Authors:
Mutia Ulfa
1
;
2
;
Rahmad Syah
1
;
2
and
Muhathir
1
;
2
Affiliations:
1
Informatics Depatrment, Faculty of Engineering, Universitas Medan Area, Medan, Indonesia
;
2
Excellent Centre of Innovations and New Science, Universitas Medan Area, Medan, Indonesia
Keyword(s):
Tea Leaf Diseases, Sift Feature Extraction, LVQ, SVM, Classification.
Abstract:
Productivity is highly dependent on healthy leaves, which are the main components of the product. However, plants are very susceptible to all kinds of disturbances. One of these disturbances is a pest that causes disease on tea leaves; the pest is helopeltis. is a type of pest that attacks young leaf shoots by piercing the part to be attacked, and then the puncture mark from the razor will show symptoms in the form of irregular spots. Based on the uniqueness of the damage pattern on the tea leaves, this study tested the classification of the types of tea leaf diseases by comparing two methods, namely support vector machine and learning vector quantization, and utilizing SIFT feature extraction. The level of accuracy produced by each method is 98% using the Support Vector Machine method with 99% precision, 98% recall, and 98% F1-Score, and 94% using the Learning Vector Quantization method with 96% precision, 94% recall, and 96% F1-Score.