IoT: Building a Raspberry Pi security system with facial recognition
Gsponer, David (2018)
Avaa tiedosto
Lataukset:
Haaga-Helia ammattikorkeakoulu
2018
Creative Commons Attribution-NonCommercial-ShareAlike 2.5 Switzerland
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2018110416616
https://urn.fi/URN:NBN:fi:amk-2018110416616
Tiivistelmä
As IoT grows, devices become more cheaply available and facial recognition develops rapidly, the need for a system using these three components arises. Combining these three technologies will allow tackling a diversity of use cases, from replacing door locks to smart CCTV systems.
Some products which are making use of these components combined exist, but none have been implemented on a low-budget system under 100 euros. For this reason, the aim of the thesis is to work only with open-source software, no webservice, and only with the cheapest hardware available and the question whether it is possible to build such a system with the given hardware.
The following thesis will first focus on facial recognition and IoT in a theoretical framework, which is then followed by its implementation throughout the project’s five phases and a
final presentation of the product.
The result of the presented project is a working system using facial recognition with OpenCV, IoT over the MQTT protocol and Clients on the Raspberry Pi as well as on an Android mobile application. The thesis builds a foundation using these three components on a system applicable in many further scenarios of facial recognition in combination with IoT on low-budget hardware.
Some products which are making use of these components combined exist, but none have been implemented on a low-budget system under 100 euros. For this reason, the aim of the thesis is to work only with open-source software, no webservice, and only with the cheapest hardware available and the question whether it is possible to build such a system with the given hardware.
The following thesis will first focus on facial recognition and IoT in a theoretical framework, which is then followed by its implementation throughout the project’s five phases and a
final presentation of the product.
The result of the presented project is a working system using facial recognition with OpenCV, IoT over the MQTT protocol and Clients on the Raspberry Pi as well as on an Android mobile application. The thesis builds a foundation using these three components on a system applicable in many further scenarios of facial recognition in combination with IoT on low-budget hardware.