Implementation and evaluation of parallel edge detector on cluster of workstations using MPI

Many image processing applications can exploit high degree of data locality and parallelism to improve performance. Customized massively parallel computing hardware provides the most efficient implementation as far as performance is concerned but ore the other hand the cost of such implementations is very high. Now-a-days high bandwidth networks with high performance workstations is becoming economical and are widely available. Hence, better performance-to-cost ratio cars be achieved by implementing message-passing paradigm of parallel processing. In this paper we develop parallel edge detector that runs on a cluster of workstations and then investigate the performance of this application in terms of execution lime, speedup and efficiency for tasks with different computational intensities. Effects of the sequential I/O component and load balancing are also discussed.