KLASIFIKASI KEMATANGAN BUAH SAWIT DENGAN JARINGAN SYARAF TIRUAN METODE PERCEPTRON
No Thumbnail Available
Date
2020-12
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The development of digital image processing science makes it possible to sort and sort
the maturity level of oil palm fruit with the help of image processing applications. Image
processing techniques are another form of visual observation. Currently, the process of
determining mortality is still using traditional methods, namely looking at the number of
loose fruit and falling from the fruit bunches and the color of the fruit on the bunches.
This study aims to design a system using the perceptron method, measure the accuracy of
the system and measure the correlation between the color of the oil palm fruit and the
level of maturity. The data used is a digital image in JPG format by extracting the RGB
and HSV values. The sample used was palm fruit which presented 2 levels of maturity
which were grouped into 5 fractions, namely F00, F0 categorized as raw fruit, F1, F2 and
F3 categorized as ripe fruit. The amount of input data used amounted to 50 palms then
processed using the single layer perceptron method and using the sigmoid bipolar and
maximal epoh activation functions used were 30 where 10 data were for the training
process and 30 data were for the testing process. The output produced is raw and ripe
palm fruit. The success rate in experiment 1 using flash was 55% and the success rate in
experiment 2 without flash resulted in an accuracy of 80%
Description
Keywords
Palm fruit, RGB, HSV, Perceptron, Artificial Neural Network, Python