基于光补偿和逐像素透射率的图像复原算法
发布时间：2018-07-26 08:55
【摘要】：针对雾霾和沙尘天气下图像降质,提出一种光补偿色彩复原和逐像素透射率估计的单幅图像可见度复原算法。沙尘颗粒对蓝光吸收导致图像偏黄,利用光补偿方法可以消除偏色现象并转换为雾图。以最小颜色通道与其高斯函数的比值关系估计透射率,用中值滤波消除透射率纹理效应,恢复的透射率深度交替明显,边缘保持良好,不需要耗时的后处理操作,最后根据大气散射模型恢复图像。实验结果表明,沙尘天气处理效果较好,雾天图像恢复的图像饱和度适宜,明亮区域恢复得更为自然,运行时间较快。
[Abstract]:A single image visibility restoration algorithm based on photo-compensated color restoration and Pixel-by-pixel transmittance estimation is proposed for image degradation in haze and dust weather. The absorption of blue light by dust particles results in yellowing of the image, which can be eliminated by using light compensation method and converted into fog map. Based on the ratio of the minimum color channel to its Gao Si function, the transmissivity is estimated by using median filter to eliminate the transmissivity texture effect. The restored transmittance depth alternates obviously, the edges are kept well, and the time consuming post-processing operation is not required. Finally, the image is restored according to the atmospheric scattering model. The experimental results show that the effect of sand and dust weather processing is better, the image saturation of fog image restoration is suitable, the bright area is recovered more naturally, and the running time is faster.
【作者单位】： 兰州交通大学电子与信息工程学院;
【基金】：国家自然科学基金资助项目(No.61561030) 甘肃省财政厅基本科研业务费基金资助项目(No.214138) 兰州交通大学教改基金资助项目(No.160012)~~
【分类号】：TP391.41
本文编号：2145497
[Abstract]:A single image visibility restoration algorithm based on photo-compensated color restoration and Pixel-by-pixel transmittance estimation is proposed for image degradation in haze and dust weather. The absorption of blue light by dust particles results in yellowing of the image, which can be eliminated by using light compensation method and converted into fog map. Based on the ratio of the minimum color channel to its Gao Si function, the transmissivity is estimated by using median filter to eliminate the transmissivity texture effect. The restored transmittance depth alternates obviously, the edges are kept well, and the time consuming post-processing operation is not required. Finally, the image is restored according to the atmospheric scattering model. The experimental results show that the effect of sand and dust weather processing is better, the image saturation of fog image restoration is suitable, the bright area is recovered more naturally, and the running time is faster.
【作者单位】： 兰州交通大学电子与信息工程学院;
【基金】：国家自然科学基金资助项目(No.61561030) 甘肃省财政厅基本科研业务费基金资助项目(No.214138) 兰州交通大学教改基金资助项目(No.160012)~~
【分类号】：TP391.41
【相似文献】
相关期刊论文 前2条
1 唐红梅;申瑾;周亚同;韩力英;王霞;;雾霾天气条件下车牌信息的识别[J];电视技术;2014年05期
2 ;[J];;年期
相关硕士学位论文 前2条
1 曾宇骏;雾天降质图像增强处理技术研究[D];国防科学技术大学;2013年
2 孙张明;数字图像去雾方法及其应用研究[D];天津大学;2014年
，本文编号：2145497
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2145497.html
