Automatska detekcija izvora svjetlosti u slici

• Main Author: Orioli, Adriano
• Other Authors: Pandžić, Igor Sunday
• Format: Bachelor Thesis
• Language: Croatian
• Published: Sveučilište u Zagrebu. Fakultet elektrotehnike i računarstva. 2016
• Subjects: svjetlo; svjetlina; osvjetljenost; sjena; automatska detekcija; rust; light; luminance; illumination; shadow; automatic detection; TEHNIČKE ZNANOSTI. Računarstvo; TECHNICAL SCIENCES. Computing; sami
• Online Access: https://zir.nsk.hr/islandora/object/fer:1684; https://urn.nsk.hr/urn:nbn:hr:168:134775; https://repozitorij.unizg.hr/islandora/object/fer:1684; https://repozitorij.unizg.hr/islandora/object/fer:1684/datastream/PDF

Uz napredak tehnologije, sve vec ́i broj korisnika je okružen ured ̄ajima koji imaju moguc ́nost fotografiranja i snimanja, te moguc ́nost obrade tih zapisa. Od mobilnih ured ̄aja do stolnih racˇunala, vec ́ina ih dolazi opremljena ured ̄ajima za snimanje slike i zvuka. Te moguc ́nosti se danas koriste, med ̄u ostalim, za marketinške, edukativne i zabavne svrhe u obliku aplikacija koje proširuju stvarnost koju korisnik vidi i umec ́u virtualne elemente poput graficˇkih sucˇelja i 3D modela. U ovom radu se analiziraju tre- nutacˇno dostupne metode razvijene u svrhu automatskog detektiranja svjetla na slici. Detektiranje svjetla, i shodno tome, kreiranje virtualnog osvjetljenja za elemente pro- širene stvarnosti bi povec ́alo kvalitetu aplikacija koje se predstavljaju korisnicima, i samim time i vrijednost koju oni sami dobivaju od njih. With the current advancements in technology, an ever increasing number of users is surrounded by devices with video and photo taking capabilities. From mobile de- vices to desktop computers, most come equipped with some sort of camera, and a processor powerful enough to process the images from that camera in realtime. Those capabilities are being used today, among other things, for marketing, educational and entertainment purposes in the form of applications for augmented reality which aug- ment the perception of the users with graphical interfaces and 3D models. This paper will analyse the currently available algorithms for automatic light source detection in images. Light source detection, combined with the creation of 3D models illuminated by that light source, can serve a great purpose in the enhancement of augmented reality applications, the increase of user immersion, and consecutively the improvement of the user experience.