Real-Time Plant Species Classification Using ResNet-18 and MobileNetV3-Large on Pl@ntNet-300K

Accurate identification and classification of plant species are essential for environmental conservation, agricultural management, and botanical research. This study explores a deep learning-based approach for automated plant species identification, leveraging the Pl@ntNet-300K dataset. Two pre-trained Convolutional Neural Network architectures, ResNet-18 and MobileNetV3-Large, originally trained on ImageNet, were evaluated based on their classification performance and computational efficiency. MobileNetV3-Large, optimized for resource-constrained devices, demonstrated its suitability for real-time, low-power applications, while ResNet-18 exhibited higher classification accuracy, making it ideal for high-precision biodiversity monitoring. The study highlights the trade-offs between model accuracy and computational efficiency, emphasizing the potential of deep learning for scalable, real-time plant species classification. These findings provide a foundation for future applications in ecology, precision agriculture, and conservation efforts.