Android aplikacija za detekciju 2D-kodova

master's thesis

• Title (english): Android application for detection of 2D barcodes
• Author: Jurica Vidak
• Mentor: Hrvoje Mlinarić (mentor)
• Committee member: Hrvoje Mlinarić (predsjednik povjerenstva)
• Committee member: Igor Čavrak (član povjerenstva)
• Committee member: Josip Knezović (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2021-07-15, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Google ML Kit i ZXing su dvije tehnologije koje nude sličnu funkcionalnost (što se tiče barkod očitanja) ali uz potpuno različit način rada. ML Kit je alat i programski okvir koji radi na principu strojnog učenja i zahtjeva nešto jače sklopovlje za efikasniji rad od jednostavnije ZXing programske biblioteke koja radi na principu usporedbe očitanih najtamnijih piksela sa kamera sa geometrijskim oblikom sličnim barkodovima.
U izradi aplikacije za skeniranje QR kodova korištene su obje tehnologije te je vršena usporedba očitavanje raznih QR kodova (višestrukih kodova, skeniranih pod raznim kutovima, djelomično nepotpuni, itd.). Aplikacija je izrađena u Android studiju i pisana u Java programskom jeziku ,a za samu analizu korišten je mobilni uređaj Samsung Galaxy S8.
Aplikacija je jednostavna sa početnim sučeljem koje nudi na izbor tehnologiju skeniranja i gdje odabirom skeniranja se aktivira prikaz kamere gdje u stvarnom vremenu odabrana metoda u obrađuje kadrove kamere te pokušava pronaći i dekodirati QR kod, a istovremeno u gornjem desnom kutu imamo brojač obrađenih kadrova u sekundi preko čega možemo pratiti performanse aplikacije. Tijekom skeniranja kodova ML Kit se pokazao puno brži naročito u situacijama gdje skeniramo višestruke kodove i pod raznim kutovima, a naročito u situacijama gdje su kodovi djelomično sakriveni.
Analizom rada u prosjeku je Google ML Kit implementacija se pokazala do 1,5 puta brža kod očitanja pojedinačnih QR kod pod pravilnim kutom snimanja te do 2,1 puta brža u prosjeku. Kod očitanja višestrukih kodova ML Kit uspijeva očitati puno više kodova ,a kod očitanja kodova snimanih nepravilnim kutovima kamere i djelomično vidljivih ZXing nije uspijevao uopće očitati niti prepoznati QR kodove dok se ML Kit ovdje se pokazao puno uspješniji sa preko 70% očitanih kodova.

Abstract (english)

Google ML Kit and ZXing are two technologies that offer similar functionality (in terms of barcode readings) but with a completely different way of working. The ML Kit is a machine learning tool and software framework that requires slightly stronger hardware for more efficient operation than the simpler ZXing software library, which works on the principle of comparing the darkest pixels read from cameras with geometric shapes similar to barcodes.
Both technologies were used in the development of the application for scanning QR codes, and a comparison of readings of various QR codes (multiple codes, scanned at various angles, partially incomplete, etc.) was performed. The application was made in the Android studio and written in the Java programming language, and the Samsung Galaxy S8 mobile device was used for the analysis.
The application is simple with an initial interface that offers a choice of scanning technology and where selecting a scan activates the camera display where the selected method processes the camera footage in real time and tries to find and decode the QR code. through which we can monitor the performance of the application. During code scanning, the ML Kit proved to be much faster, especially in situations where we scan multiple codes and at various angles, and especially in situations where the codes are partially hidden.
By analyzing the work on average, the Google ML Kit implementation proved to be up to 1.5 times faster when reading individual QR codes at the correct shooting angle and up to 2.1 times faster on average. When reading multiple codes, the ML Kit manages to read many more codes, and when reading codes recorded at irregular camera angles and partially visible, ZXing failed to read or recognize QR codes at all, while the ML Kit proved to be much more successful with over 70% of read codes.

• Keywords: ZXing; Google ML Kit; Android; Android studio; Detekcija 2D kodova; Detekcija QR kodova; mobilna aplikacija
• Keywords (english): ZXing; Google ML Kit; Android; Android studio; 2D code detection; QR code detection; mobile application
• Language: croatian
• URN:NBN: urn:nbn:hr:168:444249
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-01-15 11:24:35