Identifikasi Mekar dan Kuncupnya Bunga Sedap Malam Menggunakan Convolutional Neural Network

Abstract

The utilization of technology can aid humans across various sectors, including agriculture. This study harnesses one such technology to identify a particular agricultural commodity, tuberose flowers (Polianthes tuberosa), based on their blooming using Convolutional Neural Network (CNN). The CNN method can assist farmers in harvesting tuberose flowers by distinguishing between bloomed and budding flowers. In this research, a dataset comprised of 600 primary data points captured via smartphones is utilized, divided into 420 training sets and 180 testing sets. Three scenarios are tested, involving training epochs of 10, 15, and 20. The testing results indicate that the first scenario achieves an accuracy score of approximately 82.44%, falling below the 85% threshold. Meanwhile, the second and third scenarios achieve accuracies of approximately 91.20% and 92%, respectively