The Development of Methods for Detecting Melon Maturity Level Based on Fruit Skin Texture Using the Histogram of Oriented Gradients and the Support Vector Machine

摘要

Melon (Cucumis melo L.) has economic value and promising prospects in the marketing aspect. Export data shows that melons are in the fifth place in fruit commodities foreign exchange earner. The increasing production and demand for melons have not been balanced with the optimization of harvesting and postharvest handling. The similarity of melon skin texture between mature, undercooked and immature for people who have difficulty detecting melons from various fruit skin levels to be different from one assessor with the other assessor. From these problems, research was conducted to develop a method that can detect the maturity level of melons based on fruit skin texture. The purpose of this research is to find out whether the method for detecting the maturity level of melons can be standardized and what method has the best accuracy in detecting melon maturity levels based on fruit skin texture. The data in this study using the 450 images obtained from melon farmers. The results of this research show that the best method for detecting melon maturity levels in this study is a combination of methods (mean filtering matrix [5×5] + gamma contrast stretching 2.0) + (mean adaptive threshold) + (histogram of oriented gradients) + (support vector machine) with an accuracy of 78.67%.