Feature Importance Analysis for Local Climate Zone Classification Using a Residual Convolutional Neural Network with Multi-Source Datasets

Abstract

Global Local Climate Zone (LCZ) maps, indicating urban structures and land use, are crucial for Urban Heat Island (UHI) studies and also as starting points to better understand the spatio-temporal dynamics of cities worldwide. However, reliable LCZ maps are not available on a global scale, hindering scientific progress across a range of disciplines that study the functionality of sustainable cities. As a first step towards large-scale LCZ mapping, this paper tries to provide guidance about data/feature choice. To this end, we evaluate the spectral reflectance and spectral indices of the globally available Sentinel-2 and Landsat-8 imagery, as well as the Global Urban Footprint (GUF) dataset, the OpenStreetMap layers buildings and land use and the Visible Infrared Imager Radiometer Suite (VIIRS)-based Nighttime Light (NTL) data, regarding their relevance for discriminating different Local Climate Zones (LCZs). Using a Residual convolutional neural Network (ResNet), a systematic analysis of feature importance is performed with a manually-labeled dataset containing nine cities located in Europe. Based on the investigation of the data and feature choice, we propose a framework to fully exploit the available datasets. The results show that GUF, OSM and NTL can contribute to the classification accuracy of some LCZs with relatively few samples, and it is suggested that Landsat-8 and Sentinel-2 spectral reflectances should be jointly used, for example in a majority voting manner, as proven by the improvement from the proposed framework, for large-scale LCZ mapping.

Using OpenStreetMap (OSM) to enhance the classification of local climate zones in the framework of WUDAPT

“…transfer-learning, and new input data. The methodological development has focused on new and more sophisticated classifiers and post-processing techniques including context classifiers (Verdonck et al, 2017), residual convolutional neural networks (Qiu et al, 2018), and Markov random fields (Tuia et al, 2017b). Moreover, alternative approaches from supervised classification have been tested, in particular GIS-based methods (Gal et al, 2015;Geletič and Lehnert, 2016;Hidalgo et al, accepted;Lelovics et al, 2014;Unger et al, 2014;Zheng et al, 2018).…”

“…A particular focus was set on transferring training labels between cities (Demuzere et al, 2019;Kaloustian et al, 2017; and developing robust and transferable classifiers, which was also the challenge of the 2017 Data Fusion Contest (Tuia et al, 2017a;Yokoya et al, 2018), organized by the Image Analysis and Data Fusion Technical Committee of the IEEE Geoscience and Remote Sensing Society. Further work has been done on the upscaling and generation of large area data sets (Demuzere et al, in review;Qiu et al, 2018). On the data side different sensors have been tested in the classic approach, including Sentinel 2 (Demuzere et al, 2019;Kaloustian et al, 2017;Qiu et al, 2018), ASTER (Yong Xu et al, 2017), SAR (Bechtel et al, 2016;Demuzere et al, 2019;Kaloustian et al, 2017) and nighttime lights (Demuzere et al, 2019;Qiu et al, 2018), and various remote sensing derived parameters were used in Mitraka et al (2015).…”

“…Further work has been done on the upscaling and generation of large area data sets (Demuzere et al, in review;Qiu et al, 2018). On the data side different sensors have been tested in the classic approach, including Sentinel 2 (Demuzere et al, 2019;Kaloustian et al, 2017;Qiu et al, 2018), ASTER (Yong Xu et al, 2017), SAR (Bechtel et al, 2016;Demuzere et al, 2019;Kaloustian et al, 2017) and nighttime lights (Demuzere et al, 2019;Qiu et al, 2018), and various remote sensing derived parameters were used in Mitraka et al (2015).…”

Fusion of Heterogeneous Earth Observation Data for the Classification of Local Climate Zones

“…Furthermore, multi-source data fusion offers much potential for urban mapping. Due to the rich characteristics of natural processes and environments, it is rare for a single acquisition method to provide a complete understanding of certain phenomenon [13]- [15]. Multi-source data fusion considers the task from various points of view and then provides opportunities to view the whole picture.…”

Local Climate Zone (LCZ) Map Accuracy Assessments Should Account for Land Cover Physical Characteristics that Affect the Local Thermal Environment

“…The WUDAPT protocol has been applied successfully in many LCZ mapping studies in different geographic regions [4][5][6]9,10]. To build on the WUDAPT approach, some studies have proposed the use of additional remote sensing datasets, e.g., nighttime optical imagery [11], nighttime thermal infrared imagery [12], synthetic aperture radar data [9], or LiDAR data [13]. Other studies have proposed incorporating ancillary GIS datasets in the image classification workflow, e.g., OpenStreetMap data [11,14], and/or the use of alternative classification algorithms (e.g., convolutional neural networks [11]).…”