Agent-based Hough Transform: A Way to the Improvement of the Execution Time in the Detection of the Dominant Straight Line in an Image

Abstract

This paper focuses on the minimization of the time of the dominant straight line detection in an image using Hough Transform algorithm. The idea is a mixture between two domains, namely image processing and multi-agent systems. The importance of this work comes from the relying of image processing techniques on hardware accelerations. This paper demonstrates how the distribution of a purely sequential processing on a set of agents leads to an improvement from time execution point of view. The purpose is to reduce the execution time of Hough transform technique through the distribution of the algorithm on a set of reactive agents. This may allow the exploitation of a parallel or distributed environment. The main idea is based on the division of similar repeated processing with different parameters on several agents. It is a SIMD-like architecture according to Flynn classification. The obtained results are promising in the way that the execution time is at least divided by 4 comparatively to the use of the algorithm in its sequential form.