Claim Missing Document
Check
Articles

Found 1 Documents
Search
Journal : Jurnal Informasi dan Teknologi

Akurasi dalam Mengidentifikasi Citra Anggrek Menggunakan Backpropagation Artificial Neural Network Ardia Ovidius; Gunadi Widi Nurcahyo; Sumijan; Roni Salambue
Jurnal Informasi dan Teknologi 2021, Vol. 3, No. 3
Publisher : SEULANGA SYSTEM PUBLISHER

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37034/jidt.v3i3.115

Abstract

Orchids are ornamental flower plants in the Family Orchidaceae whose habitat is spread over almost all continents in the world, except Antarctica. There are so many orchid enthusiasts in Indonesia and this fact made orchids a promising commodity for ornamental plant cultivator. With a variety of orchid species that reach more than 25,000 species, the identification of orchid species becomes a little complicated for orchid lovers. The purpose of this study was to determine the accuracy level of orchid species identification through image recognition so that it can be used as a reference in determining the feasibility of this method. This study used 120 images of orchids in 6 species. The image of the orchid was obtained by shooting at several locations using the camera. The photo is then processed using image processing software by cropping and resizing to speed up computing time during network training. Furthermore, MatLab software is used to perform the feature extraction process in the form of color feature data and moment invariants. Data from feature extraction is used as input for training artificial neural networks using the Back Propagation method. Calculation of the level of accuracy done by testing the network using the test data that has been provided. The trial results show that 26 of 30 were successfully recognized so that the accuracy rate can be calculated, namely 86.7%. An accuracy rate of 86.7% can be considered feasible and can be used as a basis for consideration of using this tested method as the right method for identifying orchids through images.