Conference paper

Evaluation of Embedded Devices for Real-Time Video Lane Detection

K. Podbucki, J. Suder, T. Marciniak, A. Dąbrowski (Poznan Univ. of Techn., Poland)

This paper presents a comparison of the performance of embedded systems processing video sequences in real time. As part of the work, practical programs for detecting lanes located on airport areas, which allow autonomous vehicles to move around the airport, were tested. The following modules were used during the tests: Raspberry Pi 4B, NVIDIA Jetson Nano, NVIDIA Jetson Xavier AGX. For modules from the NVIDIA Jetson family, the maximum performance of video stream processing (FPS) depending on the resolution and the selected power mode has been checked. The results of the experiment show that NVIDIA Jetson modules have sufficient computing resources to effectively track lines based on the camera image, even in low power modes.

Download one page abstract

Receipt of papers:

March 15th, 2022

Notification of acceptance:

April 30th, 2022

Registration opening:

May 12th, 2022

Final paper versions:

May 20th, 2022