/RealSense_T265_evaluator

Python code to evaluate Intel RealSense T265 tracking camera.

Primary LanguagePython

RealSense_T265_evaluator

In this work we tested the Intel® RealSense™ Tracking Camera T265 functionality and evaluate data acquisition reliability in different settings. Particularly, the aim is provide a clear and simple report that can be used for future studies or experiments. The experiments were done indoors and outdoors and were classified as follows:

  • Short range experiments (1m)
  • Medium range experiments (5-15m)
  • Long range experiments (50-100m)
  • Very long range experiments (> 100m)
  • Rotation experiments

output

Getting started

To get a local copy up and running follow these simple steps.

Prerequisites

Hardware

Software

Installation

  1. Follow these instructions for the Raspberry Pi set-up
  2. Clone and import the repo in Raspberry Pi
  3. Open the Test folder and choose the right file to do your experiments (if you have an LCD touchscreen to connect to the Raspberry, you can use the GUI to manage the files, but remember to install PyQt 5.15.1)

Results

The following table shows a short description of our results, for more details see the report or presentation.

For all experiments it is advisable to wait for a camera stabilization time of 5-10 seconds

Short Range Medium Range Long Range Very Long Range Rotation
  • indoor: unstable performance, huge changes at minimal setup variation
  • outdoor: generally more stable and accurate track
  • path detected very well, very distant references allow the camera to perform better tracking
  • visible swaying of the steps
  • the extended environment improves tracking compared to previous cases
  • start and end of tracking match
  • path found, but start and end of tracking do not match
  • good sensitivity to speed and acceleration changes
  • precise pitch and roll rotation
  • a good yaw measurement requires higher rotation speed

In conclusion, the results obtained by the camera are highly dependent on the environment used. A broader view brings improvement.