Pendugaan kandungan klorofil pada daun kelor (moringa oleifera.) Berbasis image analysis dengan menggunakan metode artificial neural network

Perkasa, Titon Elang and Prof. Yusuf Hendrawan, STP.M.App.Life.Sc. Ph.D and Joko Prasetyo, STP, M.Si (2022) Pendugaan kandungan klorofil pada daun kelor (moringa oleifera.) Berbasis image analysis dengan menggunakan metode artificial neural network. Sarjana thesis, Universitas Brawijaya.

Abstract

Tanaman kelor (Moringa oleifera) memiliki manfaat dan khasiat yang terdapat pada semua bangian tanaman baik daun, batang, akar maupun biji. Kandungan nutrisi yang cukup tinggi menjadikan kelor memiliki sifat fungsional bagi kesehatan serta mengatasi kekurangan nutrisi. Oleh karena kelor disebut Miracle Tree dan Mother’s Best Friend. Ada 2 metode analisis kadar klorofil yaitu destruktif dan non-destruktif. Analisis destruktif dapat merusak sampel dan memerlukan waktu yang cukup lama sedangkan non destruktif dapat menggunakan alat bantu SPAD 502 namun harga alat terbilang mahal, sehingga dikembangkan metode lain yaitu pengolahan citra digital. Tujuan dari penelitian ini yaitu untuk mendapatkan cara mengidentifikasi kandungan klrorofil menggunakan pengolahan citra digital dengan analisa tekstur dan warna dan mendapatkan model ANN terbaik untuk mendukung sistem identifikasi kandungan klorofil pada daun kelor. Analisis tekstur menggunakan fitur warna grey, RGB, HSL, HSV, dan L*a*b* dan fitur tekstur Haralick menggunakan color co-occurrence matrix (CCM) yang meliputi entropy, energy, contrast, homogeneity, sum mean, variance, correlation, maximum probability, inverse difference moment, dan cluster tendency Tahap fitur seleksi dengan metode filter dengan beberapa atribut berbeda untuk mendapatkan fitur tekstur dan warna terbaik. Berdasarkan penelitian diketahui kandungan klorofill mengalami peningkatan seiring meningkatnya tingkat perkembangan daun. Diperoleh 10 fitur terbaik menggunakan Chi-Square sebagai atribut seleksi. Model ANN terbaik dipilih dari 75% data training dan 25% data validasi yaitu dengan struktur 10-30-40-1 (10 node input, 30 node hidden layer 1, 40 hidden layer 2, 1 node output) dengan nilai learning rate 0.1 dan momentum 0.5, traincgf sebagai fungsi pembelajaran terpilih, logsig fungsi aktivasi pada hidden layer dan tansig pada output layer. Struktur ANN terpilih menghasilkan koefisien korelasi (R) training sebesar 0.9792, koefisien korelasi (R) validasi 0.9794 MSE training 0.0100, dan MSE validasi sebesar 0.0099. Hasil penelitian menunjukkan bahwa pengolahan citra digital dengan analisa tekstur dan warna dan model ANN berpotensi sebagai sensor dalam mendeteksi kandungan klorofil daun kelor.

English Abstract

Moringa plant (Moringa oleifera) has benefits and properties found in all parts of the plant, including leaves, stems, roots and seeds. The high nutritional content makes Moringa have functional properties for health and overcome nutritional deficiencies. That's why Moringa is called Miracle Tree and Mother's Best Friend. There are 2 methods of analyzing chlorophyll content, namely destructive and non-destructive. Destructive analysis can damage the sample and take a long time, while non-destructive analysis can use the SPAD 502 tool but the price of the tool is quite expensive, so another method is developed, namely digital image processing. The purpose of this study was to find a way to identify chlorophyll content using digital image processing with texture and color analysis and to obtain the best ANN model to support the identification system of chlorophyll content in Moringa leaves. Texture analysis uses gray, RGB, HSL, HSV, and L*a*b* color features and Haralick texture features use color co-occurrence matrix (CCM) which includes entropy, energy, contrast, homogeneity, sum mean, variance, correlation, maximum probability, inverse difference moment, and cluster tendency The feature selection stage is filtered using several different attributes to get the best texture and color features. Based on the research, it was found that the chlorophyll content increased along with the increasing level of leaf development. The best 10 features were obtained using Chi-Square as a selection attribute. The best ANN model was chosen from 75% of the training data and 25% of the validation data with a structure of 10-30-40-1 (10 input nodes, 30 hidden layer 1 nodes, 40 hidden layer 2, 1 output node) with a learning rate value of 0.1 and momentum 0.5, traincgf as the selected learning function, logsig the activation function in the hidden layer and tansig in the output layer. The selected ANN structure produces a training correlation coefficient (R) of 0.9792, a validation coefficient (R) of 0.7994 MSE training 0.0100, and a validation MSE of 0.0099. The results showed that digital image processing with texture and color analysis and the ANN model has the potential as a sensor in detecting the chlorophyll content of Moringa leaves.

Item Type:	Thesis (Sarjana)
Identification Number:	0522100291
Uncontrolled Keywords:	Daun Kelor, Artificial Neural Network, Klorofil,Artificial Neural Network, Chlorophyll, Moringa Oleifera
Subjects:	600 Technology (Applied sciences) > 630 Agriculture and related technologies
Divisions:	Fakultas Teknologi Pertanian > Keteknikan Pertanian
Depositing User:	Sugeng Moelyono
Date Deposited:	03 Jan 2023 03:51
Last Modified:	03 Jan 2023 03:51
URI:	http://repository.ub.ac.id/id/eprint/196543

	Text (DALAM MASA EMBARGO) Titon Elang Perkasa.pdf Restricted to Registered users only until 31 December 2024. Download (7MB)
	Other (Generate index codes conversion from text to indexcodes) indexcodes.txt Restricted to Registered users only Download (0B)

Actions (login required)

View Item