Implementation of Support Vector Machine Method to Predict Harvest Readiness of Wonosalam Coffee Fruits

The quality of coffee is primarily determined by the maturity level of the fruit at harvest time. If the coffee fruit is picked too early, the quality and flavor will not reach its full potential. On the other hand, if the coffee fruit is overripe, it can affect the acidity and taste of the coffee produced. The first stage is picking coffee cherries from the tree by Wonosalam coffee farmers. The picking is done by selecting ripe coffee cherries because if immature cherries are mixed in, the quality and flavor of the coffee may be affected. In addition to the maturity level, fruit defects also affect the quality. Nowadays, picking is done manually by farmers. Farmers often pick unripe fruit in a hurry, resulting in the harvesting of unripe and deformed fruit. As a result, farmers have to re-sort, which reduces their productivity. Therefore, an image processing-based Wonosalam coffee fruit quality detection tool is proposed. This research uses one of the image processing methods used in classifying the maturity of coffee fruit that is ready to harvest, namely the Support Vector Machine (SVM) method. The results showed that the SVM method in classifying the readiness of coffee fruit harvest obtained a total accuracy of 85%.