基于局部线性和系数分片常数的全变差图像去雾算法OA北大核心

In order to effectively retain the details of hazy images,like edge structures,and avoid color distortion during dehazing,this paper introduces the total variation of local window coefficients and the total variation of the original image,and proposes a total variation image dehazing model based on local linearity and piecewise constant coefficients.Further,we design a fast numerical algorithm for the model by combining the alternating direction multipliers method.Finally,dehazing experiments conducted on synthetic images and real-world scene images demonstrate that,compared with the current mainstream dehazing algorithms,the proposed algorithm exhibits superior performance in both numerical indicators and visual effects.