The Image Foresting Transform: Theory, Algorithms, and Applications
IEEE Transactions on Pattern Analysis and Machine Intelligence
Digital Image Processing (3rd Edition)
Digital Image Processing (3rd Edition)
Hybrid MPI/OpenMP Parallel Programming on Clusters of Multi-Core SMP Nodes
PDP '09 Proceedings of the 2009 17th Euromicro International Conference on Parallel, Distributed and Network-based Processing
How will astronomy archives survive the data tsunami?
Communications of the ACM
Hi-index | 0.00 |
There are several projects and missions designed to strictly observe the Sun. These projects usually produce a large amount of information embedded in images. The analysis of such information is valuable for the study and monitoring of solar storms that can affect telecommunications, for instance. The databases sizes with sun image are huge. Several projects are producing images of the Sun and exists a considerable amount of stored images. Combining image processing algorithms with parallel programming techniques we can compute such information faster and a major volume. This paper describes our parallel OpenMP-MPI hybrid solutions for processing Sun images, and our results obtained in a hybrid system, i.e. a cluster with several multi-core nodes. Specifically, we present two methods to detect and categorize solar filaments in hybrid systems: Filament Diffusion-Detection based on graphs and Morph Detection, based on morphological operators. The results show that the Filament Diffusion-Detection based on graphs detects approximately 80% of the filaments, with a 326-fold speed-up over. In turn, Morph Detection detects 58% of the objects with a 54-fold increase in speed. Overall, these results show that our OpenMP-MPI combination works well for hybrid architectures, but more optimizations are needed to improve accuracy.