The paper proposes a fast and effective defogging method for single image. Based on the fact of statistical similarity between the inverted image of foggy image and the low light image, the method utilizes the Retinex algo-rithm to enhance the inverted image, and then performs the dynamic range adjustment on the result image. After that, a color recovery is made to get the de-hazed image. Experimental results of our method and other mainstream defog algo-rithm show that both the subjective and objective evaluations of the results of our algorithm are comparative to the state of art algorithm. However, the time complexity of our method is much lower than others, which makes our algorithm best suited for real-time defogging tasks.%基于雾图求反图与低光照图像之间统计特性的相似性，本文提出了一种利用Retinex图像增强方法实现图像去雾的新方法，能够实现对雾图的快速、有效去雾。该方法首先获取雾图的求反图，对求反图进行Retinex增强，并对其结果图像进行动态范围调整和色彩复原，得到去雾后的图像。对本文提出的去雾算法与当前其他去雾算法的实验结果对比显示，本文提出的去雾算法去雾效果明显，主客观评价结果与当前最好的算法相当，但本文算法的时间效率很高，能够达到实时处理的要求。
展开▼
