EXP-SA: Wavelet Feature Extraction and Pattern Recognition in Imagery Surveillance for Detecting Roadside Explosives

University Of California-San Diego, La Jolla CA

Investigators

Abstract

The project will develop an imagery analysis system based on wavelet feature extraction for roadside bomb detection based on two-dimensional wavelet transforms. The use of multiple features as described in this project, rather than individual ones, related to the shape, the surface texture and the surface brightness of known IED camouflages, are likely to better distinguish these objects from background scenes. This work offers an alternative to more standard imagery change detection. A module on imagery surveillance will be developed for a graduate level class for a new Master's Degree program established at UCSD's Structural Engineering Department jointly with the Los Alamos National Laboratory.