Automatic Target Detection And Recognition: A Wavelet Based Approach
Abstract
Wavelet based target detection and identification algorithms for radar applications are presented and tested and evaluated on computer simulated data. The algorithms make use of a scale sequential and/or scale recursive paradigm where computations are performed within and across scales in a multiresolution analysis (MRA) of the sensor data relative to a compactly supported discrete orthonormal wavelet basis. It is argued that such procedures are computationally efficient and offer promise of yielding near optimal performance with a minimum CPU time burden. Specific applications considered in the report include automatic target identification from high range resolution radar (HRR), target detection in the presence of fractal noise and the integration of multisensor data in the tracking of aircraft. Other applications addressed include scale recursive optimal filtering and the synthesis of parallel architectures for the 1-D discrete wavelet transform. The report includes a full discussion of the theory behind the various detection and identification algorithms plus results from Monte Carlo simulations.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 25, 1997
- Accession Number
- ADA329696
Entities
People
- A. J. Devaney
- E. Manolakos
- H. Lev-ari
- Mieczyslaw M. Kokar
- R. S. Raghavan
Organizations
- Northeastern University