改进HarDNet-MSEG的遥感影像水体信息提取方法  

Water Body Extraction from Remote Sensing Images Based on Improved HarDNet-MSEG

作　　者：郭慧琳 谢元礼[1,2] 胡李发 雍佳乐 李云梅 孙韶启 GUO Huilin;XIE Yuanli;HU Lifa;YONG Jiale;LI Yunmei;SUN Shaoqi(College of Urban and Environmental Science,Northwest University,Xi'an 710127,China;Shaanxi Remote Sensing and GIS Engineering Research Center,Xi'an 710127,China)

机构地区：[1]西北大学城市与环境学院,西安710127 [2]陕西省遥感与地理信息工程研究中心,西安710127

出　　处：《地球信息科学学报》2024年第7期1745-1762,共18页Journal of Geo-information Science

摘　　要：准确有效地提取水体信息,对于水资源监测、管理和应用等方面具有重要意义。由于水体形状、大小和分布的多样性以及场景的复杂性,如何高效准确地从遥感影像中提取出水体仍具有挑战性。传统的方法虽然可以从遥感图像中提取水体,但由于异物同谱其提取精度往往难以满足实际应用要求。因此,迫切需要先进的高性能技术来提高水体提取的效率和准确性。将深度学习与遥感技术相结合可以充分发挥深度学习的优势,有效帮助准确地提取水体信息。目前以深度学习的方法提取水体信息面临的挑战仍是多尺度特征融合、耗时长和参数多。HarDNet-MSEG(Harmonic DenseNet-MSEG)模型拥有较高的分割精度和较快的推理速度,为进一步充分利用来自通道和空间位置层面的相关信息以及提高模型的分割精度,本文以HarDNet-MSEG为网络框架,设计了一种名为HAM(Hybrid Attention Mechanism)的注意力机制,将其嵌入到HarDNet-MSEG网络中以探究其在网络中的最佳位置,在相同的实验环境下与其他注意力机制、经典网络算法以及传统的方法进行一系列的对比实验,并测试该模型在其他数据集上的通用性。结果表明,HAM模块在HarDNet-MSEG网络的较浅层处表现最出色。与其他注意力机制相比,HAM模块取得了更高的性能,MIoU、FWIoU和PA分别达到了94.0687%、97.7374%和99.3205%。与DeepLabV3+、U-Net和PSPNet等经典模型相比,HarDNet-MSEG-HAM1模型不仅有最好的MIoU,参数量、计算量和训练时间各方面都表现出卓越的性能。与传统方法相比,HarDNet-MSEG-HAM1模型具有显著的优势,同时该模型在其他数据集上也表现出了良好的性能。最后,成功提取了青藏高原内流流域的2013、2016、2019和2022年4期湖泊,并对其面积变化进行了分析。一系列的实验数据表明,该模型在水体提取任务中的优越性与鲁棒性。本论文预期可以为从复杂场景的�Accurately and efficiently mapping water bodies is of great significance for water resources monitoring,management,and application.It remains challenging to efficiently and accurately extract water bodies from remote sensing images due to the diversity of their shapes,sizes,distributions,and the complexity of the scenes.Although traditional methods can extract water bodies from remote sensing images,their extraction accuracy falls short of meeting the practical application requirements due to heterogeneous objects with the same spectrum.Therefore,there is an urgent demand for advanced high-performance techniques to improve the efficiency and accuracy of water body extraction.The combination of deep learning and remote sensing technology can fully exploit the advantages of deep learning and effectively contribute to the accurate extraction of water bodies.Current challenges in extracting water bodies through deep learning methods persist in issues such as multi-scale feature fusion,prolonged processing times,and the involvement of multiple parameters.The HarDNet-MSEG model possesses high segmentation accuracy and fast inference speed.In order to further fully utilize relevant information from the channel and spatial location levels,and to improve the segmentation accuracy of the model,this paper proposes a Hybrid Attention Mechanism(HAM)integrated into the HarDNet-MSEG network framework.The HAM is embedded into the HarDNet-MSEG network to explore its optimal position within the architecture,and a series of comparative experiments are conducted under the same experimental conditions,comparing it with other attention mechanisms,classical network algorithms,and traditional methods.Additionally,the generalizability of the model on other datasets is tested.The results show that the HAM module performs best at the shallower layers of the HarDNet-MSEG network.Compared with other attention mechanisms,the HAM module achieves higher performance,with MIoU,FWIoU,and PA reaching 94.0687%,97.7374%,and 99.3205%,respectively.Comp

关 键 词：水体提取 深度学习 HarDNet-MSEG 混合注意力机制 语义分割 多尺度特征 青藏高原 遥感影像 

分 类 号：P237[天文地球—摄影测量与遥感]