Image-Based Honeybee Colony Conditions Detection Using a Hybrid CNN–ANN Framework

Honeybee health is critical for agriculture and ecosystems, yet traditional hive inspections are time-consuming and prone to error. This paper presents a hybrid Deep Learning (DL) framework for image-based detection of six common honeybee conditions. The method integrates a dual-branch Convolutional Neural Network (CNN) for multi-scale feature extraction with a Multi-Layer Feedback Artificial Neural Network (MLFB-ANN) classifier, replacing the conventional Softmax layer to improve generalization on fine-grained classes. A curated dataset was used to train and evaluate the model. Experimental results show that the hybrid approach achieves 97.61\% accuracy and a macro-F1 score of 0.96, outperforming a baseline CNN+Softmax model (93.6\% accuracy). The proposed system demonstrates strong robustness across classes, particularly in reducing confusion between visually similar conditions such as Varroa and SHB. These findings highlight the potential of feedback-driven classifiers for challenging multi-class image recognition tasks and support the development of real-time, automated hive monitoring systems.