Urban vegetation mapping through pixel-based image analysis of high-resolution satellite imagery / Haslina Hashim ... [et al.]

Urban vegetation and land use information are critical for sustainable environmental management in urban areas. In general, urban vegetation plays an important role for urban planning through a balance between the natural environment and the built environment. Thus, mapping of urban vegetation is important towards sustainable urban development. Remote sensing has increasingly been used to derive such information for mapping and monitoring the changes of urban vegetation. The use of remote sensing data for urban mapping has increased along with the availability of very high resolution (VHR) satellite data such as Quickbird, Worldview and Pleiades. The aims of this study is to identify and classify using remote sensing methods in the context of a vegetation mapping in the urban environment. This paper describes the use of high-resolution Pleiades imageries to extract and classify vegetation in an urban area with the use of pixel-based image analysis. Classification types in the study area were divided into vegetation and non-vegetation classes. The pixel-based method was applied, and a support vector machine algorithm was used for classification of urban vegetation. Comparison of accuracies was made from the error matrices, overall accuracy, and kappa coefficient. The overall accuracy for classification approach was 98.980% and the kappa value was 0.9795. The result shows the ability of high-resolution imagery to extract urban vegetation accurately despite the complex surface of the urban area. This information is useful to support other research applications related with urban green spaces monitoring purposes and to the state authorities for future planning in the conservation of urban vegetation areas.