Building a dataset for vision-based localization in orienteering - project at Chalmers/TRACKS

Orienteering project at Chalmers/TRACKS!

Project in the course Digitalization in Sports - TRA300 – 2024/25
Study period 1-2 (autumn) 2024
CHALMERS University of Technology

Read more about TRACKS.

Status: 🟢 open for application - ❗ deadline May 14th

Last update: 2024-05-03

Apply Contact

Example of a screenshot from an headcam during an orienteering race

Background and scope

Localization is a common task in many fields, including robotics, computer vision, and self-driving cars. For the competitors of an orienteering race, localization is the problem to determine their own position on the map, without using any other instrument than a compass and the map itself. Orienteers can easily solve this problem at the start and at the control points, because they are marked both on the terrain and on the map. In between, however, it is a more difficult task: orienteers use a variety of techniques and skills to localize themselves, but they mainly rely on the visual comparison between the terrain and the map.

Purpose of the work

The aim of this project is to build a relevant dataset for vision-based localization in forests with orienteering maps. The dataset will consist of a set of runs in the forest. Each run will include: a video recorded with a head-cam camera (e.g., GoPro), a GNSS track synchronized with the video, and an orienteering map of the area (as geo-referenced image). In addition to planning and collecting the data, another activity is the development of a web application for accessing the dataset. The minimum set of features of the application is: (i) browsing the dataset, (ii) displaying the map and playing the video of the runs, and (iii) showing/hiding the track on the map.

The development of the dataset will serve as basis for the future development of AI algorithms for vision-based localization in the forest. Moreover, the dataset can be used for developing applications/games for orienteering training: for example, trainees can be asked to watch a video of a run and their task is to determine the location of the camera on the map at the end of the video, knowing the initial position and direction.

Additional features can be developed and will be discussed with the students.

Activities included in the project

  1. Defining the project
  2. Selecting appropriate areas and length for the runs
  3. Design the data collection protocol
  4. Design the database for storing the data
  5. Collecting the data
  6. Developing the web application for accessing the data
  7. Testing the web application

Profile of the students

Core competencies needed (each student is expected to have at least one of the following skills, the team should cover all of them): experimental planning, orienteering, course setting in orienteering, multi-media database design, software/web development.

FAQ

  1. Who can apply for this project?
    Chalmers students (including exchange students), alumni, PhD students, and professionals.

  2. How can I apply?
    By following the instructions on the course page. Please note that when you apply for the course, you have to attach your transcripts, CV, and a short motivation letter. In the motivation letter write that you are interested in this project because there are also other projects in the course.

  3. Do I get credits for this project?
    Yes, the course TRA300 - Digitalization in sports is a 7.5hp course. There is also the possibility to extend the project to the 15hp course TRA375 - Digitalization in sports - 15hp, which includes more project work.This is a possibility to be discussed with me and the examiner of the course, Dan Kuylenstierna.

  4. Is the project the entire course?
    The project work corresponds to 60% of the TRA300 course (equivalent to about three study weeks per student). In addition: 30% from quizzes about lectures (~8 lectures at the beginning of study period 1); 10% presentation skills.

  5. How many people can participate?
    The projects in TRA300/TRA375 are typically carried out in groups of 2-6 students but I'm flexible to accommodate more students if needed.

Contact

Supervisor: Marco Della Vedova, Chalmers - marco.dellavedova@chalmers.se

Please, email the supervisor if you are interested in this project.