5 October 2017 The implementation of aerial object recognition algorithm based on contour descriptor in FPGA-based on-board vision system
Author Affiliations +
Abstract
This paper describes the aerial object recognition algorithm for on-board and stationary vision system. Suggested algorithm is intended to recognize the objects of a specific kind using the set of the reference objects defined by 3D models. The proposed algorithm based on the outer contour descriptor building. The algorithm consists of two stages: learning and recognition. Learning stage is devoted to the exploring of reference objects. Using 3D models we can build the database containing training images by rendering the 3D model from viewpoints evenly distributed on a sphere. Sphere points distribution is made by the geosphere principle. Gathered training image set is used for calculating descriptors, which will be used in the recognition stage of the algorithm. The recognition stage is focusing on estimating the similarity of the captured object and the reference objects by matching an observed image descriptor and the reference object descriptors. The experimental research was performed using a set of the models of the aircraft of the different types (airplanes, helicopters, UAVs). The proposed orientation estimation algorithm showed good accuracy in all case studies. The real-time performance of the algorithm in FPGA-based vision system was demonstrated.
Conference Presentation
© (2017) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Pavel Babayan, Sergey Smirnov, Valery Strotov, "The implementation of aerial object recognition algorithm based on contour descriptor in FPGA-based on-board vision system ", Proc. SPIE 10430, High-Performance Computing in Geoscience and Remote Sensing VII, 1043006 (5 October 2017); doi: 10.1117/12.2277920; https://doi.org/10.1117/12.2277920
PROCEEDINGS
9 PAGES + PRESENTATION

SHARE
RELATED CONTENT

Automatic stereoscopic system for person recognition
Proceedings of SPIE (June 22 1999)
Neural networks for handwriting recognition
Proceedings of SPIE (September 16 1992)
A Rule Based System For The Recognition Of Line Drawings...
Proceedings of SPIE (November 20 1986)
Object-oriented recognition for automatic inspection
Proceedings of SPIE (March 11 1994)
Cascade fuzzy ART a new extensible database for model...
Proceedings of SPIE (February 27 1996)

Back to Top