Klasifikasi Kualitas Daging Kerang Darah (Anadara Granosa) Menggunakan Sensor Gas Berbasis Machine Learning

Rahman, Thariq Irfan (2026) Klasifikasi Kualitas Daging Kerang Darah (Anadara Granosa) Menggunakan Sensor Gas Berbasis Machine Learning. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5022221050-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (7MB) | Request a copy

Abstract

Kerang darah (Anadara granosa) merupakan komoditas hasil laut yang rentan mengalami penurunan kualitas akibat proses biokimia dan aktivitas mikroorganisme. Proses pembusukan ini menghasilkan berbagai emisi senyawa volatil seperti amonia (NH₃), hidrogen sulfida (H₂S), dan volatile organic compounds (VOC). Selama ini, penilaian kesegaran kerang darah yang dilakukan secara organoleptik cenderung subjektif dan kurang konsisten. Penelitian ini bertujuan untuk merancang sistem electronic nose (e-nose) menggunakan deret sensor gas tipe Metal Oxide Semiconductor untuk memantau penurunan mutu kerang darah secara objektif. Sinyal tegangan dari respons sensor diolah melalui tahapan pra-pemrosesan yang mencakup ekstraksi fitur nilai rata-rata tiap 10 detik dalam 50 detik terakhir kondisi steady-state serta nilai differensial antara nilai tegangan ketika terpapar aroma sampel terhadap nilai tegangan ketika udara bersih (baseline) yang kemudian data diskalakan ke rentang 0 hingga 1 menggunakan MaxAbs Scaler. Rentang ini terbentuk karena seluruh nilai data awal bersifat non-negatif. Data tersebut diklasifikasikan ke dalam tiga kelas kesegaran yaitu segar, sedang, busuk menggunakan algoritma Artificial Neural Network (ANN) dan Support Vector Machine (SVM). Hasil evaluasi komputasi menunjukkan bahwa model ANN dengan konfigurasi 7 neuron pada hidden layer pertama dan 2 neuron pada hidden layer kedua dengan akurasi pengujian sebesar 92% serta model SVM dengan kernel rbf berhasil mencapai tingkat akurasi pengujian sebesar 90%. Lebih lanjut, implementasi algoritma klasifikasi secara langsung pada mikrokontroler Arduino Mega 2560 terbukti mampu memprediksi 30 sampel uji di lapangan secara akurat dengan persentase keberhasilan 100%.
===================================================================================================================================
Blood shellfish (Anadara granosa) is a seafood commodity that is prone to deterioration in quality due to biochemical processes and microorganism activities. This decay process produces various emissions of volatile compounds such as ammonia (NH₃), hydrogen sulfide (H₂S), and volatile organic compounds (VOCs). So far, the assessment of the freshness of blood shells carried out organoleptically tends to be subjective and less consistent. This study aims to design an electronic nose (e-nose) system using a series of Metal Oxide Semiconductor type gas sensors to objectively monitor the decline in the quality of blood shells. The voltage signal from the sensor response is processed through a pre-processing stage that includes the extraction of the average value feature every 10 seconds in the last 50 seconds of steady-state conditions as well as the differential value between the voltage value when exposed to the sample aroma to the voltage value when the air is clean (baseline) which is then scaled to the range of 0 to 1 using the MaxAbs Scaler. This range is formed because all the initial data values are non-negative. The data is classified into three classes of freshness, namely fresh, medium, rotten using the Artificial Neural Network (ANN) and Support Vector Machine (SVM) algorithms. The results of the computational evaluation showed that the ANN model with a configuration of 7 neurons in the first hidden layer and 2 neurons in the second hidden layer with a test accuracy of 92% and the SVM model with the rbf kernel managed to achieve a test accuracy level of 90%. Furthermore, the implementation of the classification algorithm directly on the Arduino Mega 2560 microcontroller was proven to be able to accurately predict 30 test samples in the field with a 100% success percentage.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	electronic nose, kerang darah, kualitas pangan, ANN, SVM, electronic nose, blood clamps, food quality, ANN, SVM
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) T Technology > TA Engineering (General). Civil engineering (General) > TA1573 Detectors. Sensors
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20201-(S1) Undergraduate Thesis
Depositing User:	Thariq Irfan Rahman
Date Deposited:	21 Jul 2026 06:00
Last Modified:	21 Jul 2026 06:00
URI:	http://repository.its.ac.id/id/eprint/135873

Actions (login required)

View Item