Research On Sand Dust Image Enhancement Algorithm Based On Color Deviation Correction And Color Restoration

Sand dust weather is a disastrous weather phenomenon in which the wind rolls dry dust and sand particles in the atmosphere,causing great inconvenience to people’s travel.In addition,when shooting an image outdoors,the image will appear yellow as a whole,the image details are not prominent,and the foreground and background are not separated clearly,leading to image target information hiding and less useful information obtained,which will have many impacts on video monitoring and image analysis and processing.In view of these problems in the sand dust image,this thesis designs corresponding algorithms for image clarity research:(1)Aiming at the problems of serious image color offset and low contrast in dust weather,a sand-dust degraded image enhancement algorithm based on histogram equalization and multi-scale retinex with color restoration(MSRCR)is proposed,which mainly restores the image through two steps: color deviation correction and image enhancement.Firstly,after preprocessing each channel of RGB image,the corrected image is processed by using the contrast limited adaptive histogram equalization method(CLAHE);Then,the image is denoised by bilateral filtering.MSRCR algorithm is used to further solve the problem of color imbalance;Finally,because the contrast of the processed image is still low and there is still a certain color deviation,the final result is obtained by further processing with gamma correction and color deviation detection and color correction methods based on image analysis.Through the simulation of a large number of sand dust degraded images,the results show that the algorithm can effectively process sand dust images with different color deviation degrees,not only improve the image contrast,but also effectively avoid the phenomenon of image color offset.In addition,compared with other comparison algorithms,the average time efficiency of this algorithm is improved by 73.5%.(2)To address the problems of low time efficiency of existing de-dusting algorithms and insignificant image contrast enhancement affected by haze,this thesis proposes a de-dusting enhancement algorithm to improve saturation and contrast in HSI space on the basis of color correction.Firstly,the green channel is used to correct other color channels in RGB color space to achieve color balance of sand dust images and complete color correction.Secondly,the adaptive saturation enhancement algorithm is designed for the S channel using the property that each channel in HSI space does not interfere with each other,and the contrast limited adaptive histogram equalization(CLAHE)algorithm is used to enhance the contrast for the I channel.Finally,the combined channels are transferred to RGB color space to obtain a clear dedust image.The experimental results show that the de-dusted images have better performance in both subjective and objective quality evaluation.The proposed algorithm can remove the influence of haze,and has greatly improved the clarity and contrast restoration of image details.Moreover,compared with other better de-dusting algorithms,the algorithm in this thesis has significant improvement in time efficiency,and the average algorithm efficiency is improved by about 77% based on the original algorithm,which ensures the real-time performance of the algorithm.(3)A sand dust image clarity algorithm based on improved dark channel prior is proposed for the problems of yellowish tones,lack of color richness and low clarity of sand dust images acquired by outdoor imaging devices.For the problem of image color bias,firstly,the adaptive normalization method is adopted to adjust the image dark pixels to improve the Gaussian model,and weighted fusion of a color correction method based on the gray world to make the histogram distribution of each channel more concentrated,so as to remove the color bias effect;then the multi-scale retinex with color restoration(MSRCR)algorithm is used for color recovery;for the fog effect existing after processing,the atmospheric light value is re-selected and the dark pixels are compensated for brightness using the dark channel-based prior method;finally,for the problems of insufficient image saturation and low contrast,the images are transferred to HSI space and enhanced using the adaptive adjustment function and improved 2D gamma correction algorithm,respectively.In order to verify the effectiveness of the algorithm,qualitative and quantitative comparative analyses are performed with other mainstream algorithms on the basis of extensive experiments.The results show that the method can not only effectively correct the color shift and better improve the image contrast and sharpness,but also has an obvious effect on the image color richness enhancement.