<![CDATA[Cayenne pepper (Capsicum frutescens L.) is a horticultural commodity of high economic value, so determining its quality is an important factor in determining the selling price and suitability for consumption. So far, quality assessment is still mostly done manually, but this method tends to be subjective and less efficient. To overcome this, this research evaluates the quality classification of cayenne pepper based on digital image processing using the XGBoost algorithm with three types of features, namely Local Binary Pattern (LBP), Gray Level Co-occurrence Matrix (GLCM), and Hue, Saturation, Value (HSV). The primary dataset used consists of 1,200 images of six quality classes (raw, undercooked, cooked, dry, rotten, and anthracnose). The methodology stages include pre-processing in the form of background removal, resizing, and data augmentation. Next, LBP, GLCM, and HSV feature extraction is carried out, then classification by dividing the test training data by 80:20. The test results show that the best configuration is obtained with the HSV feature, using learning rate parameters 0.1, n_estimators 100, and max depth 12, which produces accuracy (98.92%), higher than using GLCM (88.08%) or LBP (79.17%). These findings confirm that color information is more dominant than texture in supporting automatic quality classification of cayenne peppers.]]>
Copyrights © 2025
