image.By the simulation result of timing sequence,we can see,the computation time is less than the previous algorithm. |
