PROTOTYPE PEMILAH BUAH JERUK KEPROK LOKAL MENGGUNAKAN CONVOLUTIONAL NEURAL NETWORK (CNN) BERDASARKAN UKURAN FISIK DAN CITRA JERUK
Agung Sidiq Santoso, Hidayat Nur Isnianto, S.T., M.Eng.
2023 | Tugas Akhir | D4 Teknologi Rekayasa Instrumentasi dan Kontrol
Today, citrus fruits are among the fruits that are much loved by all levels of society. Citrus fruits are rich in antioxidants and vitamin C which effectively increase the body's resistance to fight various types of diseases. As a result of the high demand for citrus fruits in the market which continues to increase, it causes the sorting of fruit in a fast time but still uses a lot of manual processes carried out by workers who are subjective, meaning that they are only in accordance with the eyes of the workers who carry out the sorting process. In addition, sorting time also becomes longer and requires large costs for workers.
In this study, a citrus fruit sorter prototype was created to overcome these problems that arise due to demand factors and market needs. The citrus fruit sorter prototype in this study was only for sorting one type of local tangerine fruit. The sorting on this prototype uses the SNI (Indonesian National Standard) 3165:2009 so that the expected results are more objective which means according to existing standards not from the eyes of each individual. The citrus fruit sorting method used in the local tangerine fruit sorting prototype based on SNI is the CNN (Convolutional Neural Network) classification method. The CNN method is used because it uses image data as data input and has the advantage of being able to study data to process new data, is more accurate, and can process complex data. With the utilization of agricultural technology, one of which is the citrus fruit sorting prototype, it is hoped that it can overcome all of these problems starting from subjective sorting to objective sorting, faster and more efficient time, and not requiring large sorting costs.
Kata Kunci : Buah Jeruk Keprok Lokal, CNN, SNI, Objektif