Urban Changes with Satellite Imagery and Lidar Data

It is important to know how urban areas change for monitoring the evolution of demographics, lifestyles, and economic trends. The objective of this paper is to present a method for determining the occurrence of building change between two dates. The information on buildings for the first date comes from digital cartography and for the second date comes from spatial imagery and LIDAR data. A Support Vector Machine algorithm with automatic training was applied to detect the buildings on the second date and then the results were compared with the buildings of the first date, represented by vectorial data from the digital cartography. It is argued how urban vegetation change could also be derived from the building change study. Two areas of four square kms each, from central Spain, were considered to test the proposed approach. The first area corresponds to the university Campus of Alcala, where few changes have happened in the last years; and the second area corresponds to a new residential area in the suburb of the city of Alcala, with many constructions in the last years. Results showed a 1% and 7% increase of buildings respectively. The proposed method was evaluated for efficacy and suitability using ground truth obtained manually for the second area. Some discussion and conclusions are stated about the approach to be used and the best data entry in order to obtain an optimal performance. * Corresponding author.