Fast Morphological Image Processing on GPU using CUDA

Mugdha A. Rane
Department of Computer Engineering and Information Technology, College of Engineering, Pune
College of Engineering, Pune, 2013


   title={Fast Morphological Image Processing on GPU using CUDA},

   author={Rane, Mugdha A},



Download Download (PDF)   View View   Source Source   



A mathematical morphology is used as a tool for extracting image components that are useful in the representation and description of region shape. The mathematical morphology operations of dilation, erosion, opening, and closing are important building blocks of many other image processing algorithms. The data parallel programming provides an opportunity for performance acceleration using highly parallel processors such as GPU. NVIDIA CUDA architecture offers relatively inexpensive and powerful framework for performing these operations. However the generic morphological erosion and dilation operation in CUDA NPP library is relatively naive, but it provides impressive speed ups only for a limited range of structuring element sizes. The vHGW algorithm is one of the fastest for computing morphological operations on a serial CPU. This algorithm is compute intensive and can be accelerated with the help of GPU. This project implements vHGW algorithm for erosion and dilation independent of structuring element size has been implemented for different types of structuring elements of an arbitrary length and along arbitrary angle on CUDA 5.0 programming environment with GPU hardware as GeForce GTX 480. The results show maximum performance gain of 20 times than the conventional serial implementation of algorithm in terms of execution time.
No votes yet.
Please wait...

* * *

* * *

HGPU group © 2010-2021 hgpu.org

All rights belong to the respective authors

Contact us: