Mapping impervious surface for new jersey watershed management area 5 using aster imagery

Document Type

Conference Proceeding

Date of Original Version



In this paper we present methods to estimate and map impervious surface distribution in an urban watershed using Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) imagery for Watershed Management Area 5 in the Northern New Jersey. Spectral Mixture Analysis (Unmixing) technique was used on a vegetationimpervious surface-soil (V-I-S) model to separate the urban pixels into a linear combination of three typical land cover types: vegetation, impervious surface, and bare soil. Endmembers of the V-I-S components were selected directly from the image and used to model the mixture of urban land surface. We also adopted a method to estimate impervious surface using normalized difference vegetation index (NDVI). The accuracy of estimated impervious surface was evaluated using New Jersey DEP 1-foot infrared Orthophoto. Totally 100 polygons with areas between 8 and 13 acres were randomly selected in five different urban categories and the percentage of impervious surface area of each polygon was digitized and calculated. The results show that NDVI method obtained higher accuracy with approximately 10% estimation error. The modified two-endmember unmixing approach had similar accuracy of 12 % estimation error, but it tended to underestimate impervious surface in this study.

Publication Title, e.g., Journal

American Society for Photogrammetry and Remote Sensing - Annual Conference 2005 - Geospatial Goes Global: From Your Neighborhood to the Whole Planet



This document is currently not available here.