Enhancing Apiculture through Deep Learning: A Vision-based Honey Bee Classification

The work Honey Bee Classification aims to develop an advanced system for identifying and classifying various species of honey bees. Given the critical role honey bees play in pollination and maintaining biodiversity, accurate identification is essential for monitoring their populations and understanding the factors affecting their decline. Traditional methods of classification, which rely on morphological characteristics, are time-consuming and often prone to error due to the subtle differences between species. This project leverages the power of deep learning to provide a more efficient and accurate solution. The implementation of this deep learning-based approach not only enhances classification accuracy but also significantly reduces the time and effort required compared to traditional methods. The honey bee dataset used in the work has six classes of honey bee. The deep learning models DenseNet121, MobileNetv2 and ResNet50v2 used in the work were consistently trained over 10 epochs with 80:20 data split. DenseNet121 model achieved a high accuracy of 99.8%. The work successfully achieved its goal classifying honey bee using deep learning techniques. The results confirmed the efficacy data augmentation and advanced deep learning models in enhancing classification accuracy.