Urban Land Cover Maps for Two African Cities from High-Resolution Satellite Imagery

2020 urban land cover maps for Mekelle, Ethiopia and Polokwane, South Africa. The dataset was generated by NASA using an object-based image classification approach on Planet SuperDove and Worldview-3 satellite imagery at multiple spatial resolutions. It supports fine-scale urban analysis and comparative assessments of different high-resolution imagery sources.

Use Cases

• Comparing land cover classification accuracy between 3.125-m Planet SuperDove and 0.5-m/2-m Worldview-3 imagery.
• Analyzing the distribution of urban land cover classes (e.g., built-up, vegetation) at a fine scale for specific city districts.
• Training or validating machine learning models for object-based land cover classification from satellite data.
• Conducting temporal urban growth studies if combined with land cover data from other years.

Strengths

• Provides very high-resolution data with spatial resolutions of 0.5 m, 2 m, and 3.125 m from two satellite sources.
• Covers two distinct African cities, enabling comparative urban studies.
• Uses an object-based image classification approach, which can improve accuracy over pixel-based methods for land cover.

Limitations

• Row/feature count, dataset size, and specific column names are unknown across all sources.
• Conflicting last updated dates: datagov lists 2026-03-12, while nasa_earthdata lists 2020-12 31.
• License is specified as 'other-license-specified' on datagov but is null on nasa_earthdata, requiring user verification.

Provenance

Source

National Aeronautics and Space Administration (NASA)

Collection Method

Object-based image classification applied to Planet SuperDove and Worldview-3 (Maxar Technologies) satellite imagery.

Time Range

2020

Freshness

2026-03-12 20:29:28.378325

Geography

Mekelle, Ethiopia and Polokwane, South Africa.