Abstract

This paper describes the design and implementation of two image-processing algorithms using the custom computing platform. is a reconfigurable system that can be tailored to perform a wide variety of tasks. The particular tasks discussed here are the Hough transform, a well-known technique for detecting lines in an image, and pyramid generation, the process of transforming a single image into a set of filtered images with successively lower spatial resolution. This paper describes how these computationally intensive processes have been mapped onto hardware. Both processes have been designed to operate at high speed. In particular, the generation of both Gaussian (low-pass) and Laplacian (band-pass) pyramids can occur concurrently in real time using images from a video camera, assuming the standard frame rate of 30 images per second. Results are presented to illustrate the efficacy of reconfigurable FPGA-based machines to image processing applications.

Citations