A comparative analysis of the functionality of stereo vision algorithms on various hardware architectures has been carried out. The quantitative results of stereo vision algorithms implementation are presented, taking into account the specifics of the applied hardware base. The description of the original algorithm for calculating the depth map using the summed-area table is given. The complexity of the algorithm does not depend on the size of the search window. The article presents the content and results of the implementation of the stereo vision method on standard architecture computers, including multi-threaded implementation, a single-board computer and FPGA. The proposed results may be of interest in the design of vision systems for applied applications.

synthetic vision, stereo algorithms, hardware architectures, FPGA

Efimov Aleksey Igorevich,  ,  orcid.org/0000-0002-4014-8718,  Ryazan State Radio Engineering University named after V.F. Utkin

Ustukov Dmitry Igorevich,  ,  orcid.org/0000-0002-4848-8936,  Ryazan State Radio Engineering University named after V.F. Utkin