Publication of the journal

The section is currently being updated

The use of systems on a chip (SoC) including the processor part (PS part) and programmable logic (PL part) in embedded systems lowers power consumption and increases data processing speed. The PYNQ project provides a Jupyter Notebook-based platform with a Python API that allows the interaction with elements placed in the PL part. Using PYNQ makes it possible to separate the parts of the code responsible for user interaction and the hardware blocks implemented in the PL part. In this work, the problem of moving a robot along a line using a computer vision system is considered as an example of PYNQ technology use. The video stream enters the PL part via the TMDS interface. When using only processor cores for line allocation on image, data is stored in RAM by the DMA Video controller (VDMA) and processed using the OpenCV library. Several configurations were prepared that demonstrate the phased transfer of image processing operations to the PL part. As a result, a pipeline has been formed that includes the stages of video compression, conversion from RGB color space to HSV, color extraction and counting of image moments. Using moments, the position of the center of the selected object (colored line) in the field of view of the camera was determined. The performance of processing a video stream with only processor cores (PS part) and a combination of processor cores and FPGAs (PL and PS parts) of the SoC under consideration was compared. The substantial performance increase in PL part has been detected in relation to variant of full processing in PS part. The presented colored line detection module can be used as part of a robot oriented by lidar and stereogram camera data.
  • Key words: FPGA, PYNQ, OpenCV, computer vision
  • Published in: INFORMATION-COMMUNICATION TECHNOLOGIES
  • Bibliography link: Konchenkov V. I., Mangushev A. V., Markov A. E. Computer vision system for a mobile robot based on the PYNQ platform. Izv.vuzov. Elektronika = Proc. Univ. Electronics. 2025;30(2):217–228. (In Russ.). https://doi.org/10.24151/1561-5405-2025-30-2-217-228.
Vladimir I. Konchenkov
Volgograd State Technical University (Russia, 400005, Volgograd, Lenin Ave, 28); Volgograd State Socio-Pedagogical University (Russia, 400005, Volgograd, Lenin Ave, 27)
Alexander V. Mangushev
Volgograd State Technical University (Russia, 400005, Volgograd, Lenin Ave, 28)
Alexey E. Markov
Kaspersky Lab JSC (Russia, 125212, Moscow, Leningradskoe Hwy, 39a, bld. 3)

124498, Moscow, Zelenograd, Bld. 1, Shokin Square, MIET, editorial office of the Journal "Proceedings of Universities. Electronics", room 7231

+7 (499) 734-62-05
magazine@miee.ru