Klasifikasi Kematangan Buah Tomat Berdasarkan Fitur Warna Menggunakan Metode Watershed

NOVI, Maria (2024) Klasifikasi Kematangan Buah Tomat Berdasarkan Fitur Warna Menggunakan Metode Watershed. Undergraduate thesis, Universitas Katolik Widya Mandira Kupang.

	Text ABSTRAK.pdf Download (450kB)
	Text BAB I.pdf Download (294kB)
	Text BAB II.pdf Restricted to Repository staff only Download (261kB)
	Text BAB III.pdf Restricted to Repository staff only Download (662kB)
	Text BAB IV.pdf Restricted to Repository staff only Download (871kB)
	Text BAB V.pdf Restricted to Repository staff only Download (860kB)
	Text BAB VI.pdf Download (184kB)
	Text DAFTAR PUSTAKA DAN SUKET BEBAS PLAGIAT.pdf Download (996kB)

Abstract

In this research, a classification system for tomato fruit ripeness has been developed based on color features using the Watershed method. Images of tomatoes were taken with a Nikon D3100 camera at three different distances, namely 40 cm, 50 cm and 60 cm, for a total of 1,530 images. Each distance is divided into 300 training data images and 210 test data images, each for ripe, half-ripe and unripe tomatoes. This research aims to determine the classification of ripeness levels of tomatoes automatically, with a preprocessing process to resize and standardize the image, followed by color-based feature extraction through conversion from RGB to HSV. This is done to maintain the quality and selling value of tomatoes, considering that the manual method currently used still has many weaknesses, such as non-uniformity of results and dependence on human subjectivity. The research results showed that the classification system developed was able to identify the level of ripeness of tomatoes with sufficient accuracy. At a distance of 40 cm, the majority of tomatoes are classified as “Semi Ripe”, followed by “Raw” and “Ripe”. At a distance of 50 cm, “Semi Ripe” dominates, followed by “Raw” and “Ripe”. Meanwhile, at a distance of 60 cm, the majority of tomatoes were also classified as "Ripe", followed by "Raw" and "Semi Ripe". The highest accuracy was achieved at a distance of 40 cm with a value of 75.7142%, indicating that the method used was effective in classifying the level of ripeness of tomato fruit. This research confirms that the adopted computer vision approach can be successfully applied for tomato fruit ripeness classification, providing a more efficient and consistent solution compared to manual methods.

Actions (login required)

View Item