当前位置: X-MOL 学术Remote Sens. Environ. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
An automated, generalized, deep-learning-based method for delineating the calving fronts of Greenland glaciers from multi-sensor remote sensing imagery
Remote Sensing of Environment ( IF 13.5 ) Pub Date : 2020-12-24 , DOI: 10.1016/j.rse.2020.112265
Enze Zhang , Lin Liu , Lingcao Huang , Ka Shing Ng

In the past two decades, the data volume of remote sensing imagery in the polar regions has increased dramatically. The calving fronts of many Greenland glaciers have been undergoing substantial variations, and a comprehensive front dataset is necessary for better understanding such frontal dynamics. Therefore, there is a need for an automated approach to identifying glaciological features such as calving fronts. In 2019, three deep-learning-based methods were applied to calving front delineation, but were restricted to a specific area or dataset. Here, we develop a more generalized method that can be applied to a major outlet glacier or remote sensing datasets that are not included in the training. We integrate seven remote sensing datasets into a single deep learning network. The core datasets include optical (Landsat-8 and Sentinel-2) and synthetic aperture radar images (Envisat, ALOS-1 TerraSAR-X, Sentinel-1, and ALOS-2) taken over Jakobshavn Isbræ, Kangerlussuaq, and Helheim, spanning from 2002 to 2019. We evaluate four neural network architectures (e.g., U-Net, DeepLabv3+ with ResNet, DRN, and MobileNet as the backbones) and three histogram modification strategies (e.g., histogram normalization, linear stretching, and no histogram modification). We find that the combination of histogram normalization and DRN-DeepLabv3+ has the lowest test error, at 86 m. These promising results show that our method has a high generalization ability on various glaciers and data types.



中文翻译:

一种基于深度学习的自动,通用方法,用于从多传感器遥感影像中描绘格陵兰冰川的冰崩锋线

在过去的二十年中,极地地区遥感影像的数据量急剧增加。许多格陵兰冰川的产犊锋线都在经历很大的变化,因此有一个全面的锋线数据集对于更好地了解这种锋线动力学是必要的。因此,需要一种自动方法来识别冰川特征,例如产犊锋。在2019年,将三种基于深度学习的方法应用于划分前线轮廓,但仅限于特定区域或数据集。在这里,我们开发了一种更通用的方法,可将其应用于训练中未包括的主要出口冰川或遥感数据集。我们将七个遥感数据集集成到一个深度学习网络中。核心数据集包括从JakobshavnIsbræ,Kangerlussuaq和Helheim拍摄的光学(Landsat-8和Sentinel-2)和合成孔径雷达图像(Envisat,ALOS-1 TerraSAR-X,Sentinel-1和ALOS-2),范围从2002年至2019年。我们评估了四种神经网络架构(例如,以ResNet,DRN和MobileNet为骨干的U-Net,DeepLabv3 +)和三种直方图修改策略(例如,直方图归一化,线性拉伸和无直方图修改)。我们发现直方图归一化和DRN-DeepLabv3 +的组合具有最低的测试误差,为86 m。这些有希望的结果表明,我们的方法对各种冰川和数据类型具有很高的概括能力。从2002年到2019年。我们评估了四种神经网络体系结构(例如,U-Net,以ResNet,DRN和MobileNet为骨干的DeepLabv3 +)和三种直方图修改策略(例如,直方图归一化,线性拉伸和无直方图修改) 。我们发现直方图归一化和DRN-DeepLabv3 +的组合具有最低的测试误差,为86 m。这些有希望的结果表明,我们的方法对各种冰川和数据类型具有很高的概括能力。从2002年到2019年。我们评估了四种神经网络架构(例如,U-Net,以ResNet,DRN和MobileNet为骨干的DeepLabv3 +)和三种直方图修改策略(例如,直方图归一化,线性拉伸和无直方图修改) 。我们发现直方图归一化和DRN-DeepLabv3 +的组合具有最低的测试误差,为86 m。这些有希望的结果表明,我们的方法对各种冰川和数据类型具有很高的概括能力。我们发现直方图归一化和DRN-DeepLabv3 +的组合具有最低的测试误差,为86 m。这些有希望的结果表明,我们的方法对各种冰川和数据类型具有很高的概括能力。我们发现直方图归一化和DRN-DeepLabv3 +的组合具有最低的测试误差,为86 m。这些有希望的结果表明,我们的方法对各种冰川和数据类型具有很高的概括能力。

更新日期:2020-12-25
down
wechat
bug