基于直方图均衡化与MSRCR的沙尘降质图像增强算法

针对沙尘天气下图像色彩偏移严重及对比度低等问题，提出一种基于直方图均衡化与带色彩恢复的多尺度视网膜（MSRCR）增强的沙尘降质图像增强算法。通过偏色校正和图像增强两个步骤进行图像恢复，将RGB图像各通道预处理后利用限制对比度自适应直方图均衡方法得到校正后的图像，对图像采用双边滤波进行降噪处理，通过MSRCR算法进一步解决色彩失衡问题。由于处理后的图像对比度较低，存在一定色偏，利用伽马校正和基于图像分析的偏色检测及颜色校正方法进行处理得到最终结果。对大量沙尘降质图像进行仿真实验，结果表明，该算法能够有效处理不同偏色程度的沙尘图像，不仅提高了图像的对比度，而且有效避免了图像颜色偏移现象，相比GCANet、MSRCR等算法，平均时间效率提升了46.2%~94.7%。

Sand-Dust Degraded Image Enhancement Algorithm Based on Histogram Equalization and MSRCR

To solve the problems of severe image color offset and low contrast in dusty conditions, a sand-dust degraded image enhancement algorithm based on histogram equalization and Multi-Scale Retinex with Color Restoration (MSRCR) is proposed, which primarily restores an image via two procedures:color deviation correction and image enhancement.First, after preprocessing each channel of an RGB image, the corrected image is processed using the Contrast Limited Adaptive Histogram Equalization (CLAHE) method.Subsequently, the image is denoised via bilateral filtering.Next, the MSRCR algorithm is used to solve color imbalance. Finally, because the contrast of the processed image is low and color deviation remains, the final result is obtained via further processing through gamma correction and color deviation detection, as well as via color correction methods based on image analysis.Results obtained from simulating numerous sand and dust degraded images show that the algorithm can effectively process sand and dust images of different color deviation degrees, improve the image contrast, and effectively avoid image color offsets.In addition, the average time efficiency of this algorithm is 46.2%~94.7% higher than those of the GCANet and MSRCR algorithms.