Research On Improved Defogging Algorithm For Image And Surveillance Video

Under foggy conditions,due to the absorption,scattering and reflection of water droplets,particles,fog and haze suspended in the atmosphere,the light from the scene radiation is attenuated,resulting in severe degradation of the image captured outdoors,reduced image contrast,and color distortion.The lack of details and the degradation of outdoor surveillance video quality seriously affect the intelligent traffic monitoring system and people's daily life.The clear recovery technology of foggy image and video surveillance has important theoretical and practical significance.Based on the existing theory of dehazing algorithm,this paper proposes an improved dehazing algorithm for foggy images and surveillance video.The main research contents and specific schemes are as follows:for the dark channel prior algorithm,the overall color of the restored image will be dark,sky area is prone to emerge problems such as halo and color distortion.In this paper,an image dehazing algorithm based on open dark channel and optimized boundary constraint is proposed.The foggy image is divided into non-sky area and sky area,and processed according to different characteristics of each area.Finally,the two regions are merged to obtain a defog image.For night foggy images.,which have low contrast,uneven brightness and high noise,a nighttime image dehazing algorithm based on transmittance and Retinex algorithm is proposed.After the night foggy image is inversely transformed,the transmittance is solved according to the liuminance propagation map,and optimized by the multi-scale Retinex algorithm,and the anti-color night foggy image is restored by the atmospheric scattering physical model.Reverse again,remove the noise by bilateral filtering and use the partial block Shade of Gray color correctioin algorithm to get the final night defog image.Aiming at the fixed and real-time characteristics of surveillance video background,a surveillance video dehazing algorithm based on universal background method and frame difference method is proposed.Firstly,the general background image is obtained by statistical average method,and the corresponding transmittance image and atmospheric ambient light value are calculated.Then use the frame difference method to judge and update the general background image and atmospheric ambient light value in time.Finally,the restored defog frame images are sequentially connected to obtain the final defog video.At the same time,the algorithm is compared with other related dehazing algorithms.The experimental results demonstrate the effectiveness of the proposed algorithm from both subj ective and obj ective evaluation.