BdWaterSurfaceSet: Satellite Images for Water Detection in Bangladesh

232.3 MB of satellite imagery samples for automated inland water surface detection in Bangladesh. Kona Moni created this dataset, which includes lakes, canals, ponds, and rivers, and was last updated on April 16, 2026. It provides annotations in COCO Segmentation Format and is pre-split into training, validation, and test subsets.

Use Cases

• Binary classification of water vs. non-water land regions based on image samples.
• Image segmentation tasks based on annotations in COCO Segmentation Format.
• Land-cover analysis based on diverse geographic locations and surface conditions.
• Change detection studies based on satellite imagery representing inland water bodies.

Strengths

• Dataset size is 232.3 MB, indicating a substantial collection of image samples.
• Includes annotations in COCO Segmentation Format, enabling segmentation and detection tasks.
• Pre-divided into training, validation, and test subsets for reproducible benchmarking.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.

Provenance

Source

figshare

Collection Method

Developed through a structured preprocessing pipeline extracting region-level samples from raw satellite images.

Freshness

Last updated 2026-04-16 07:28:34; freshness should be verified.

Geography

Bangladesh