Classification of Vegetable Types Using Singular Value Decomposition (SVD) and K-Nearest Neighbor (KNN) Algorithms

Abstract

Vegetables are widely grown in Indonesia, but sometimes they can be prepared poorly and pose risks to consumers. To solve this problem, we need a high-quality system that can identify good and safe vegetables. This study aims to create a vegetable classification system using pictures and computer algorithms. The system analyzes different types of vegetable images, including color and shape. It uses special techniques called Singular Value Decomposition (SVD) and K- Nearest Neighbor (KNN) to classify the vegetables based on their features. The researchers used a dataset of 121 vegetable images, which were divided into 73 training images and 48 test images. The results showed that the system was able to classify the vegetables with a high accuracy rate of 85.42%. This study has the potential to help improve the quality of vegetables and contribute to the development of automated systems in the agricultural industry.