SMT201: Geographic Information Systems for Urban Planning
In this project, we are given two major themes and sub-themes. The topic our group has chosen is Theme 1 Sub-theme 2: Land cover change detection using high resolution multi-spectral satellite data.
Singapore’s urbanization has progressed rapidly due to rising demand for public housing. This increasing demand for housing estates, especially for Build-To-Order (BTO) flats puts greater pressure on HDB to ramp up the supply of flats to meet the increasing housing needs.
With more land reserved for residential purposes, biodiversity loss and pressures on natural ecosystems are becoming more pertinent, especially with increasing land cover change due to changing land uses. Punggol being Singapore’s first eco-town, we are interested in what is the extent to which recent urban developments in Punggol have changed its land cover, starting from the opening of Waterway Point in 2016, to implementations of smart initiatives for Northshore District in 2020.
In this project, we analysed the land cover changes and impact on Punggol from 2016 to 2020 through the use of remote sensing data, satellite imagery and image processing techniques.
Using Sentinel-2 satellite imagery data, we created True and False Color Composites of the map. By defining seven land cover classes; namely Water body, Inland water, Bare land, Natural vegetation, Managed vegetation, Built-up and Impervious surfaces, we digitised the training and testing samples to perform Supervised Classification. Lastly, we performed the land cover change detection using the Semi-automatic classification Plugin (SCP) to identify changes in vegetation.
From 2016 to 2020, Managed vegetation had the highest decrease in land cover, while Water body, Bare land and Built-up increased significantly. The results of the land cover classification showed a high overall accuracy of 86% and Kappa Statistic of 0.88 for 2016, and an overall accuracy of 72% and Kappa Statistic of 0.67 for 2020.
Based on the Land Cover Change analysis of Punggol, we discovered several land cover changes from 2016 to 2020. The proportion of water increased by 2.43% in proportion to the total area for Water. Vegetation, on the other hand, experienced moderate reduction of 13.92% between 2016 and 2020, where majority of vegetation was converted to bare land which were primarily used for the future urban developments. Lastly, land terrain such as bare land and built-up experienced a significant increase of 45.71% and 43.28% respectively in proportion to their respective total land area in 2016.