夜间有雾图像去雾算法的研究

Research on defogging algorithm for foggy images at night

夜间有雾图像对比度较低、噪声明显且细节不突出,极大的影响了后期的数据处理与信息提取。而大多数针对于白天图像的去雾算法,并不适用于夜间有雾图像,因此在大气散射模型的基础上,通过分析当前主流相机频段同常见散射介质之间的作用机理,构建了夜雾图像成像模型,再结合对原始导向滤波透射率的改进,进一步突出了图像的细节。最后通过使用CLAHE算法对图像进行增强从而得到了复原后的图像。通过对大量实测的夜雾图像进行处理,并与现有的除雾算法进行比较,验证了提出的夜雾条件下图像去雾算法的有效性与优异性。

Foggy images at night have low contrast, obvious noise, and unobtrusive details, which greatly affects later data processing and information extraction. Most defogging algorithms for daytime images are not suitable for foggy images at night, so based on the atmospheric scattering model, by analyzing the mechanism of action between the current mainstream camera frequency band and common scattering media, an night fog imaging model is constructed, and combined with the improvement in the original guided filter transmittance, the details of the image are further highlighted. Finally, the CLAHE algorithm is used to enhance the image to obtain the restored image. By processing a large number of measured night fog images and comparing with the existing algorithms, the effectiveness and superiority in the proposed image dehazing algorithm under night fog conditions are verified.