Dehazing Method for the Remote Sensing and Positioning Image of UAV with Bright Area
As the traditional dark channel prior(DCP)dehazing algorithm processes the remote sensing positioning image with hazeof UAV,the color of such bright areas as the sky or white is prone to distortion,the overall contrast of the image is reduced,and the areas of close-up scenery in the image are prone to such problems as noise patches and blurred details after processing.An adaptive threshold segmentation dark channel prior dehzaing method is proposed.Firstly,the grayscale image Igray(x)is used to obtain the adaptive threshold ThrB of the bright and non-bright areas of the image.Then,according to the adaptive threshold ThrB,the bright and non-bright areas on the image can be segmented,and the adaptive correction function M is designed for these regions.Finally,the coarse estimate of the atmospheric dissipation function generated by the dark channel image is optimized,and the transmittance is refined again with the bilateral filtering to complete the image dehazing restoration.The experimental results show that the proposed method can effectively avoid the problem of color distortion after image restoration when the sky area or the bright area with strong reflection is processed,and can further improve the processing effects of the ground area in the remote sensing image.After restoration,the color saturation and contrast of the entire remote sensing image are obviously improved,the subjective visual effect is improved to a certain extent,and the objective parameters such as PSNR,FC,SSIM,and CR are all improved,which is conducive to the subsequent remote sensing positioning image analysis.
万方数据
维普
