基于半逆法和暗原色先验的单一图像去雾新方法

Single image dehaze based on semi-inverse approach and dark channel prior

针对现有图像去雾方法的缺陷,依据雾天成像的理论模型,提出了一种基于半逆法和暗原色先验相结合的单一图像去雾新方法。对一幅待去雾图像,首先通过半逆法检测并标记出图像浓雾区域,结合暗原色先验,随机多次选取浓雾区域中具有较大暗原色值的候选像素点集,并通过计算原图中这些候选点集的像素点平均亮度值估算大气光常数。其次通过暗原色先验信息获取图像像素点的透射率分布。最后在得到大气光常数和透射率参数的基础上,根据雾天成像的光学理论模型反解出去雾后图像。通过图像去雾的无参考客观质量评价(NQA)体系对有效细节强度、色调还原程度和结构信息三项指标进行定量评估,评估结果表明,相比目前去雾效果最好的He算法,这一新方法平均能提高了15%。此外,该方法能减少去雾后图像的halo效应,更能满足人们的主观观测需求。

Aiming at existing image dehaze methods＇ problems in dehaze,a novel method for single image dehaze based on the semi-inverse approach and the dark channel prior was studied according to the theoretical model for image formation in hazy weather.For an image to be processed,firstly,the semi-inverse approach was used to detect and label its foggy region,and randomly select the candidate pixel set with larger values of dark channel prior in the detected foggy region for several times,thus,the Airlight value can be estimated by averaging the brightness values of candidate pixels in origin image.Then,the prior information was used to obtain the transmission distribution of image pixels.Finally,with the Airlight constant and the transmission distribution,the image to be processed can be restored according to the given optical model.The experimental results demonstrated that compared to the best algorithm so far proposed by He,the new method＇ s performance averaging over effective strength,tone reproduction extent and structure information,which are the three indicators of the no-reference quality assessment （NQA） system for image quantitative evaluation,was increased by more than 15％.Moeover,since it can effectively reduce the halo effect,the restored images are natural to human needs.