<![CDATA[The popularity of coffee-based dishes is now being favored by many people, especially Indonesians. Not a few people also want to try to make coffee dishes (roasted coffee beans) with their own flavors but are constrained by such expensive equipment for cooking coffee beans with sophisticated systems purchased in the market. With so many color standards for the level of ripeness of coffee beans available, an innovation has emerged to determine the ripeness level of the coffee bean roast automatically, more efficiently, and as desired. However, in taking pictures there are constraints on excessive light intensity so that the resulting image looks biased. From these problems, an approach can be made using the Hue and Saturation color features which are able to detect and reduce excess light intensity from the outside and become the main focus of this research. The features used are the Hue and Saturation color spaces using the values ​​from the normalization results of the Red, Green, and Blue (RGB) color features, and using 500 coffee roast images. This study uses two methods, the first uses RGB values ​​without normalizing them by using the features Hue and Saturation which are able to get 76% accuracy with 3 classes of data matching results (Green Roast, Yellow Roast, and Dark Roast), and the second feature Hue and Saturation normalization are able to get an accuracy of 70% with data matching results of 4 classes (Green Roast, Yellow Roast, Light Roast, and Dark Roast).]]>
Copyrights © 2021
