Research On Image Dehazing Based On Passive Millimeter-wave And Visible Image Fusion

Under hazy weather,the outdoor visual system is affected by the atmospheric particles,which results in problems of blur,low contrast and color distortion of perceived image.The study of image dehazing algorithm helps to improve the performance of outdoor vision system and make it robust in adverse weather conditions.It is of great significance to the development of automated image / video processing technology in computer vision system.Based on passive millimeter-wave(PMMW)and visible image fusion,this paper proposes a new image dehazing method which further improves the image dehazing effect.Image degradation mechanism under hazy weather is the basis for image dehazing.Based on the theory of atmospheric scattering,we introduce the incident light attenuation model and atmospheric light imaging model.The two models above constitute the foggy image degradation model,which lay the physical foundation for the subsequent dehazing algorithm.Secondly,millimeter wave radiation theory shows that the millimeter-wave has a strong penetrating power to haze.Therefore,the millimeter wave imaging system is less affected by haze and is an effective dehazing method nevertheless with a limitation of resolution.We propose a dehazing method based on the fusion of millimeter wave and visible image.The fusion process is realized by Laplacian Pyramid decomposition and reconstruction.The fusion image has more abundant information as dehazing object,which make up for the loss of the scene information in the visible foggy imageFinally,the fused image is dehazed based on the foggy image degradation model.The dehazing process involves the solution of two unknown parameters in the model: the transmission is solved quickly by regarding the millimeter-wave image as a haze-free image;a dark channel prior algorithm is used to solve atmospheric light values.Then the scene radiation is restored.In addition,we perform a comparative analysis between our experimental results and some existing dehazing methods to further verify its effectiveness.