Sistem Monitoring Analog Gauge Otomatis Pada Tangki Nitrogen Menggunakan Metode Convolutional Neural Network

Athaya, Muhammad Farras Nabil (2025) Sistem Monitoring Analog Gauge Otomatis Pada Tangki Nitrogen Menggunakan Metode Convolutional Neural Network. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2040211007-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

Pengukuran dan pemantauan parameter fisik dalam industri seperti laju aliran, tekanan, dan suhu, sering kali masih mengandalkan analog gauge yang dibaca secara manual oleh operator. Metode konvensional ini memiliki keterbatasan dalam efisiensi dan efektivitas, terutama dalam lingkungan ekstrem atau lokasi yang sulit dijangkau. Penelitian ini bertujuan untuk mengembangkan sistem pemantauan analog gauge otomatis menggunakan metode Convolutional Neural Network (CNN) dengan algoritma You Only Look Once (YOLO). Sistem ini dirancang untuk menangkap gambar analog gauge menggunakan ESP32-CAM dan diproses melalui Raspberry Pi menggunakan metode image processing seperti CLAHE dan sharpness untuk meningkatkan akurasi deteksi jarum dan skala pengukuran. Implementasi algoritma YOLO memungkinkan sistem untuk mendeteksi serta membaca nilai dari analog gauge secara real-time, sehingga dapat meningkatkan efisiensi dalam pemantauan dan pengambilan keputusan berbasis data. Penelitian ini diharapkan dapat memberikan kontribusi dalam otomatisasi pemantauan alat ukur di industri, khususnya pada tangki nitrogen di PT. Flextronics Technology Indonesia, guna meningkatkan akurasi, keandalan, dan efisiensi proses produksi. Hasil dari pengujian sistem deteksi titik tengah dan titik jarum analog gauge dengan 92 sampel gambar mendapatkan score precision 98%; score akurasi 97%, F1-Score 97% dan pengujian pembacaan jarum pada analog gauge dengan 20 gambar menghasilkan nilai MAE 0,163. Sehingga sistem pembacaan analog gauge dapat digunakan untuk membantu memantau nilai parameter pada analog gauge. ====================================================================================================================================
The measurement and monitoring of physical parameters in industries, such as flow rate, pressure, and temperature, often still rely on analog gauges that require manual reading by operators. This conventional method has limitations in efficiency and effectiveness, particularly in extreme environments or hard-to-reach locations. This research aims to develop an automatic analog gauge monitoring system using Convolutional Neural Network(CNN) method with You Only Look Once (YOLO) algorithm. The system captures images of analog gauges using an ESP32-CAM and processes them through a Raspberry Pi using image processing techniques such as CLAHE and sharpness to improve needle detection accuracy and measurement scale. The identification process involves pre-processing stages such as grayscaling, noise reduction, and edge detection to enhance the accuracy of needle and scale detection. The implementation of the YOLO algorithm enables the system to detect and read values from analog gauges in real time, improving monitoring efficiency and data-driven decision-making. This research is expected to contribute to the automation of measurement monitoring in industries, particularly for nitrogen tanks at PT Flextronics Technology Indonesia, enhancing accuracy, reliability, and production efficiency. The results of testing the analog gauge center and needle point detection system with 92 image samples get a precision score of 98%; accuracy score of 97%, F1-Score 97% and testing the needle reading on the analog gauge with 20 images produces an MAE value of 0.163. So that the analog gauge reading system can be used to help monitor the value of the parameter on the analog
gauge.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Analog Gauge, Convolutional Neural Network, You Only Look Once Analog Gauge, Convolutional Neural Network, You Only Look Once
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > QA Mathematics > QA336 Artificial Intelligence Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Divisions:	Faculty of Vocational > 36304-Automation Electronic Engineering
Depositing User:	Muhammad Farras Nabil Athaya
Date Deposited:	14 Aug 2025 07:54
Last Modified:	14 Aug 2025 07:54
URI:	http://repository.its.ac.id/id/eprint/128108

Actions (login required)

View Item