Evaluation of Different Satellite Images for Urban Land Use Analysis: A Case Study of Greater Manchester, United Kingdom

Abstract

Background: The complexity of an urban area makes mapping it very difficult as its surface materials are highly spatial and spectrally diverse. This study evaluates the problems associated with remote sensing of urban land cover types. Methodology: Three satellite image sensors; Disaster Monitoring Constellation (DMC), Landsat TM and colour infrared were used to investigate their potential in mapping and characterizing land cover in a part of Greater Manchester. Supervised and unsupervised image classifications were used to map urban land use and cover. Results: The satellite image sensors and their accuracy were statistically tested to see if there is a significant relationship between them. The colour infrared image was the best in discriminating among different types of land cover with an overall accuracy of 80% followed by the Landsat image with an overall accuracy of 61% while the DMC image had the least potential in discriminating among different types of land cover with an overall accuracy of 55%. Conclusion: The colour infrared image is the most suitable for urban land cover analysis as the misclassifications are minimal compared to the other two and the features can be vividly recognized due to its spatial resolution.