Research On Image Dehazing Algorithm Based On Atmospheric Scattering Model

The foggy imaging system is very sensitive to outdoor work.The reflected light passes through different sizes of suspended particles and scatters before reaching the imaging device,resulting in low visibility and contrast of the image.There are immeasurable consequences in the field of image-dependent technologies such as transportation,military,and aerospace.Therefore,it is of great theoretical and industrial significance to improve the quality of foggy degraded images and restore original high quality images.Firstly,this paper makes a rigorous analysis of the images under outdoor haze weather.Based on the theory of digital image processing and optics,this paper discusses the influence of atmospheric scattering model on foggy images by analyzing the imaging mechanism of foggy degraded images,and introduces in detail the characteristics of foggy images and several image dehazing algorithms based on atmospheric scattering model.Secondly,this paper makes some improvements to the deficiency of the dark channel prior algorithm.The specific work is as follows: A fusion dark channel is proposed for the shortcomings of the filter windows,because the filter windows are too small to conform to the dark channel prior principle.By combining the minimum operation and the dark channel operation on two different sized neighborhood filtering windows,a fused dark channel image is obtained as a new dark channel image.This improvement improves the edge block effect of the original dark channel.In order to prevent the transmission rate from being too rough,this paper proposes an adjustable medium transmission rate.Indirectly solve the initial estimate of the transmission rate by white balance estimation,and make a difference with the original medium transmission rate as the adjustment factor of the transmission rate.Then formulate an adjustment parameter to obtain an adjustable medium transmission rate.This method improves the transmission rate is too rough and the halo phenomenon generated during the solution process.Aiming at the incorrect estimation of atmospheric light by the white objects in foggy image scene,an adaptive atmospheric light method is proposed.This misjudgment can be eliminated by opening the operation from white objects in the scene target,and the adaptive coefficient can be set to determine the estimation of atmospheric light.The adjustable medium transmission rate is optimized by improved guided filter,filtering the obtained medium transmission rate by using the gray image of the Gaussian filter as a guide image.This improvement improves the overestimation of atmospheric light values and results in more accurate transmission rate images.Finally,in this paper,the image defogging interface is used to realize each algorithm.The experimental results show that the improved algorithm can remove the foggy image.The experimental results are better than the original dark channel prior algorithm,at the same time,the defogging effects of several classical algorithms are compared and analyzed,and their effectiveness is verified by objective data.