Deep Learning-Based Localization and Recognition of Diseases and Pests in Coffee Plants

Coffee cultivation plays a vital role in the global economy, yet its productivity is frequently threatened by leaf diseases and pests that reduce yield and quality. Traditional disease identification relies on manual inspection, which is labor-intensive, time-consuming, and prone to error. To address this challenge, we propose an automated deep learning approach for coffee leaf disease detection using MobileNetV2, a lightweight convolutional neural network architecture. Our framework leverages efficient feature extraction for real-time disease recognition, making it suitable for deployment on resource-constrained devices like smartphones. The model is trained on a carefully curated dataset ensuring class balance across multiple disease categories including Coffee Leaf Rust, Brown Spot, and Leaf Miner damage. Data augmentation techniques enhance robustness against varying environmental conditions. The system achieves high classification accuracy while maintaining computational efficiency, enabling farmers to obtain instant diagnoses in the field. By facilitating early detection and precise identification, our approach supports timely intervention, reduces crop losses, and minimizes unnecessary pesticide usage. This practical solution empowers farmers with actionable insights, reduces dependency on expert agronomists, and contributes to sustainable agricultural practices.