Implementacija sustava za optičko prepoznavanje znamenki u ugradbenom računalu

master's thesis

• Title (english): Embedded System Implementation for Optical Digit Recognition
• Author: David Romić
• Mentor: Hrvoje Džapo (mentor)
• Committee member: Hrvoje Džapo (predsjednik povjerenstva)
• Committee member: Željka Lučev Vasić (član povjerenstva)
• Committee member: Ana Sović Kržić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2018-07-17, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Electrical Engineering

Abstract

U ovom radu prikazani su osnovni postupci predobrade slike i korištenja metoda strojnog učenja za prepoznavanje znamenaka s brojila na ugradbenim računalnim sustavima. Istražene su mogućnosti korištenja mikrokontrolerskih sustava (STM32F4) i ugradbenih računalnih sustava temeljenih na aplikacijskim procesorima i operacijskom sustavu Linux (Raspberry Pi 2). Uspoređene su mogućnosti programskih biblioteka FANN i Tensorflow. Implementirano je rješenje prikladno za ugradbeni računalni sustav temeljen na aplikacijskom procesoru i korištenju operacijskog sustava Linux, a rješenje se bazira na CNN mreži treniranoj korištenjem Tensorflow biblioteke.

Abstract (english)

This paper elaborates image preprocessing and machine learning algorithms for digits recognition on various types of meters by means of embedded systems. The research was conducted for microcontroller based systems (STM32F4) and embedded systems based on application processors and Linux operating system (Raspberry Pi). The features and capabilities of FANN and Tensorflow libraries were explored. The implemented solution is applicable for embedded systems based on application processors and Linux operating system. The solution is based on CNN trained by using Tensorflow library.

• Keywords: prepoznavanje znamenaka; obrada slike; ugradbeni računalni sustavi
• Keywords (english): digit recognition; image processing; embedded systems
• Language: croatian
• URN:NBN: urn:nbn:hr:168:688799
• Study programme: Title: Electrical Engineering and Information Technology; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka elektrotehnike i informacijske tehnologije (magistar/magistra inženjer/inženjerka elektrotehnike i informacijske tehnologije)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-03-14 00:30:34