Extraction of impervious surface area using orthophotos in rhode island

Abstract

Suburban sprawl has been identified as the most important social and environmental issue facing Americans in their community. Suburban development consumes green space, widens urban fringes, increases impervious surface areas (ISA), and puts pressure on environmentally sensitive inland and coastal areas. Urban runoff, mostly through ISA, is the leading source of pollution in the Nation's estuaries, lakes, and rivers. ISA can serve as a key environmental indicator due to its impacts on water systems and its role in transportation and concentration of pollutants. Quantifying precise spatial locations and distributions of ISA has become increasingly important with growing concern over water quality in this country. Classification of high spatial resolution remote sensing data is an important method to obtain ISA information. In this study we used 1-meter spatial resolution true color digital aerial photography data for extracting ISA information in the coastal Rhode Island. We developed a synthetic algorithm based on the multiple agent segmentation and classification. In this algorithm, the indices describing shape feature of objects were introduced in multiple agent segmentation. The shadow problem that is common in high spatial resolution remote sensing images has also been considered. Existing GIS data were used in the classification and post-classification process. The testing result indicates that this synthetic algorithm performs well in obtaining precise ISA areas and the result is better than that from pixel-based classification.