Sistem Deteksi Asap Pada Smart Building Berbasis Internet Of Things Dan Convolutional Neural Network

Muhammad, Intro Brilliant (2024) Sistem Deteksi Asap Pada Smart Building Berbasis Internet Of Things Dan Convolutional Neural Network. Other thesis, institut teknologi sepuluh nopember.

[thumbnail of 0721940000062 - Undergraduate_Thesis.pdf] Text
0721940000062 - Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 2 February 2026.

Download (26MB) | Request a copy

Abstract

Deteksi asap dini dapat memberikan informasi yang krusial terutama untuk menghindari kebakaran. Namun, tingginya biaya sistem deteksi asap yang disediakan vendor, serta kapabilitas deteksi asap dengan sensor yang kurang akurat dan memiliki jangkauan deteksi yang sempit, membuat adanya celah keamanan. Selain itu, merokok atau vaping di lingkungan kampus dilarang keras dan dikenakan sanksi. Untuk mengatasi masalah tersebut, Tugas Akhir ini mengusulkan pembuatan perangkat yang dapat digunakan untuk mendeteksi asap di Smart Building terutama di Tower 2 ITS menggunakan CNN dengan arsitektur MobileNetSSD-V2 sebagai pendeteksi citra asap dan Internet of Things. Sistem tersebut mendeteksi asap melalui video menggunakan komponen Single Based Computer Raspberry Pi 4 Model B dan kamera. Tangkapan asap yang terdeteksi ditampilkan ke emph Dashboard menggunakan MQTT sehingga pihak berwenang dapat menindaklanjutinya. Dilakukan penelitian untuk membandingkan arsitektur CNN yang memiliki performa tinggi dengan beban komputasi yang sesuai dengan komponen SBC. Sistem yang dikembangkan memiliki potensi untuk memberikan solusi yang terjangkau, lebih presisi, dan akurat untuk deteksi asap. Hasil yang dari tugas akhir adalah sebuah perangkat yang dapat digunakan untuk mendeteksi asap dengan skor mAP 94.10% , citra asap yang tergangkap akan dikirimkan ke server dan ditampilkan ke web mengenai adanya objek tersebut
=======================================================================================================================================
Early smoke detection can provide crucial information, especially to avoid fires. However, the high cost of smoke detection systems provided by vendors, as well as smoke detection capabilities with sensors that are less accurate and have a narrow detection range, create security gaps. In addition, smoking or vaping on campus is strictly prohibited and subject to sanctions. To overcome this problem, this Final Project proposes the creation of a device that can be used to detect smoke in Smart Buildings, especially in ITS Tower 2 using a CNN with MobileNetSSD-V2 architecture as a smoke image detector and the Internet of Things. The system detects smoke via video using Single Based Computer Raspberry Pi 4 Model B components and a camera. Detected smoke captures are displayed to Dashboard using MQTT so that authorities can act on them. Research was carried out to compare CNN architectures that have high performance with computational loads that match the SBC components. The developed system has the potential to provide an affordable, more precise and accurate solution for smoke detection. The result of the final project is a device that can be used to detect smoke with a score of mAP 94.10\% , the captured smoke image will be sent to the server and displayed on the web regarding the presence of the object.

Item Type: Thesis (Other)
Uncontrolled Keywords: Convolutional Neural Network, Deteksi Asap, Internet of Things, Kamera, Raspberry Pi, Smart Building
Subjects: Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines.
Q Science > QA Mathematics > QA76.6 Computer programming.
Q Science > QA Mathematics > QA76.758 Software engineering
Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5103.2 Wireless communication systems. Two way wireless communication
T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5105.546 Computer algorithms
T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7882.P3 Pattern recognition systems
Divisions: Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Computer Engineering > 90243-(S1) Undergraduate Thesis
Depositing User: Intro Brilliant Muhammad
Date Deposited: 12 Feb 2024 02:27
Last Modified: 12 Feb 2024 02:27
URI: http://repository.its.ac.id/id/eprint/106883

Actions (login required)

View Item View Item