Building Footprints in Chittagong, Bangladesh with 40 cm Satellite Imagery

A geospatial dataset contains 5,229 high-resolution image tiles and corresponding vector labels for 38,096 individual buildings. The data covers Chittagong and parts of the Kutupalong Refugee Camp, sourced from Maxar ODP satellite imagery at 40 cm resolution. It was published by MLHUB on the NASA Earthdata platform and last updated in January 2023.

Use Cases

• Train building footprint detection models based on high-resolution 40 cm satellite imagery.
• Analyze settlement patterns and density in peri-urban and rural areas based on the labeled building locations.
• Develop and test localized remote sensing models for humanitarian applications, given the inclusion of a refugee camp area.
• Conduct land-use classification in agricultural and mixed-use landscapes using the image and label pairs.

Strengths

• Contains 5,229 image-label tile pairs, providing a substantial volume of training data.
• Includes labels for 38,096 individual buildings, enabling detailed analysis.
• Imagery has a high spatial resolution of 40 cm, allowing for precise feature identification.
• Described as a 'Tier 1' dataset, indicating it has been thoroughly reviewed and improved.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Last updated 2023-01-01 00:00:00; freshness should be verified.
• The specific file formats for the imagery and labels are not explicitly listed in the provided metadata.

Provenance

Source

MLHUB via NASA Earthdata platform.

Collection Method

Satellite imagery sourced from Maxar ODP (105001001AC98900), with corresponding building footprint labels created for model training.

Freshness

Last updated 2023-01-01 00:00:00

Geography

Chittagong, Bangladesh, including parts of the Kutupalong Refugee Camp.