Semi-Automated Method to Extract Urban Areas From Barren Land/ Bare Soil, Case Study: Idku, Nile Delta Coast, Egypt

Optical remote sensing offers a time and cost-effective manner to monitor urban-form expansion over time. The availability of recent and historical remote sensing data together with the wide coverage of these data make remote sensing application further favorable for the purpose. However, urban areas surrounded by barren land/ bare soil tend to be overestimated due to the similar spectral behavior of both urban areas and their surroundings. This paper aims to develop a methodology that overcomes the interference between urban areas and barren land/ bare soil and thus enhances the urban areas’ extraction in a semi-automated manner. For this purpose, SAVI, NDBI and MNDWI indices, PCA and thermal band were employed to extract urban areas from their surroundings in two Landsat images for the years 1998 and 2016. The developed methodology was applied to extract urban form of Idku Town, one of the coastal urban centers of the Nile Delta, with an overall accuracy of 96.48% and 93.39% for the images of 1998 and 2016, respectively. The results have revealed that the urban form of Idku Town was tripled over the last two decades with an annual growth rate of about 12%.

Keywords: urban-form; Barren land; bare soil; remote sensing

*Corresponding author: Tel.: (203)4249290 Fax: (203)4249290  

                                           E-mail: hanady.h.khalil@gmail.com