Object-based classification of residential land use within Accra, Ghana based on QuickBird satellite data

Area of Science:

• Remote Sensing
• Geographic Information Systems (GIS)
• Urban Planning

Background:

• Accurate mapping of urban land use and socio-economic status is crucial for effective urban planning and resource allocation.
• Satellite imagery offers a valuable tool for large-scale urban environmental assessment.

Purpose of the Study:

• To delineate residential land use polygons in Accra, Ghana.
• To identify low and high socio-economic status neighborhoods using QuickBird multispectral satellite data.
• To evaluate two object-based classification strategies for urban socio-economic mapping.

Main Methods:

• A segmentation and hierarchical classification approach was applied to QuickBird multispectral satellite data.
• Two object-based classification strategies were tested: one based on spatial frequency, the other on Vegetation-Impervious-Soil (VIS) sub-object proportions.
• Accuracy assessment was performed using test objects derived from visual interpretation of QuickBird panchromatic imagery.

Main Results:

• Both classification approaches yielded comparable results for residential land-use mapping.
• Overall percentage accuracy was 75%, with a kappa index of agreement of 0.62.
• The study successfully delineated residential areas and differentiated socio-economic statuses.

Conclusions:

• Object-based image analysis is effective for mapping urban land use and socio-economic status from high-resolution satellite data.
• The developed methodology provides a valuable tool for urban planners and researchers in Accra and similar urban environments.
• Further research could explore integrating additional data sources to enhance classification accuracy.