A Comparison Of Remote Sensing Change Detection Methods For Urban Creep Identification In Norwich

Change detection is one of the most active research areas in remote sensing, driven by the desire to monitor the highly dynamic world around us. Modern very high resolution (VHR) images from satellite and aerial platforms present an opportunity to reveal landscape change in great detail. However, there is little research focusing on detailed change detection. This research addresses this problem, by investigating the utility of remote sensing change detection to detect fine urban changes called ‘urban creep’. Urban creep is the addition of impermeable surface to an existing property, after its initial construction. This is problematic because cumulatively, urban creep significantly increases flood risk. Moreover, up-to-date urban creep statistics are not readily available. Therefore, urban creep identification is a challenge to change detection research because of its subtle expression, small extent, and complicated contextual setting. The investigation of urban creep is conducted over the city of Norwich using aerial images from 2006 and 2010. The research focuses on three methodological areas. Firstly, statistical sampling is employed to quantify a baseline rate of urban creep. Secondly, a direct classification of the remotely sensed data is undertaken to assess the utility of the state-of-the-art in this application. And lastly, as a counter-point, change vector analysis (CVA) is applied to explore the utility and relevance of differencing methods in a complex urban change detection setting. The results have shown that 1) The rate of urban creep in Norwich is increasing; 2) Quantitative urban creep evaluation remains beyond the current state-of-the-art, but qualitative detection is possible; 3) Simple spectral CVA has almost no utility in this application. These findings contribute to knowledge by clarifying the capabilities of change detection techniques in detailed urban applications; building upon, and extending existing CVA research; and finally by adding to urban creep knowledge.