[IDN] Ground Cover Change Dataset

Published: 21 September 2022| Version 3 | DOI: 10.17632/mzp3k6fmtz.3
Martin Garcia Fry


The Global Ground Cover Change (GGCC) dataset is a reference-based raster collection of bi-annual land change maps. Data here was produced with a random forest model used to classify strata-based ground cover change trends in a disaster-affected area of Java, Indonesia. The associated research comprehends (1) data collection: ICESat-2, ATLAS08 LiDAR datasets, Land Cover datasets, and ancillary DEM-derived subsets of data for quality-based sample refinements used for model training; (2) data processing, using local temporal and static metrics to parse localized filtration ramifications in a filtering code; (3) model optimization using machine learning techniques with the sci-kit learn package; (4) preliminary classifications and accuracy assessments; (5) informed post-processing; and, (5) recovery trend evaluations on bi-annual ground change maps to inform livelihood recovery and monitor the underlying effects of relocation and resettlement in peri-urban environments. These are valuable assets to understanding how we should rebuild affected areas for humanitarian purposes where communities require restored environments that efficiently contribute to diminishing the overall human footprint on Earth.


Steps to reproduce

Steps to Collect and reproduce the data sets are available within a .pdf in: Model > Filtering Code > PDF. All steps to classify and reproduce preliminary classifications are available through the Google Colab link below. The code to extract the 2019 image composites corresponding to Java east and west is available through the second link below. The associated article with detailed steps to reproduce the entire method will be available later.


Tohoku Daigaku - Aobayama Shin Campus


Machine Learning, Land Cover Change, Land Use Change