基于全局及局部优势特征融合的遥感图像去雾

Global and local advantageous feature fusion for remote sensing image dehazing

摘要:

由于大气中颗粒物质的散射和吸收，遥感图像通常存在细节模糊和对比度降低等问题，严重影响其视觉质量。针对这些问题，本文提出了一种基于全局及局部优势特征融合的遥感图像去雾方法。具体而言，首先利用暗通道先验对原始图像进行去雾预处理。随后，采用多曝光融合策略以及积分和平方积分方法整合图像区域的优势特征信息，提升全局及局部对比度。最后，通过金字塔融合自适应选择全局及局部对比度增强的显著特征以获得清晰化图像。实验结果表明，该方法在遥感图像去雾领域优于其他方法，处理后的图像在黑暗区域曝光、全局对比度增强和局部细节提升等方面表现出了良好性能。

Abstract:

Due to the scattering and absorption of particulate matter in the atmosphere, remote sensing images often suffer from problems such as blurred details and reduced contrast, which seriously affect their visual quality. To address these issues, this paper proposes a remote sensing image dehazing method based on the fusion of global and local advantages features. Specifically, we employ the dark channel prior to haze removal preprocessing on the raw image. Subsequently, we utilize a multi-exposure fusion strategy and integration methods such as integral and square integral to combine dominant feature information from image regions, enhancing global and local contrast. Finally, a pyramid fusion approach is employed to adaptively select salient features to enhance global and local contrast, resulting in a clarified image. The experimental results show that this method outperforms other methods in remote sensing image dehazing, and the processed image exhibits good performance in terms of dark region exposure, global contrast enhancement, and local detail enhancement.