•  
  •  
 

Turkish Journal of Electrical Engineering and Computer Sciences

Author ORCID Identifier

MUHAMMED ADIYAMAN: 0000-0003-2400-9897

İSMAİL BAŞKAYA: 0000-0001-6743-3992

Abstract

Real-time depth estimation is crucial in many vision-related tasks, including autonomous driving, 3D reconstruction, robotics, and simultaneous localization and mapping. In recent years, many methods have been proposed to solve depth maps from images by utilizing different modality setups like monocular vision, binocular vision, or sensor fusion. However, for real-time deployment on edge devices, complex methods are not suitable due to latency constraints and limited computation capacity. For edge implementation, models should be simple, minimal in size, and hardware-friendly. Considering these factors, we implemented MiDaSNet, which works on the simplest setup of monocular vision and utilizes hardware-friendly convolutional neural network-based architecture, for real-time depth estimation on the edge. Furthermore, since the model is trained on diverse datasets, it shows stable performance across different mediums. For edge implementation, we quantized the model weights down to an 8-bit fixed-point representation and deployed the quantized model on an inexpensive field-programmable gate array card, Kria KV260, utilizing predefined deep learning processing units embedded in the programmable logic. The results demonstrate that our quantized model achieves 82.6% zero-shot δ1 accuracy on the NYUv2 test dataset, with an inference speed of 50.7 FPS on the card.

DOI

10.55730/1300-0632.4197

Keywords

MiDaSNet, real-time, depth estimation, indoor robotics, field-programmable gate array, FPGA

First Page

682

Last Page

698

Publisher

The Scientific and Technological Research Council of Türkiye (TÜBİTAK)

Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

Download

Share

COinS