基于引导滤波的单幅图像自适应去雾算法

发布时间：2019-06-14 22:02

【摘要】：在雾、霾等天气条件下,由于空气中悬浮粒子的大气散射作用导致获得的户外图像严重降质。针对该问题,提出一种单幅图像自适应去雾算法。在分析大气散耗函数特性的基础上,利用引导滤波估计大气散耗函数,间接求得透射率,通过自适应选取阈值调整明亮区域透射率,改善复原图像中的假彩色,并根据大气散射模型恢复图像。实验结果表明,与传统引导滤波算法和取固定阈值的算法相比,该算法获得的去雾图像近景更清晰,明亮区域恢复更自然,同时提高了运算速度。
[Abstract]:Under the weather conditions of fog, haze and so on, the outdoor images obtained are seriously degraded due to the atmospheric scattering of suspended particles in the air. In order to solve this problem, an adaptive fog removal algorithm for single image is proposed. Based on the analysis of the characteristics of atmospheric dissipation function, the atmospheric dissipation function is estimated by guided filtering, and the transmittance is obtained indirectly. The transmittance of bright region is adjusted by adaptive threshold selection, the false color in the restored image is improved, and the image is restored according to the atmospheric scattering model. The experimental results show that compared with the traditional guided filtering algorithm and the fixed threshold algorithm, the close range of the fog removal image obtained by this algorithm is clearer, the bright region recovery is more natural, and the operation speed is improved at the same time.
【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(61561030) 甘肃省科技厅自然科学基金资助项目(1310RJZA050) 甘肃省财政厅基本科研业务费基金资助项目(214138) 兰州交通大学研究生校内教改立项基金资助项目(160012)
【分类号】：TP391.41

【相似文献】