ANALYSIS OF THE COMBINATION OF K NEAREST NEIGHBOR (KNN) AND K-MEANS IN THE CLASSIFICATION OF RICE LEAF DISEASES USING AN IMAGE SEGMENTATION METHOD

Afandi, Rizal (2026) ANALYSIS OF THE COMBINATION OF K NEAREST NEIGHBOR (KNN) AND K-MEANS IN THE CLASSIFICATION OF RICE LEAF DISEASES USING AN IMAGE SEGMENTATION METHOD. Undergraduate thesis, UPN "Veteran" Jatim.

	Text (Cover) 19081010146_cover.pdf Download (740kB)
	Text (Bab 1) 19081010146_bab1.pdf Download (23kB)
	Text (Bab 2) 19081010146_bab2.pdf Restricted to Repository staff only until 21 July 2029. Download (165kB)
	Text (Bab 3) 19081010146_bab3.pdf Restricted to Repository staff only until 21 July 2029. Download (264kB)
	Text (Bab 4) 19081010146_bab4.pdf Restricted to Repository staff only until 21 July 2029. Download (4MB)
	Text (Bab 5) 19081010146_bab5.pdf Download (11kB)
	Text (Daftar Pustaka) 19081010146_daftarpustaka.pdf Download (72kB)

Abstract

Rice is a primary food commodity in Indonesia that plays an essential role in national food security. However, rice productivity is often reduced due to leaf diseases such as Bacterial Blight, Blast, Brown Spot, and Tungro. Manual disease identification is considered inefficient as it requires expert knowledge and considerable time. Therefore, an automated system based on image processing and machine learning is needed to assist in the fast and accurate classification of rice leaf diseases. This study aims to implement a combination of K-Means and K-Nearest Neighbor (KNN) algorithms for classifying rice leaf diseases using image segmentation methods. The dataset used consists of 5,932 rice leaf images divided into four disease classes: Bacterial Blight, Blast, Brown Spot, and Tungro. The research stages include image preprocessing (grayscale conversion, resizing, and flattening), segmentation using K-Means, and classification using KNN. Testing is conducted using various training and testing data ratios as well as different K values to determine the optimal performance. The results show that the combination of K-Means and KNN is capable of classifying rice leaf diseases with high accuracy. The best performance is achieved using a 90:10 training-to-testing ratio with an optimal K value, producing highly accurate predictions based on confusion matrix evaluation. Therefore, the proposed method is effective in supporting automatic identification of rice leaf diseases. This research is expected to contribute to the development of image-based plant disease detection systems and assist farmers in diagnosing diseases more quickly and accurately, thereby improving agricultural productivity.

Item Type:

Thesis (Undergraduate)

Contributors: