Land use/land cover mapping using deep neural network and sentinel image dataset based on google earth engine in a heavily urbanized area, China

Abstract

Accurate and detailed identification of land cover types is beneficial to the ecological environment and sustainable urban development, yet the extraction of urban land use/land cover (LULC) information with high accuracy is challenged by the high degree of landscape fragmentation. Therefore, this study is based on the Google Earth Engine (GEE) cloud platform and uses the U-Net model combined with spectral image data and SAR data to conduct a LULC classification study of highly heterogeneous urban areas in central China. The results indicated that the overall accuracy of classification result by using the U-Net model and the optimal combination of image features was 95.58%, which was 1.37%, 4.84%, and 7.85% higher than that of random forest (RF), support vector machine (SVM) and k- Nearest Neighbor (kNN), respectively. It showed that the U-Net model can effectively extract LULC information and obtain better classification results in urban areas than the machine learning algorithms. The results of this study could provide technical support to improve the accuracy of information extraction in urban areas with fragmented features.

Keywords:

• Urban agglomeration
• Land cover/land use
• Google Earth Engine
• Spatio-temporal fusion
• U-Net model

Data availability statement

Sentinel-1, Sentinel-2, MOD13Q1, ESA WorldCover 10 m v100 and CGLS-LC100 data are accessible from Google Earth Engine (https://earthengine.google.com/). Chinese Ministry of Resources (Globeland30) is accessible from the website http://globallandcover.com.

Acknowledgements

This study was funded by the National Natural Science Foundation of China (41901385), the China Postdoctoral Science Foundation (2019M652815; 2020T130731) and in part by the Key Laboratory of Natural Resources Monitoring and Supervision in Southern Hilly Region, Ministry of Natural Resources (NRMSSHR-2022-Y06). The source code for U-Net can be downloaded from https://github.com/Shu-Dan/LULC-Classification.git.

Declaration of competing interest

The authors declared that they have no conflicts of interest to this work. We declare that we do not have any commercial or associative interest that represents a conflict of interest in connection with the work submitted.

Additional information

Funding

This study was funded by the National Natural Science Foundation of China (41901385), the China Postdoctoral Science Foundation (2019M652815; 2020T130731). The source code for U-Net can be downloaded from https://github.com/Shu-Dan/LULC-Classification.git.