改进暗通道遥感影像去雾方法及效果分析

遥感影像的预处理工作是遥感数据应用的基础。去除云雾是影像预处理工作的重要组成部分。针对遥感影像雾霾浓度分布不均匀的问题,提出一种改进的暗通道遥感影像去雾方法。以"高分一号"(GF-1)卫星为例,根据影像灰度图中的灰度值对影像雾霾浓度区域进行划分,对每个区域中暗原色值的获取方式进行改进,使用导向滤波优化大气传输率,以归一化植被指数(NDVI)为基础,设计用于评价影像去雾质量的定量指标。结果表明:所提出的方法能明显去除雾霾干扰,有效改善卫星影像数据的视觉效果,增强影像细节。该方法去雾处理后的遥感数据能应用于定量遥感,提高遥感影像的可用性及有效性。

Defogging Effect Evaluation of Remote Sensing Image Based on Improved Dark Channel

Remote sensing image preprocessing is the foundation of remote sensing data application. Removing the influence of cloud and fog on image data is an important part of image preprocessing. In this paper, an improved dark channel fog removal method is proposed based on the GF-1 satellite, in view of the uneven distribution of haze concentration in remote sensing images. Firstly, the image haze concentration region is divided according to the gray value of image, and the acquisition mode of dark original color in each region is improved. Then, the atmosphere transmission rate is optimized by guided filtering. Finally, based on the NDVI (normalized difference vegetation index), a set of image defogging quality evaluation indicators is designed and compared with the fog-free image data. The quantitative analysis on the results of fog removal is carried out. The results show that the method can obviously remove the fog interference, effectively improve the visual effect of the satellite image data and enhance image details, and the remote sensing data after processed by the improved dark channel fog removal algorithm can be applied to the quantitative remote sensing, improving the availability and effectiveness of the remote sensing data.