A hybrid system is proposed for automatic target recognition. The system developed consists of an optical module and an electronic/computer module. The electronic module uses a DatacubeTM image processing system to carry out real-time filtering, detection, and segmentation operations. The optical module is composed of two LCTVs (liquid crystal television), a He-Ne laser, a collimator, and a CCD camera. This module calculates specific energy spectra in the Fourier plane of a potential target identified by the electronic module. These values are fed in a neural network implemented on Sun SPARC 10TM model 41 for the classification of the current target. Experimental results involving special classes of target and their distribution in the features space are presented.
[1]
Bir Bhanu,et al.
Automatic Target Recognition: State of the Art Survey
,
1986,
IEEE Transactions on Aerospace and Electronic Systems.
[2]
Jaakko Astola,et al.
On computation of the running median
,
1989,
IEEE Trans. Acoust. Speech Signal Process..
[3]
D. R. Hush,et al.
Improving the learning rate of back-propagation with the gradient reuse algorithm
,
1988,
IEEE 1988 International Conference on Neural Networks.
[4]
William K. Pratt,et al.
Pseudomedian Filter
,
1985,
Photonics West - Lasers and Applications in Science and Engineering.