基于环境光差异的图像去雾算法

针对环境光差异引起图像去雾亮度值存在偏差、图像质量下降的问题，提出一种基于环境光差异的图像去雾算法。把图像分为环境光差异较大的天空区域和非天空区域，根据环境光差异设置不同的阈值对两类区域再进行分割；采用对比度先验去雾算法对天空区域块去雾，通过暗通道先验去雾算法对非天空区域块处理；最后用动态阈值白平衡算法调整恢复后的图像亮度和色彩。实验结果表明，本算法比其他算法能更有效解决环境光差异引起的图像去雾色彩失真、质量下降问题，鲁棒性较高，恢复后的图像色彩真实度和清晰度较高，更符合人眼视觉观察特性。

An image dehazing algorithm based on atmospheric difference

The atmosphere lighting difference in outdoor image would cause the dehazed outdoor image quality to occur the color deviation, image quality degradation and other issues. An image dehazing algorithm based on atmosphere lighting difference is proposed. The image is segmented into sky area and non-sky area. The sky area and non-sky area are further segmented according to atmosphere lighting difference. The contrast prior dehazing method is applied to the sky areas, while the dehazing approach based on dark channel is applied to dehaze the non-sky areas. The white balance adjustment algorithm is employed for slightly changing the brightness and color. Experiments demonstrate that the presented method is better than the conventional methods in color fidelity preservation, uniform illumination generation and image quality maintenance. The algorithm is robust, which makes the frog-free images bear more realistic and sharp detail, and accord to the human visual characteristics.