FPGA Implementation of Gradient Based Edge Detection Algorithms for Real Time Image

摘要

In this paper a novel technique to implement gradient based edge detectors on FPGA platform is demonstrated. Three most popular gradient based edge detection algorithms viz. Sobel, Prewitt and Robert Cross are modelled in hardware description language and implemented on Terasic VEEK-MT FPGA development board. The RGB image is first transformed into gray scale using luminosity method which is fed to line buffers to apply appropriate edge detection mask. An edge detection processor block is designed wherein the relevant filter mask is applied followed by generation of horizontal and vertical gradient. The root mean square value of the two gradient components is computed. The pixel value so obtained is compared against a user defined threshold before applying to the LCD module for display. The proposed implementation is tested on both still and video images. Implementation result endorses the successful design of gradient based edge detection algorithm on FPGA platform.