- Research Article
- 10.1088/1755-1315/1586/1/012049
Deep learning-based analysis of urban expansion in South Jakarta: A ConvLSTM approach with multitemporal satellite data
- Feb 01, 2026
- IOP Conference Series: Earth and Environmental Science
- Dimas Nugroho + 1 more +1
Abstract The South Jakarta City area is one of the five administrative cities of the special capital region of Jakarta. This area has a history of development for hundreds of years until it reached the first major planned development during the Dutch East Indies era. This area then continued to develop as a residential area in recent decades. This study aims to analyze the dynamics of land cover changes in the South Jakarta area from 2000 to 2024 with a five-year interval. Using Landsat satellite imagery that has been processed and classified based on a combination of NDVI and NDBI indices, labelling is carried out on three main categories of land cover: vegetation, built-up areas, and water bodies. Furthermore, a deep learning model based on Convolutional Long Short-Term Memory (ConvLSTM) is applied to predict land cover classification spatially and temporally. Multitemporal satellite imagery with six time steps (2000, 2005, 2010, 2015, 2020, and 2024) is used as input into the ConvLSTM model, labelling the results of index extraction as the target. The training process is carried out by patching 32×32 pixel images, producing a model that can effectively recognize land change patterns in urban areas. Model evaluation using a confusion matrix and a classification report shows good classification performance in built-up areas. However, vegetation and water body classes still show accuracy challenges due to unbalanced data distribution. The prediction results are visualized as classification maps and inter-annual change maps, which are then masked with the administrative boundaries of South Jakarta. This study contributes to using spatial-temporal deep learning methods for long-term urbanization analysis. It supports formulating policies for controlling development and preserving the urban environment.
Read more