| In recent years,remote sensing technology has been widely used in land resources survey,disaster relief,geological disaster detection and other fields,bringing great convenience.However,due to the imaging mechanism of optical sensors,images collected outdoors are easily affected by hazy weather.The image will show reduced contrast,reduced sharpness,and even some details disappear.And with the advancement of technology,high-resolution image equipment is increasingly applied to production and life,especially in the field of remote sensing.Although there has been some progress in the research of fast dehazing algorithms for conventional images,the progress is slow and it cannot be effectively applied to fast defogging algorithms for large-size,high-resolution images of remote sensing images.Therefore,based on the idea of fast de-fogging algorithm of conventional images,this paper deeply studies the mechanism behind fast de-fogging algorithm and proposes a fast de-fogging algorithm for remote sensing images.This paper first studies the dehazing algorithm based on the priori theory of dark colors,and compares three different methods of transmittance refinement,and proposes to use guided filter method instead of the "soft map" method.The degree of transmission thinning and dehazing effect were studied.It was found that the smoother the transmittance,the higher the image quality after defogging.Then,based on the research of the basic theory of image defogging basic theoretical results and the existing quick defogging algorithm for conventional images,a new transmittance is proposed for the fast defogging algorithm in which the transmittance is not sufficiently smooth after the upsampling.The method of sampling restoration,and the problem of the long time-consuming problem of the improved upsample recovery method,proposes an improved steerable filtering method.And for the situation that the atmospheric light value estimated by its atmospheric light value estimation method is susceptible to a single pixel,the method of using the average value of the atmospheric light value candidate values as the atmospheric light value is proposed.Secondly,the relationship between the time-consuming defogging process and the restoration of image quality after defogging and defogging is further studied.It is found that there is a strong linear correlation between the size of the image to be defogged and the minimum time required for the quick dehazing process.Moreover,when the size of the image after the downsampling is less than a certain threshold,the process time of the quick defogging algorithm can reach the minimum time-consuming.On the other hand,the entropy value of the defogged restored image is related to the entropy value of the transmittance map participating in the arithmetic processing.The smaller the entropy value of the transmittance map is,the larger the entropy value of the restored image is.Based on the above findings,this paper presents an adaptive fast defogging algorithm for remote sensing images.Experiments show that compared with the existing fast defogging algorithm,this method can effectively reduce the defogging and time-consuming of remote sensing images and has better defogging effect. |
