Expediting Infrastructure Condition Assessment for Disaster Response and Emergency Management Using Remote Sensing Data

This paper will discuss how urban development and population trends show a significant growth of the built environment in the coastal United States. About 60% of the USA population lives in urban coastal communities. The impact of built up areas and timely evaluation of environment disasters are keys to planning sustainable future developments. Recent advances in spaceborne remote sensing and geospatial analysis technologies offer new opportunities to expedite infrastructure inventory and condition assessment. This paper presents the results of a satellite imagery-based surface classification methodology using spectral reflectance criteria derived from groundtruth samples of different built and non-built surface types. To minimize misclassification between non-built natural surfaces and built-up areas, polygons are used to buffer water bodies and built-up area pixels prior to their classification. The methodology has been validated for sample areas from Gulfport, Mississippi, using pre-Hurricane Katrina 1-m Ikonos multispectral imagery. The maximum residual error of < ±7% for all surface types, with respect to the groundtruth, is significantly more accurate compared to traditional supervised classification methods. The surface classification results have been used to develop a GIS-based decision support system for assessing storm debris and erosion damages by analyzing post-disaster imageries. Additionally, the geospatial analysis using high resolution satellite imagery is being implemented to create inventory of transportation infrastructure and building footprints.

Language

  • English

Media Info

  • Media Type: CD-ROM
  • Features: Figures; Maps; Photos; References; Tables;
  • Pagination: 10p
  • Monograph Title: International Conference on Maintenance and Rehabilitation of Pavements and Technological Control (MAIREPAV6), Sixth Proceedings

Subject/Index Terms

Filing Info

  • Accession Number: 01139721
  • Record Type: Publication
  • Files: TRIS
  • Created Date: Sep 18 2009 7:07AM