BD Open LULC Map: High-resolution land use land cover mapping & benchmarking for urban development in Dhaka, Bangladesh

Abstract

Land Use Land Cover (LULC) mapping using deep learning significantly enhances the reliability of LULC classification, aiding in understanding geography, socioeconomic conditions, poverty levels, and urban sprawl. However, the scarcity of annotated satellite data, especially in South/East Asian developing countries, poses a major challenge due to limited funding, diverse infrastructures, and dense populations. In this work, we introduce the BD Open LULC Map (BOLM), providing pixel-wise LULC annotations across eleven classes (e.g., Farmland, Water, Forest, Urban Structure, Rural Built- Up) for Dhaka metropolitan city and its surroundings using high-resolution Bing satellite imagery (2.22 m/pixel). BOLM spans 4,392 km2 (891 million pixels), with ground truth validated through a three-stage process involving GIS experts. We benchmark LULC segmentation using DeepLab V3+ across five major classes and compare performance on Bing and Sentinel-2A imagery. BOLM aims to support reliable deep models and domain adaptation tasks, addressing critical LULC dataset gaps in South/East Asia.

Cite

@INPROCEEDINGS{11084396,
  author={Hossain, Mir Sazzat and Paul, Ovi and Raihan Iftee, Md Akil and Hasan Rajib, Rakibul and Siddik Nayem, Abu Bakar and Sarker, Anis and Momen, Arshad and Amin, Md. Ashraful and Ahsan Ali, Amin and Rahman, Akm Mahbubur},
  booktitle={2025 IEEE International Conference on Image Processing (ICIP)},
  title={BD Open LULC Map: High-Resolution Land Use Land Cover Mapping & Benchmarking For Urban Development In Dhaka, Bangladesh},
  year={2025},
  volume={},
  number={},
  pages={2808-2813},
  keywords={Deep learning;Image segmentation;Annotations;Urban areas;Land surface;Benchmark testing;Satellite images;Reliability;Socioeconomics;Monitoring;Land Use Land Cover;Satellite Imagery;Deep Learning;BD Open LULC Map;Segmentation},
  doi={10.1109/ICIP55913.2025.11084396}
}