Updating land-cover maps by using texture information from very high-resolution space-borne imagery

The process of updating existing land-cover databases is an issue of great concern and is needed to maintain the value of existing and expensive information. Often, this problem is one of integrating appropriately multitemporal data sets that are completely different in nature, such as the updating of old vector data with new image data. In this paper, a methodology is proposed for the updating of land-cover databases of rural areas by using very high-resolution panchromatic space-borne imagery. The task is to identify the extension of urban and industrial areas. At a given scale, the homogeneity of texture derived from newly acquired imagery is measured within the polygons to update by means of a knowledge-based segmentation approach. In a successive step, the new region boundaries are compared with the existing polygons. This basically transforms the change-detection problem into one of image segmentation. Experimental results for the updating of land-cover maps by IRS-1C panchromatic imagery indicate the usefulness of the direction that is adopted