Primjena tehnika dubokog učenja u okruženju pametnih zgrada u sustavima interneta stvari

master's thesis

• Title (english): Deep Learning Applications in IoT Environments for Smart Building Automation
• Author: Jure Knezović
• Mentor: Josip Knezović (mentor)
• Committee member: Josip Knezović (predsjednik povjerenstva)
• Committee member: Igor Čavrak (član povjerenstva)
• Committee member: Ivana Bosnić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2019-07-03, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

U ovom projektu su povezane tehnike dubokog učenja sa sustavima interneta stvari tako da udaljeni modul koji detektira pokret dojavljuje poslužitelju da se dogodio pokret, te poslužitelj u ovisnosti o periodu dana i dostupnosti kamere na odgovarajućem modulu odlučuje želi li dohvatiti sliku s kamere i provesti prepoznavanje osobe, nakon kojeg može dopustiti, odnosno zabraniti pristup osobi.
Osim kontrole pristupa putem prepoznavanja osobe, u projektu su izvedene i kontrola pristupa korištenjem senzora pokreta, koja u određenom periodu dana dozvoljava pristup, kontrola pristupa korištenjem NFC tehnologije putem aplikacije za pametne telefone s operacijskim sustavom Android, te kontrola pristupa putem Web sučelja.
Za ispunjenje svih zahtjeva projekta korištena je mreža mikrokontrolera ESP32.
Jezgra projekta se nalazi na poslužitelju te uključuje prepoznavanje lica, API i web sučelje s podrškom za administraciju i komunikaciju s udaljenim modulima.

Abstract (english)

In this project, deep learning techniques are connected with the Internet of Things, in a way that the remote module that detects movement notifies the server about movement, and the server, depending on the period of the day and the availability of the camera on the remote module, decides whether to retrieve a picture from the camera and perform a person's identification using face recognition, after which it may or may not allow access to a person which triggered motion sensor.
In addition to access control using face recognition, this project also implements so called unattended access control using the motion sensor, which allows access during a certain period of the day, access control using NFC technology via smartphone application for the Android operating system, and access control via the Web interface.
To fulfill all project requirements, a network of remote modules which use ESP32--based microcontroller is used. The core of the project is located on the server, which includes face recognition, API and web interface with support for administration and communication with remote modules.

• Keywords: duboko učenje; kućna automatizacija; internet stvari; kontrola pristupa
• Keywords (english): deep learning; home automation; Internet of Things; access control
• Language: english
• URN:NBN: urn:nbn:hr:168:067755
• 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: 2020-01-07 19:07:22