HERNANDA KHOIRIYAH PUTRI, . (2024) KLASIFIKASI KUALITAS CANGKANG TELUR AYAM MENGGUNAKAN CONVOLUTIONAL NEURAL NETWORK (CNN) BERBASIS CITRA DIGITAL. Sarjana thesis, UNIVERSITAS NEGERI JAKARTA.
Text
COVER.pdf Download (1MB) |
|
Text
BAB I.pdf Download (223kB) |
|
Text
BAB II.pdf Restricted to Registered users only Download (1MB) | Request a copy |
|
Text
BAB III.pdf Restricted to Registered users only Download (351kB) | Request a copy |
|
Text
BAB IV.pdf Restricted to Registered users only Download (1MB) | Request a copy |
|
Text
BAB V.pdf Restricted to Registered users only Download (142kB) | Request a copy |
|
Text
Daftar Pustaka.pdf Download (234kB) |
|
Text
Lampiran.pdf Restricted to Registered users only Download (3MB) | Request a copy |
Abstract
Keretakan telur sering terjadi selama proses distribusi. Retakan pada cangkang telur menjadi perhatian serius karena berpotensi menyebabkan kontaminasi dan risiko kesehatan bagi konsumen. Pengelompokan kualitas cangkang telur bertujuan untuk meningkatkan keamanan pangan dan efisiensi distribusi dalam industri perunggasan. Pada penelitian ini, keretakan cangkang telur ayam diklasifikasi berdasarkan citra digital dengan menggunakan Convolutional Neural Network (CNN). Eksperimen dilakukan dengan sampel 210 gambar telur dalam tiga kondisi; terlindungi, cukup terlindungi, dan tidak terlindungi, masing-masing kondisi terdiri dari 70 gambar. Dataset yang diujikan adalah dataset gabungan yang terdiri dari dataset non-augmentasi (asli) dan augmentasi data, serta pengujian hanya pada dataset non-augmentasi. Pra-pemrosesan data meliputi cropping, resize, dan augmentasi. Optimizer yang digunakan adalah Adam dengan splitting data sebesar 80:20 dan epoch sebanyak 50 kali untuk dataset gabungan dan 100 kali untuk dataset non-augmentasi. Performa model dievaluasi menggunakan loss function (sparse categorical cross entropy), confusion matrix, dan metrik evaluasi yang mencakup akurasi, presisi, recall, dan f1-score. Hasilnya menunjukkan bahwa model EfficientNet-B2 mencapai metrik evaluasi yang tinggi pada kedua dataset, dengan akurasi sebesar 98,30% pada dataset gabungan dan 92,86% pada dataset non-augmentasi. Berdasarkan hasil yang diperoleh, augmentasi data dapat meningkatkan akurasi dibandingkan dengan dataset non-augmentasi. Hasil penelitian menunjukkan bahwa citra keretakan cangkang telur ayam dapat dimanfaatkan untuk identifikasi kualitas telur dan dikembangkan untuk klasifikasi kualitas telur ayam. ****** Egg cracks often occurs during the distribution process. Cracks in eggshells are a serious concern as they have the potential to cause contamination and health risks to consumers. Classifying eggshell quality aims to improve food safety and distribution efficiency in the poultry industry. In this study, chicken eggshell cracks are classified based on digital images using Convolutional Neural Network (CNN). Experiments were conducted with a sample of 210 egg images in three conditions; protected, quite protected, and unprotected, each condition consisting of 70 images. The dataset is a combined dataset consisting of non-augmented (original) and augmented data, as well as testing only on the non-augmented dataset. Pre-processing data includes cropping, resizing, and augmentation. The optimizer used was Adam with data splitting of 80:20 and epochs of 50 iterations for the combined dataset and 100 iterations for the non-augmented dataset. Model performance was evaluated using loss function (sparse categorical cross entropy), confusion matrix, and evaluation metrics including accuracy, precision, recall, and f1-score. The results show that the EfficientNet-B2 model achieves high evaluation metrics on both datasets, with an accuracy of 98,30% on the combined dataset and 92,86% on the non-augmented dataset. Based on the results obtained, data augmentation can improve accuracy compared to the non-augmented dataset. The results show that the image of chicken egg shell cracks can be utilized for egg quality identification and developed for chicken egg quality classification.
Item Type: | Thesis (Sarjana) |
---|---|
Additional Information: | 1). Dr.rer.nat. Bambang Heru Iswanto, M.Si. ; 2). Haris Suhendar, M.Sc. |
Subjects: | Sains > Fisika |
Divisions: | FMIPA > S1 Fisika |
Depositing User: | Users 24763 not found. |
Date Deposited: | 13 Aug 2024 03:45 |
Last Modified: | 13 Aug 2024 03:45 |
URI: | http://repository.unj.ac.id/id/eprint/49745 |
Actions (login required)
View Item |