UDC 004.932, 004.042
PROCESSING OF ADJACENT STEREO FRAMES USING MULTI-CORE COMPUTING SYSTEMS
A. K. Linkov, post-graduate student, RTU MIREA, Moscow, Russia;
orcid.org//0009-0004-3725-508X, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
This paper presents the process of adjacent stereo frames analysis, specifically the method of their pre processing, as well as an algorithm for matching image fragments (local features) and generating a depth map. The structure of the algorithm, which consists of nested loops, is shown. This implies the possibility of executing such algorithms on various multi-core computing systems, using parallel computing techniques. These techniques would significantly improve processing performance, bringing it closer to real time scale. The existing classes of multi-core computing systems are analyzed. Their features and applicability to the task of stereo frame processing are reviewed. Experimental studies were carried out using the developed program in Python language and auxiliary libraries. The execution time of the algorithm was measured in single-threaded mode, using software optimi zation, and on a massively parallel device – graphics processor. The aim of the work is a comparative analysis of different types of computing systems, as well as demonstration of optimization possibilities for nested loops.
Key words: : Istereo image, depth map, filtering, matching algorithms, parallel computing, computer ar chitecture, graphics processor, FPGA.