基于ViT-D-UNet的双分支遥感云影检测网络
Bi-branch Remote Sensing Cloud and Shadow Detection Network Based on ViT-D-UNet
李远禄 1王键翔 2范小婷 2周昕 2吴明轩2
作者信息
- 1. 南京信息工程大学自动化学院,南京 210044;江苏省大气环境与装备技术协同创新中心,南京 210044
- 2. 南京信息工程大学自动化学院,南京 210044
- 折叠
摘要
云及其阴影的有效分割是遥感图像处理领域中重要的问题,它对于地表特征提取、气候检测、大气校正等有很大帮助.然而云和云影遥感图像特征复杂,云分布多样不规则,且边界信息模糊易受背景干扰等特点,导致其特征难以准确提取,也少有专门为其设计的网络.针对以上问题,本文提出一种ViT(vision Transformer)和D-UNet双路网络.本文网络分为两个分支:一路是基于卷积的局部特征提取模块,在D-UNet的膨胀卷积模块基础上,引入深度可分离卷积,提取多尺度特征的同时,减少参数;另一路通过ViT在全局上理解上下文语义,加深对整体特征提取.两支路间存在信息交互,完善提取的特征信息.最后通过独特设计的融合特征解码器,进行上采样,减少信息丢失.模型在自建的云和云影数据集以及HRC_WHU公开数据集上取得优越的性能,在MIoU指标上分别领先次优模型0.52%和0.44%,达到了92.05%和85.37%.
Abstract
Effective segmentation of clouds and their shadows is a critical issue in the field of remote sensing image processing.It plays a significant role in surface feature extraction,climate detection,atmospheric correction,and more.However,the complex features of clouds and cloud shadows in remote sensing images,characterized by their diverse,irregular distributions and fuzzy boundary information that is easily disturbed by the background,make accurate feature extraction challenging.Moreover,there are few networks specifically designed for this task.To address these issues,this study proposes a dual-path network combining vision Transformer(ViT)and D-UNet.The network is divided into two branches:one is a convolutional local feature extraction module based on the dilated convolution module of D-UNet,which introduces a multi-scale atrous spatial pyramid pooling(ASPP)to extract multi-dimensional features;the other branch comprehends the context semantics globally through the vision Transformer,enhancing feature extraction.Finally,the study performs an upsampling through a feature fusion decoder.The model achieves superior performance on both a self-built dataset of clouds and cloud shadows and the publicly available HRC_WHU dataset,leading the second-best model by 0.52%and 0.44%in the MIoUmetric,achieving 92.05%and 85.37%,respectively.
关键词
遥感/云检测/语义分割/特征融合Key words
remote sensing/cloud detection/semantic segmentation/feature fusion引用本文复制引用
出版年
2024
NETL
NSTL
万方数据