The article contains materials on the digital image representation in computer systems as a matrix of a finite set of pixels. Materials on spatial and frequency domain methods for digital image processing are also given. As an example, spatial methods of digital filtering with Sobel, Prewitt operators and frequency domain method of optimal filtering are compared. The described methods can be effectively implemented in a heterogeneous computing environment of parallel-conveyor data processing using graphics processors or high-capacity FPGAs. It is shown that the implementation of such a parallelconveyor system provides the required performance for large data processing by parallelizing the computations between the nodes of the system, each of which is optimized for a specific task to be solved. In turn, this can significantly improve the digital image processing speed or increase the number of parallel processing data streams compared to systems that use only the CPU for computations.
METHODS OF DIGITAL IMAGE PROCESSING FOR IMPLEMENTATION IN PARALLEL– CONVEYOR DATA PROCESSING SYSTEMS IN THE HETEROGENEOUS COMPUTING ENVIRONMENT
2017
Aufsatz (Zeitschrift)
Elektronische Ressource
Unbekannt
Metadata by DOAJ is licensed under CC BY-SA 1.0
New parallel vision environment in heterogeneous networked computing [3452-06]
British Library Conference Proceedings | 1998
|Methods and systems for a data marketplace in a conveyor environment
Europäisches Patentamt | 2022
|