Infrastructure growth assessment of urban areas based on multi-temporal satellite images and linear features. Issue 3 (3rd July 2017)

• Record Type: Journal Article
• Authors: Al-Ruzouq, Rami; Hamad, Khaled; Shanableh, Abdallah; Khalil, Mohamad
• Is Part Of: Annals of GIS. Volume 23:Issue 3(2017)
• Page Start: 183
• Page End: 201
• Publication Date: 2017-07-03
• Subjects: Urbanization -- change detection -- image registration -- land use -- infrastructure growth -- multi-temporal Landsat; Geographic information systems -- Periodicals; Periodicals; 910.285
• DOI: 10.1080/19475683.2017.1325935
• Languages: English
• ISSNs: 1947-5683

Abstract

ABSTRACT: Urbanization is typically demonstrated by expansion of a city's infrastructure, mainly development of its roads and buildings. In particular, transportation infrastructure is a key indicator of growth in the city since transportation is the backbone of economic development and prosperity. Recent advances in satellite imagery, in terms of improved spatial and temporal resolutions, enable efficient identification of change patterns and prediction of built-up areas. In this study, two approaches were adopted to quantify and assess the pattern of urbanization. The first approach relied on extracting linear features (buildings and roads) from multi-temporal satellite images, where image-to-image registration was utilized based on linear features with Modified Iterated Hough Transform as matching criteria. The second approach relied on extracting linear features from vector (digitized) data. The latter approach complemented the first one by distinguishing between roads and buildings. The proposed methodology was then applied to Sharjah City, United Arab Emirates, as a case study. Results show that the urbanized area of the city almost quadrupled during 1976–2016. Growth in buildings and roads was generally consistent until 2005, after that the spatial growth witnessed a steep increase due to vertical expansion. To assess the accuracy of the utilized edge images and change detection, error matrices were prepared for the case study. An overall accuracy of more than 84% was achieved. The proposed methodology was successful in quantifying urban growth in the study area.