Multitask Learning for Human Settlement Extent Regression and Local Climate Zone Classification

作     者：Qiu, Chunping Liebel, Lukas Hughes, Lloyd Haydn Schmitt, Michael Korner, Marco Zhu, Xiao Xiang 

作者机构：Tech Univ Munich TUM Data Sci Earth Observat SiPEO D-80333 Munich Germany Tech Univ Munich TUM Remote Sensing Technol LMF D-80333 Munich Germany German Aerosp Ctr DLR Remote Sensing Technol Inst IMF D-82234 Wessling Germany 

出 版 物：《IEEE GEOSCIENCE AND REMOTE SENSING LETTERS》 (IEEE Geosci. Remote Sens. Lett.)

年 卷 期：2022年第19卷

核心收录：

学科分类：0808[工学-电气工程] 1002[医学-临床医学] 08[工学] 0708[理学-地球物理学] 0816[工学-测绘科学与技术] 

基　　金：China Scholarship Concil [ERC-2016-StG-714087] European Research Council (ERC) under the European Union [ERC-2016-StG-714087] Helmholtz Association Helmholtz Excellent Professorship "Data Science in Earth Observation -Big Data Fusion for Urban Research" German Federal Ministry of Education and Research (BMBF) Federal Ministry of Transport and Digital Infrastructure (BMVI) [16AVF2019A] 

主　　题：Task analysis Training Meteorology Earth Remote sensing Europe Urban areas Convolutional neural networks (CNNs) human settlement extent (HSE) land cover (LC) local climate zones (LCZs) multitask learning (MTL) sentinel-2 

摘      要：Human settlement extent (HSE) and local climate zone (LCZ) maps are both essential sources, e.g., for sustainable urban development and Urban Heat Island (UHI) studies. Remote sensing (RS)- and deep learning (DL)-based classification approaches play a significant role by providing the potential for global mapping. However, most of the efforts only focus on one of the two schemes, usually on a specific scale. This leads to unnecessary redundancies since the learned features could be leveraged for both of these related tasks. In this letter, the concept of multitask learning (MTL) is introduced to HSE regression and LCZ classification for the first time. We propose an MTL framework and develop an end-to-end convolutional neural network (CNN), which consists of a backbone network for shared feature learning, attention modules for task-specific feature learning, and a weighting strategy for balancing the two tasks. We additionally propose to exploit HSE predictions as a prior for LCZ classification to enhance the accuracy. The MTL approach was extensively tested with Sentinel-2 data of 13 cities across the world. The results demonstrate that the framework is able to provide a competitive solution for both tasks.