A complete and efficient real time system for line segment detection based on the binary Hough transform

Author

da Fontoura Costa, Luciano ; Sandle, Mark B.

Author_Institution

Dept. of Electron. & Electr. Eng., London Univ., UK

fYear

1990

fDate

6-8 Jun 1990

Firstpage

205

Lastpage

213

Abstract

A complete and efficient system for detection of line segments in real-time (video rate) based on the binary Hough transform is described. This system is able to operate in video rate for images with dimensions up to 1024×1024 for a clock of 32 ns. Special attention is paid to the descriptions of the main processes (image acquisition, edge detection. Hough transform, peak detection, and connectivity analysis) as well as to their balancing, interrelation, and implementation in dedicated systolic hardware. Expressions are derived for the total number of basic operators, as well as the execution rate for each main process

Keywords

computer architecture; computer vision; fast Fourier transforms; real-time systems; balancing; binary Hough transform; connectivity analysis; dedicated systolic hardware; edge detection; image acquisition; interrelation; line segment detection; peak detection; real time system; video rate; Cameras; Computer vision; Educational institutions; Hardware; Image analysis; Image edge detection; Image segmentation; Proposals; Real time systems; Robotics and automation;

fLanguage

English

Publisher

ieee

Conference_Titel

Real Time, 1990. Proceedings., Euromicro '90 Workshop on

Conference_Location

Horsholm

Print_ISBN

0-8186-2076-5

Type

conf

DOI

10.1109/EMWRT.1990.128252

Filename

128252