Classification of Starfruit Maturity Using the KNN Algorithm Based on Color

Abstract

Starfruit is a high-value economic tropical fruit widely used in the food industry. The determination of the fruit's maturity level is typically done manually through visual observation, but this method is subjective and less accurate. This research aims to develop a system to classify the maturity level of starfruit using the K-Nearest Neighbor (KNN) algorithm based on color features in the Hue, Saturation, and Value (HSV) color space. A dataset of 90 images was obtained through an acquisition process using a smartphone camera. This was followed by pre-processing stages, including resizing, segmentation, and HSV feature extraction. The data was split into 80% for training and 20% for testing.The results indicate that the value of the K parameter influences the model's performance, with the highest accuracy obtained at K=7, achieving an accuracy of 94.71%. The overall average accuracy reached 84.54%. These results prove that the KNN algorithm is effective for classifying the maturity level of starfruit based on color characteristics. The developed system has the potential to be an automated sorting solution that is more objective, efficient, and accurate for farmers and agricultural processing industries.