三种雾霾图像去雾技术分析

Analysis of Image Defogging Technology for Unmanned Vehicle

摘要: 在现实生活中，我国许多城市在秋冬季节都饱受雾霾天气的困扰。雾霾天气的频繁出现，对车辆安全驾驶、道路交通管控都是一个严峻的问题。由于大气中粒子的散射作用，机器拍摄的照片质量严重下降，这不仅影响了照片的视觉效果，也对图像中关键信息的提取产生了不利影响。针对上述问题，通过采取直方图均衡化、Retinex图像增强算法、暗原色先验去雾算法等三种去雾技术，对在雾霾天实际采集的交通路况图片进行处理，并通过比较处理前后图片中车牌信息和图片灰度直方图对结果进行评估，论述不同去雾技术的有效性和实时性。

Abstract: In real life, many cities in China suffer from haze in autumn and winter. The frequent occurrence of haze weather is a serious problem for vehicle safety driving and road traffic control. Due to the scattering effect of particles in the atmosphere, the quality of the photos taken by the machine is seriously reduced, which not only affects the visual effect of the photos, but also adversely affects the extraction of key information in the images. In view of the above problems, students adopt three common defogging technologies (histogram equalization, Retinex image enhancement algorithm and dark channel prior defogging algorithm) to process the actual traffic condition pictures collected in haze days. The results were evaluated by comparing the license plate information and gray histogram of the images before and after processing.