【摘要】 传统的图像去雾算法在处理复杂景深场景下的物体时，往往因为透射率估计的不准确而产生伪影和色彩失真现象.结合颜色衰减先验和上下文正则，并用改进的引导滤波器对透射率优化的方法，能够准确计算出各种场景下图像的透射率，最后通过对比度拉伸，从而使恢复出来的图像更具有可视性，并且在峰值信噪比(PSNR)、结构相似性(SSIM)、平均梯度(AG)评价中要优于以往的算法.更多还原

【Abstract】 Traditional image defogging algorithms often produce artifacts and color distortion due to inaccurate estimation of transmittance when dealing with objects in complex depth of field scenes. Combining color attenuation priori and context regularization, and using an improved guided filter to optimize the transmittance, the transmittance of images in various scenes can be accurately calculated. Finally, through contrast stretching, the restored image has more visibility, and is superior to the previous algorithms in peak signal-to-noise ratio(PSNR), structural similarity(SSIM) and average gradient(Ag) evaluation.更多还原