Muskmelon Maturity Stage Classification Model Based on CNN

How to quickly and accurately judge the maturity of muskmelon is very important consumers sorting staff. This paper presents a novel approach solve difficulty stage classification in greenhouse other complex environments. The color characteristics were used as main feature discrimination. A modified 29-layer ResNet was applied with proposed two-way data augmentation methods for stages using indoor outdoor datasets create robust model that can generalize better. results showed code which first way caused more performance degradation than input image augmentation—the second way. established effectiveness compared augmentation. Nevertheless, augmentations including combination revealed an excellent F1 score ?99%, hence applicable computer-based platform quick classification.