Research On Image Defogging Processing Method Of Small USV Platform

At present,many researches in the field of computer vision focus on the follow-up processing such as the classification and target recognition of clear images collected.However,in real life,"haze" weather caused by environmental pollution or extreme weather frequently occurs.In this case,the quality of the image collected is not high,which affects the follow-up work.Therefore,it is of very high processing value and very strong necessity for foggy images to be defogged in advance,which is also a research hotspot in the field of image science and computer vision.First,in this paper,we first introduce the Unmanned Surface Vehicle(USV)and the characteristics of images collected on the platform.Based on the analysis of the imaging principle of foggy images,the atmospheric scattering physics model is established.Firstly,the establishment process of the single scattering atmosphere model is described,and the characteristics of the model are analyzed.Secondly,the relevant fog removal algorithm for sea scene is analyzed and studied.Related knowledge of dark channel prior is described,the main parameters involved in the algorithm are explained,and the restoration theory is applied based on the single scattering atmosphere model.This paper introduces the effects of multiple scattering caused by multiple scattering particles on imaging,the application of radiation equation and atmospheric point diffusion function in multi-scattering atmosphere model,and then explains the spot effect which cannot be described by single scattering model.After that,the improved fog removal algorithm is given.Aiming at the estimation of atmospheric light value,the accuracy of estimation value is increased by quartering method.In order to estimate the transmittance,the algorithm complexity and algorithm time are reduced by guiding filtering method.The subjective and objective evaluation indexes are introduced,and the corresponding comparative analysis is made on the haze map,basic algorithm restoration map,improved algorithm restoration map and standard haze free map.Finally,the flow and related design of video defogging algorithm are studied.Based on the relationship with the single image defogging algorithm mentioned above,the main parameter design methods involved are described.The algorithm is optimized from the running time according to the real time requirement in practical engineering.At last,the results are compared and analyzed.