Title :
Electro-optical systems for image recognition
Abstract :
Detection and classification of targets by means of electro-optical sensors has been of interest for both defense and civilian applications for a number of years. In this presentation we emphasize two areas: extraction of more information from the sensors by polarimetric and frequency diversity, and reducing the processing and storage requirements by exploiting invariant attributes for detection and classification
Keywords :
computational complexity; image classification; image colour analysis; image sensors; infrared imaging; invariance; object detection; object recognition; optical radar; radar imaging; radar polarimetry; radar target recognition; sensor fusion; surveillance; 3D target surface orientation; Bayesian approach; Doppler imagery; Laplacian pyramids; Stokes vector parameters; automatic target recognition; composite image; computational complexities; dissimilar sensor types; electro-optical sensors; frequency diversity; image contrast; image fusion; image recognition; image understanding; infrared images; invariant algebra; invariant attributes; ladar imagery; multispectral sensing; polarimetric diversity; polarimetric imaging sensors; similar sensor types; situational awareness; solid-state color vision system; surveillance; target classification; target detection; visible images; Diversity reception; Electromagnetic wave polarization; Image fusion; Image recognition; Image sensors; Laplace equations; Layout; Optical imaging; Real time systems; Sensor phenomena and characterization;
Conference_Titel :
Lasers and Electro-Optics Society, 2001. LEOS 2001. The 14th Annual Meeting of the IEEE
Conference_Location :
San Diego, CA
Print_ISBN :
0-7803-7105-4
DOI :
10.1109/LEOS.2001.968933
