EVALUATION OF THE RICE SEED QUALITY CLASSIFICATION SYSTEM  BASED ON THE DEEP LEARNING METHOD - Dalam bentuk buku karya ilmiah

FANIESA SAUFANA HANAFI

Informasi Umum

Kode

23.05.314

Klasifikasi

621.382 - Communication Engineering, signal processing, network protocol, signal processing, communication network analysis

Jenis

Karya Ilmiah - Thesis (S2) - Reference

Subjek

Deep Learning

Dilihat

326 kali

Informasi Lainnya

Abstraksi

<p>The success of the plant and its ability to produce for a long time is very dependent on the quality. Balai Pengawasan dan Sertifikasi Benih Tanaman Pangan dan Hortikultura (BPSBTPH) West Java Province is entrusted with selecting high-quality seeds by performing a certification program as one of the stakeholders in Indonesia's agricultural ecosystem. High-quality seeds are pure and have viability, and the capacity to continue growing into healthy plants. The certification process is still handled by the conventional strategy, absolutely depending on the knowledge and observations of visually by staff who are experts in their professions. Due to the limits of humans, this may result in conflicting observations. The staff at Balai lack experience in processing data based on technology, and there are currently no datasets available for rice seeds. However, this data can also be used to simulate the system process. A system to make human activities easier is needed considering the fast-developing technology. This research creates and evaluates a rice seed classification system based on deep learning methods with VGG-GoogleNet architecture. The VGG-GoogleNet was created by utilizing two types of datasets taken exclusively from the BPSBTPH laboratory, which are the rice seed growth dataset and the rice seed quality dataset. VGG-GoogleNet in this research shows the most accurate results with respect to other architectures. The first dataset result gets an accuracy of 98.25% using the rice seed growth dataset. The second dataset of the rice seed quality gets 94.99% accuracy and 99.30% accuracy using the Random Over Sampling technique for a balancing dataset.</p>

<p>Keywords: Deep Learning, Convolutional Neural Network, Rice Seeds, VGGNet, GoogleNet and Random Over Sampling.</p>