Prabowo, Muhammad Yonanta Cahyo and Prof. Yusuf Hendrawan,, STP M. App.Life.Sc. Ph. D and Prof. Dr. Ir. Sumardi Hadi Sumarlan,, M.S (2022) Identifikasi Kualitas Daging Ayam Dengan Metode Convolutional Neural Network. Sarjana thesis, Universitas Brawijaya.
Abstract
Daging ayam adalah komoditas pertanian non nabati yang masuk ke dalam kebutuhan pokok masyarakat Indonesia yang sangat tinggi akan permintaannya. Di saat permintaan tinggi dan ketersediaan barang yang terbatas menimbulkan masalah di masyarakat, di mana maraknya perdagangan ayam – ayam yang tidak layak konsumsi di pasaran. Dengan adanya kekhawatiran tersebut, mendasari munculnya pemikiran untuk dapat mengidentifikasi perbedaan ayam segar, ayam tiren, ayam busuk dan ayam berformalin memanfaatkan AI deep learning. Convolutional Neural Network (CNN) termasuk ke dalam jenis algoritma deep learning yang perancangannya dipergunakan untuk melakukan pengolahan data menjadi bentuk dua dimensi. Tujuan penelitian ini digunakan untuk mengidentifikasi daging ayam dengan metode CNN. Variasi sampel yang digunakan 500 data (training-validation) dimana, 350 data pengujian (training) dan 150 data pelatihan (testing) dari masing-masing jenis sampel daging yang dibagi dua set trainig-validation dan testing.Arsitektur yang digunakan adalah model pre-trained AlexNet dan GoogLeNet. Pengamatan ini bertujuan untuk membedakan warna dan tektur dari daging ayam. Hasil akurasi data training yang diperoleh dari 500 citra tiap kelas bervariasi. Pada arsitektur GoogLeNet, model terbaik diperoleh dengan tingkat akurasi training-validation tertinggi 99.00% pada optimizer Adam dengan learning rate 0,0001. Pada hasil data testing, confusion matrix dengan nilai akurasi tertinggi diperoleh pada GoogLeNet dengan mengunakan optimizer Adam, learning rate 0.00005, epoch 30 dan mini-batch size 20 dan AlexNet dengan mengunakan learning rate 0.0001. Dari penelitian yang relah dilakukan dapat disimpulkan bahwa Identifikasi daging ayam berbagai kondisi mampu dilakukan dengan menggunakan metode Convolutional Neural Network dengan Pre-Trained (AlexNet dan GoogLeNet). Saran pada penelitian berikutnya agar digunakan microscope dengan resolusi yang lebih tinggi dan pastikan keseragaman citra.
English Abstract
Chicken meat is a non-vegetable agricultural commodity that is included in the basic needs of the Indonesian people, which are very high in demand. At a time when demand is high and goods are limited, it creates problems in the community, where there is a rampant trade in chickens that are not fit for consumption in the market. With these concerns, the underlying idea emerged to be able to identify the differences between fresh chicken, tiren chicken, rotten chicken and formalin chicken using AI deep learning. Convolutional Neural Network (CNN) is a type of deep learning algorithm whose design is used to process data into twodimensional form. The purpose of this study was to identify chicken meat using the CNN method. The variation of the sample used was 500 data (training-validation) in which 350 training data and 150 training data (testing) from each type of meat sample were divided into two sets of training-validation and testing. The architecture used is the pre-trained AlexNet and GoogLeNet models. This observation aims to distinguish the color and texture of chicken meat. The results of the training data accuracy obtained from 500 images for each class vary. In GoogLeNet architecture, the best model is obtained with the highest trainingvalidation accuracy rate of 99.00% on the Adam optimizer with a learning rate of 0.0001. In the data testing results, the confusion matrix with the highest accuracy value is obtained on GoogLeNet using the Adam optimizer, learning rate 0.00005, epoch 30 and mini-batch size 20 and AlexNet using a learning rate of 0.0001. From the research that has been done, it can be concluded that the identification of chicken meat in various conditions can be done using the Convolutional Neural Network method with Pre- Trained (AlexNet and GoogLeNet). Suggestions for future research are to use a microscope with a higher resolution and ensure image uniformity.
| Item Type: | Thesis (Sarjana) |
|---|---|
| Identification Number: | 0522100405 |
| Uncontrolled Keywords: | Convolutional Neural Network, Daging Ayam, Deep Learning, Pre-Trained,Chicken Meat, Convolutional Neural Network, Deep Learning, Pre-Trained |
| Subjects: | 600 Technology (Applied sciences) > 630 Agriculture and related technologies |
| Divisions: | Fakultas Teknologi Pertanian > Keteknikan Pertanian |
| Depositing User: | soegeng sugeng |
| Date Deposited: | 01 Feb 2023 02:38 |
| Last Modified: | 01 Feb 2023 02:38 |
| URI: | http://repository.ub.ac.id/id/eprint/197199 |
![[thumbnail of DALAM MASA EMBARGO]](http://repository.ub.ac.id/style/images/fileicons/text.png) |
Text (DALAM MASA EMBARGO)
Muhammad Yonanta Cahyo Prabowo.pdf Restricted to Registered users only until 31 December 2024. Download (7MB) |
Actions (login required)
 |
View Item |