Affiliation:
1. Hubei University of Technology, Wuhan, China
Abstract
In order to optimize the workflow of iterative 3D reconstruction and support the goal of massive image data processing, high performance and high scalability, this article proposes an image distributed computing framework FIODH. It is a distributed computing framework based on distributed hash algorithm, which accomplishes the task of storing, calculating and merging the image data in multiple nodes. A SIFT algorithm is used to extract feature points from the original images which are distributed on the hash nodes. During the process of image clustering computation, the agent nodes are responsible for task management and intermediate result calculation. The clustering results in hierarchical trees which can be converted into computational tasks and assigned to the appropriate nodes. The experimental analysis shows that the algorithm has achieved satisfactory results in efficiency and error adjustment. In a large amount of experiment data, the advantage of the algorithm is more obvious.
Reference26 articles.
1. Min-Wise Independent Permutations
2. Crandall, D., Owens, A., Snavely, N., & Huttenlocher, D. (2011, June). Discrete-continuous optimization for large-scale structure from motion. In 2011 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 3001-3008). IEEE.
3. A dependability layer for large-scale distributed systems
4. Locality-sensitive hashing scheme based on p-stable distributions.;M.Datar;Proceedings of the twentieth annual symposium on Computational geometry,2004
5. Enterprise distributed service platform–network architecture and topology optimisation.;I.Demydov;International Journal of Space-Based and Situated Computing,2012