Cordic architecture for Hough Transform applications

Abstract

The Hough transform is a widely used technique by converting the original spatial information in an image into a parameter space representation. It can be applied to detect straight lines, circles, ellipses and various other curves in two dimensional scenes as well as in the recognition of three dimensional objects. In this article, the trigonometric cosine and sine functions required in the computation of Hough transform are realized using a CORDIC based floating-point arithmetic scheme. The Hough transformation is simulated by Simulink simulation tools. The architecture of image enhance processing based on Laplace filtering is verified and simulated with Cadence design tools. The image enhancement system is also demonstrated on the Cyclone II FPGA device.