A global land cover training dataset from 1984 to 2020
-
Published:2023-12-07
Issue:1
Volume:10
Page:
-
ISSN:2052-4463
-
Container-title:Scientific Data
-
language:en
-
Short-container-title:Sci Data
Author:
Stanimirova RadostORCID, Tarrio KatelynORCID, Turlej Konrad, McAvoy Kristina, Stonebrook Sophia, Hu Kai-Ting, Arévalo Paulo, Bullock Eric L., Zhang Yingtong, Woodcock Curtis E., Olofsson Pontus, Zhu ZheORCID, Barber Christopher P., Souza Carlos M., Chen Shijuan, Wang Jonathan A.ORCID, Mensah FosterORCID, Calderón-Loor Marco, Hadjikakou MichalisORCID, Bryan Brett A.ORCID, Graesser Jordan, Beyene Dereje L., Mutasha Brian, Siame Sylvester, Siampale Abel, Friedl Mark A.ORCID
Abstract
AbstractState-of-the-art cloud computing platforms such as Google Earth Engine (GEE) enable regional-to-global land cover and land cover change mapping with machine learning algorithms. However, collection of high-quality training data, which is necessary for accurate land cover mapping, remains costly and labor-intensive. To address this need, we created a global database of nearly 2 million training units spanning the period from 1984 to 2020 for seven primary and nine secondary land cover classes. Our training data collection approach leveraged GEE and machine learning algorithms to ensure data quality and biogeographic representation. We sampled the spectral-temporal feature space from Landsat imagery to efficiently allocate training data across global ecoregions and incorporated publicly available and collaborator-provided datasets to our database. To reflect the underlying regional class distribution and post-disturbance landscapes, we strategically augmented the database. We used a machine learning-based cross-validation procedure to remove potentially mis-labeled training units. Our training database is relevant for a wide array of studies such as land cover change, agriculture, forestry, hydrology, urban development, among many others.
Funder
National Aeronautics and Space Administration
Publisher
Springer Science and Business Media LLC
Subject
Library and Information Sciences,Statistics, Probability and Uncertainty,Computer Science Applications,Education,Information Systems,Statistics and Probability
Reference51 articles.
1. Foody, G., Pal, M., Rocchini, D., Garzon-Lopez, C. & Bastin, L. The sensitivity of mapping methods to reference data quality: Training supervised image classifications with imperfect reference data. ISPRS Int. J. Geo-Inf. 5, 199 (2016). 2. Elmes, A. et al. Accounting for training data error in machine learning applied to earth observations. Remote Sens. 12, 1034 (2020). 3. Colditz, R. An evaluation of different training sample allocation schemes for discrete and continuous land cover classification using decision tree-based algorithms. Remote Sens. 7, 9655–9681 (2015). 4. Foody, G. M. & Mathur, A. The use of small training sets containing mixed pixels for accurate hard image classification: Training on mixed spectral responses for classification by a SVM. Remote Sens. Environ. 103, 179–189 (2006). 5. Zhou, Q., Tollerud, H., Barber, C., Smith, K. & Zelenak, D. Training data selection for annual land cover classification for the Land Change Monitoring, Assessment, and Projection (LCMAP) Initiative. Remote Sens. 12, 699 (2020).
|
|