Towards a novel backdating strategy for creating built-up land time series data using contemporary spatial constraints

摘要

For retrospective land cover change analysis extending back in time earlier than approximately the year 2000, multispectral remote sensing data such as the Landsat archive are often the only available and systematically archived data source. However, the spatial resolution of such observations often impedes the change analysis of geographic phenomena that may occur at sub-pixel level, such as changes in built-up land. In such cases, the integration of these data products with increasingly available contemporary high resolution settlement data employed as spatial constraints can potentially mitigate this drawback. Little research has been done regarding the quantitative potential of such integrative approaches, often due to the lack of multi-temporal reference data about changes in built-up area at sufficient spatio-temporal granularity and extent. In this contribution, we present and evaluate a time-series based approach for built-up area change analysis using Landsat time series data, spatially constrained by contemporary building footprint data. We evaluate the potential of our approach using a highly accurate multi-temporal reference database of built-up areas created through integrating publicly available building footprint and cadastral data in selected regions of the United States of America.