Automatic localization of military tactical plans on the digital map

Organisatsiooni nimi
Collaboration project with Estonian National Defence Academy
Kokkuvõte
Military commanders of different ranks produce tactical plans consisting of NATO standardized symbols, which describe the orders of different units in space and time. These plans are drawn on a plastic film, which is placed on the map and aligned with the map via drawing a section of the underlying map grid (N-S and W-E lines) on the film, as well as numeric information about which map lines the lines on the film correspond to.

In collaboration with the Estonian National Defence Academy, we are working on automatizing the reading of information from these films directly to digital form. This means automatic detection of the symbols on the plan, but crucially also localizing the whole film in space (on the world map). Based on a photo taken of such film, for example by phone (could consider scanner as simplifying assumption first), this thesis should localize the map grid on the film, detect the lines it is made of and their intersection points, as well as (if time allows) the handwritten numbers. Based on the angles of the detected grid lines and/or intersection points, the solution should correct for the photo not being taken from a correct angle (perspective correction). In all, the solution should be able to localize the film in space and overlay its contents on a map at the correct location (without detecting individual objects on it).

At the completion of this initial goal, automatic detection of the handwritten information can be developed. Also, a standard operating procedure for taking photos of such films (limits on the angle) can be defined as a result of the work.

Team: Ardi Tampuu (Lecturer of AI), Mihkel Lepson (MSc student), Karl-Kristjan Kõverik (MSc student) + you
Timeline: the entire solution’s prototype will be tested in April 2023. This will set a perfect deadline for the code to be completed, so you have time to write the thesis in May.

Remuneration:
The student working on this thesis will receive a scholarship. Approximate sum 300eur/month in autumn semester, 600-900 eur/month in spring semester. 2023 scholarships are is dependent on the budget allocated for the project (in negotiations).
Lõputöö kaitsmise aasta
2022-2023
Juhendaja
ardi tampuu
Suhtlemiskeel(ed)
eesti keel, inglise keel
Nõuded kandideerijale
Has completed data science or coputer vision or computer graphics courses
Tase
Magister
Märksõnad

Kandideerimise kontakt

 
Nimi
ardi tampuu
Tel
E-mail
ardi.tampuu@ut.ee