Akademik Veri Yönetim Sistemi
9th International Artificial Intelligence and Data Processing Symposium, IDAP 2025, Malatya, Türkiye, 6 - 07 Eylül 2025, (Tam Metin Bildiri)
Soybeans, an important agricultural crop worldwide, are an important source of protein and oil in human nutrition and livestock production. For effective agricultural and long-term food production, it is important to quickly and accurately check the quality of seeds. This study suggests a hybrid method that uses both deep learning-traditional machine learning to automatically categorizing soybean seeds into five different quality groups (skin-damaged, broken, immature, spotted, and intact). First, four distinct Convolutional Neural Network (CNN) models, including MobileNetV2, DarkNet53, AlexNet, and ResNet18, are trained and evaluated using a collection of labeled seed images. Deep features are extracted from each of these models. These extracted features are then classified separately by classical classifiers such as Support Vector Machines (SVM), Artificial Neural Networks (ANN), and K-Nearest Neighbor (KNN). The experimental findings suggest that the hybrid strategy enhances classification performance considerably when compared to CNN models employed alone. In particular, the DarkNet53+SVM combination shows the best performance with 93.10% accuracy. These findings reveal that the proposed method provides high accuracy and generalizability in agricultural product classification.