Praćenje položaja ljudskog kostura korištenjem fuzije više senzora

undergraduate thesis

• Title (english): Human skeleton pose tracking using multiple sensor fusion
• Author: Jelena Marić
• Mentor: Ivan Marković (mentor)
• Committee member: Ivan Marković (predsjednik povjerenstva)
• Committee member: Filip Marić (član povjerenstva)
• Committee member: Juraj Peršić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2020-09-02, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Cilj ovog rada bio je postojećim estimatorom stanja izvršiti procjenu položaja ljudskog kostura koji je dobiven mjerenjima četiriju Kinect senzora. Odabrani estimator stanja, Kalmanov filtar, radi po principu koraka predviđanja stanja i korekcije tog predviđanja. Prije same procjene, mjerenja koja su na raspolaganju grupirana su po trenucima u kojima su dobivena te su spojena u otežanu uzoračku srednju vrijednost. Pri fuziji mjerenja trebalo je obratiti pažnju na trenutke koji nedostaju u trajanju jedne sekvence kretanja i na zrcaljenje koordinatnih sustava senzora kojima je osoba okrenuta leđima. Za analizu uspješnosti Kalmanova filtra napravljena je i procjena stanja jednostavnom aritmetičkom sredinom svih mjerenja. Ta metoda često rezultira položajem kostura koji nije takav u stvarnosti dok predloženi estimator stanja postiže bolje rezultate prikazujući dijelove kostura u realističnim omjerima.

Abstract (english)

The aim of the presented work was to estimate the position of person's skeleton gathered from measurements obtained by four Kinect sensors using an existing estimator. The chosen estimator, Kalman filter, works in a two stepped process that includes prediction and update of the prediction. Before the estimation, available measurements are grouped together considering the moment in which the measurements were acquired and are fused together with the weighted measurement fusion. Missing frames as well as flipping left and right sides (if a person is opposite od facing the sensor) had to be taken into account when fusing these measurements. Estimation was also provided using simple average method. This method often generates results that represent an unnatural pose of the tracked skeleton while the proposed estimator on the other hand accomplishes better results regarding bones ratios

• Keywords: detekcija kostura; Kinect senzori; fuzija mjerenja; procjena stanja; Kalmanov filtar
• Keywords (english): skeleton tracking; Kinect sensors; measurement fusion; state estimation; Kalman filter
• Language: croatian
• URN:NBN: urn:nbn:hr:168:600442
• Study programme: Title: Computing; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-01-14 09:24:07