pengukuran kinerja k-nearest neighbors dan self organizing maps menggunakan fast fourier transform untuk identifikasi penyakit tanaman (studi kasus : tanaman padi dan anthurium)
Abstract
Plant diseases can reduce the plant sale price, resulting in losses on farmers especially in the economic field. Eradication of the diseases has been done, but determining the type of disease is still difficult. This research proposed a new study on measuring for automatically identification of plant disease based on leaf image with the input of digital image. Detection process started with feature extraction using Fast Fourier Transform and the identification process using the k-Nearest Neighbor (kNN) and Self Organizing Maps (SOM). This study, used the image which consists of six hundred pieces of images, that divided into six classes consisting of 70% for train data and 30% for test data. The experimental result showed that identification of plant disease using k-NN is better than that of SOM. The accuracy achieved 76% and 62% for k-NN and SOM respectively.
Collections
- UT - Computer Science [2322]