Use of Remote Sensing to Screen Earthen Levees

Aanstoos, J.V., Hasan, K., O'Hara C., Prasad, S., Dabbiru, L., Mahrooghy, M., Nobrega, R. A. A., Lee, M. L., & Shrestha, B. (2010). Use of Remote Sensing to Screen Earthen Levees. Proc. 2010 39th IEEE Applied Imagery Pattern Recognition Workshop. Washington, DC: IEEE.

Earthen levees protect large areas of populated and cultivated land in the US from flooding. The potential loss of life and property associated with the catastrophic failure of levees can be extremely large. Over the entire US, there are over 100,000 miles of levee structures of varying designs and conditions. Currently, there are limited processes in place to prioritize the monitoring of large numbers of dam and levee structures. Levee managers and federal agencies need to assess levee health rapidly with robust techniques that identify, classify and prioritize levee vulnerabilities with lower costs than traditional soil-boring programs, which can cost many of millions of dollars and provide information about the subsurface only in the immediate vicinity of a small-diameter borehole. This paper gives an overview and presents early results of an ongoing project studying the use aperture radar (SAR) in conjunction with other remote imagery as an aid to the levee screening process. SAR sensors being utilized include: (1) The NASA UAVSAR (Uninhabited Aerial Vehicle SAR), a fully polarimetric L-band SAR which is specifically designed to acquire airborne repeat track SAR data for differential interferometric measurements. The instrument is capable of sub-meter ground sample distance. (2) Recent space-based SAR platforms including the Canadian Radarsat-2 and the German TerraSAR-X. These both have spatial resolutions approaching that of the airborne SAR and also have multi-polar options but use different carrier frequencies. Our test study area is a stretch of 230 km of levees along the lower Mississippi River. SAR interferometric mode is capable of identifying vertical displacements on the order of a few millimeters. The L-band measurements can penetrate soil to as much as one meter depth. Thus it is valuable in detecting changes in levees that will be key inputs to a levee vulnerability classification system. We report on the use of various feature detection algorithms being applied to the polarimetry data, including entropy-anisotropy decomposition, wavelet transforms, and methods based on the Grey Level Co-occurrence Matrix (GLCM). The features detected are compared with various ground truth data including soil conductivity measurements, soil sample tests, and on site visual inspections.