Penerapan Fitur Warna dan Tekstur untuk Identifikasi Kerusakan Mutu Biji Kopi Arabika (Coffea Arabica) di Kabupaten Bondowoso

Zilvanhisna Emka Fitri (Politeknik Negeri Jember)
Brilyan Andi Syahbana (Politeknik Negeri Jember)
Abdul Madjid (Politeknik Negeri Jember)
Arizal Mujibtamala Nanda Imron (Universitas Jember)

Jurnal ilmiah teknologi informasi Asia
Vol 15 No 2 (2021): Volume 15 Nomor 2 (8)

Plantation crops are also a source of foreign exchange Indonesia is coffee. There are only two types of coffee that have economic value for cultivation, namely Arabica coffee and Robusta coffee. Bondowoso is a district in East Java that develops Arabica coffee. The problem is that farmers still use direct observation (manual) on each coffee bean to determine the quality of coffee beans so that this research is expected to be able to assist farmers in sorting the damage to the quality of coffee beans based on color and texture. The features used are color features and GLCM texture features at 0̊ and 45̊ angles. The total number of data is 198. The Backpropagation method is able to classify quality damage to Arabica coffee beans with a training accuracy rate of 100% and a testing accuracy rate of 97.5% at a learning rate variation of 0.5.