Abstract
In this paper, we present a comparative study on parallel edge detection algorithms upon high-resolution satellite images, implemented on OpenACC, Hybrid OpenMP/MPI, OpenMP, and MPI models on which the Sobel, Prewitt and Canny algorithms were developed using C++ language and OpenCV. The performance of these computing models were measured in terms of speedup and execution time by implementing the edge detection algorithms using various sized images and programming constructs. The program implementations using OpenACC display the largest speedup in CPU time, which is followed by the Hybrid OpenMP/MPI model. The parallel detection algorithms using OpenACC obtain the greatest speedup of around 6.5 over OpenMP model. The parallel Sobel and Prewitt algorithms are relatively 2 times faster than the Canny in all respects.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call