Identifikasi Kematangan Buah Jambu Biji Merah dengan Teknik Jaringan Syaraf Tiruan Metode Backpropagation

Abstract

FAHMIL IKHSAN HIDAYAT : Identification of Guava Maturity with Artificial Neural Network Backpropagation Method, supervised by LUKMAN ADLIN HARAHAP and SULASTRI PANGGABEAN. Identification of guava maturity is generally done manually by the farmers. Fruit seen visually by eyes and responded to by the brain to distinguish the level of maturity. In large quantities it will be difficult to maintain the performance of the brain due to the fatigue factor. This study was a non-conventional method of measurement that used digital image processing to produce data that will be proce5ssed by artificial neural networks and then processed using computer software that can be used to determine the level of maturity of guava. Guava are identified based on the histrogram input image color ( RGB ) that obtained from the results of the capture which then application built by using Visual Basic software. Some sample of the learning pattern guava data had different weighted values as input to the neural network by using backpropagation method to distinguish raw, ripe and rotten fruits. This identification system was capable to identify the entire category of fruit which were 83.3 % correct identification. From the identification that had been done, resulting the identification of the three outputs 85 % ripe citrus, over ripe 75 %, and 90 % raw. Results of the identifications were affected by the shooting fruit process.