Translator Disclaimer
15 December 2000 Image processing and high-capacity hardware for detection and recognition
Author Affiliations +
This contribution presents a high-capacity computing device for the detection and recognition of objects appearing or moving in an infrared movie. A solution is shown as to how to extract objects from a data flow of up to 20 million pixels per second in real-time and how to display a target immediately after its appearance in the image observed. The hardware capability requirement is pointed out and the corresponding realization is explained on the basis of the algorithms. Currently, 17 high-performance microprocessors are combined for this special task. Modular design based on VMEbus technology, parameterization of hardware and the constraints to realize as many of the algorithms as possible in software guarantee a considerable growth potential by simply redoubling modules and adding new functions. Part of this paper discusses the human interface of the device, for it is evident that the value of the system depends on the quality of the information given to the operator. Results obtained from simulation and from functional tests on the device and possible future upgrades of the system are presented. The work performed to date has shown that image- processing software, implemented in a powerful hardware is the key issue for passive automatic target detection and recognition.
© (2000) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Johann Lang "Image processing and high-capacity hardware for detection and recognition", Proc. SPIE 4130, Infrared Technology and Applications XXVI, (15 December 2000);


Multi-target tracking in clutter
Proceedings of SPIE (June 25 2002)
A novel shot boundary detection framework
Proceedings of SPIE (June 24 2005)
Vehicle real-time attitude-estimation system (VRAES)
Proceedings of SPIE (June 07 1996)
Handoff between track-while-scan and image tracking
Proceedings of SPIE (July 15 1999)

Back to Top