Study On Image Quality Improvement Of Laser Active Imaging In Fog

Foggy weather has great influence on our daily life and economy. Imagingthrough the fog could be realized using infrared laser active imaging system.However, the image obtained by the system has the features of low contrast andserious noise. Image processing technology could effectively improve the imagequality and has the advantages of low cost and easy realization. In this paper, imageprocessing technology was studied and adopted to improve the image quality oflaser active image system in fog.Firstly, laser active imaging system was analyzed integrally. The absorptionand scattering during the propagation of laser in atmosphere have great effect on theimaging distance and image quality. Range-gated technology and image processingtechnology could significantly overcome the effect of atmosphere and the later hasobvious advantages. On the basis of above analysis, the composition and parametersof the system were introduced, the imaging distance formula was derived and usedto estimate the imaging distance range of the system in certain visibility conditions.Secondly, the noise characteristics of infrared image were studied. Thetraditional infrared image enhancement methods and image dehazing methods wereintroduced briefly. Bilateral filtering, Retinex theory and image dehazing methodbased on dark channel prior were discussed in detail.Thirdly, the influence of weather condition on active imaging system wasstudied experimentally. Different methods were adopted to process the foggyinfrared image. Both objective evaluation and subjective evaluation of the processedimage show that multi-scale Retinex and dark channel prior could obtain betterresults compared with traditional methods. The standard deviation was more thandoubled and the information entropy and information capacity had about a twentypercent improvement. Thus, multi-scale Retinex and dark channel prior are effectiveways to improve the image quality of laser active imaging system in fog. Moreover,dark channel prior technology has the potential of real-time application. The visualeffect could be further improved, if the dehazed image was processed by bilateralfiltering. However, the process is time consuming.Lastly, the summary and conclusion of this paper were given. The existingproblems and the further research work were also discussed.