Razvoj sustava za detekciju i praćenje pikado igre primjenom vizijskih tehnologija

master's thesis

• Title (english): Development of a dart game detection and tracking system using vision technologies
• Author: Jurica Cvetić
• Mentor: Filip Šuligoj (mentor)
• Granter: University of Zagreb; Faculty of Mechanical Engineering and Naval Architecture; Zagreb
• Defense date and country: 2024-05-16, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Mechanical Engineering; General Mechanical Engineering (Construction)

Abstract

U današnjem je svijetu naglasak na konstantan napredak različitih segmenata društva. Napredak društva uvjetuje manjak dostupnog slobodnog vremena koje je potrebno svakom pojedincu radi odmora od napornog tempa života. Shodno tome, potreba za poboljšanjem te optimiranjem rekreacijskih aktivnosti sve više raste. Jedan od načina optimiranja je mogućnost automatizacije aktivnosti kako bi se uštedjelo na vremenu i omogućila bezbrižnost pojedinaca. U tom je kontekstu razvijen vizijski sustav za jedan od popularnijih sportova današnjice – pikado. Cilj rada je korištenje vizijskog sustava za automatsko bodovanje bačene strelice. Za potrebe rada vizijskog sustava razvijen je fizički postav koji omogućuje stabilnost i robusnost u radu algoritma. Fizički postav sadrži 3 kamere koje omogućuju razne kadrove pikado mete, prikladno osvjetljenje koje osigurava konstantne uvjete važne za rad algoritma te Jetson Nano računalo. Kako bi algoritam mogao ispravno funkcionirati, izvršena je intrinzična kalibracija kamera uz obuhvaćanje distorzijskih koeficijenata. Zatim je razvijen algoritam koji omogućuje izračun matrica homografija za svaku kameru. Tako je osigurano pravilno pronalaženje pozicije vrha strelice te prikladno bodovanje s obzirom na pronađenu poziciju. Za bodovanje je razvijen algoritam koji u polarnim koordinatama definira udaljenost i kut točke za precizno određivanje bodova. Kako bi prethodni algoritmi funkcionirali, razvijen je ključan algoritam pronalaska vrha na temelju detekcije i obrade konture strelice u kadru kamera. U konačnici je napisan algoritam koji objedinjuje sve procese i omogućava paralelan rad triju kamera koje glasovanjem odlučuju o postignutom rezultatu. Na kraju je izvršena evaluacija rezultata koja daje pojam o točnosti i nedostacima razvijenog sustava.

Abstract (english)

Today's world requires the constant advancement of various segments of society. This progress leads to a lack of leisure time, which is necessary for each individual's mental health. Hence, the need for recreational activity optimization keeps rising. One way of optimizing is by using automation techniques in order to save time and enable the careless engagement of individuals during these activities. Hence, a vision system has been developed for a very popular sport - darts. The goal of this master thesis is to use a vision system for automatic determination of scored points. In order to achieve that, a physical setup has been developed to provide stability and robustness of the algorithm. The physical setup consists of three cameras placed around the dartboard, offering different fields of view. Suitable lighting has also been secured in order to ensure constant conditions. Finally, Jetson Nano provides the necessary integration of all process participants. The next key step was the intrinsic camera calibration in order to ensure correct algorithmic operation. An algorithm for each camera's homography matrix has been developed in order to precisely detect the dart's tip position. Furthermore, an algorithm for scoring that uses polar coordinates to describe the point position has been developed. Finally, an algorithm for detecting the dart's tip, which consists of different image processing techniques, has also been developed. In order to enable parallel execution of the mentioned algorithms, a main algorithm using multiprocessing techniques integrates all algorithms and offers real-time score tracking. Finally, an evaluation of these algorithms has been conducted to assess the quality and drawbacks of this vision system.

• Keywords: vizijski sustav; intrinzična kalibracija; matrica homografije; ispravljanje perspektive kamere; detekcija konture; obrada slike; paralelno procesiranje
• Keywords (english): vision system; intrinsic calibration; homography matrix; camera perspective correction; contour detection; image processing; multiprocessing
• Language: croatian
• URN:NBN: urn:nbn:hr:235:698165
• Study programme: Title: Mechanical Engineering; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka strojarstva (magistar/magistra inženjer/inženjerka strojarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2024-05-09 09:46:17