The integration of Artificial Intelligence into Cyber-Physical Systems (CPS) is transforming precision agriculture by enabling automated, real-time plant health monitoring. AI-driven models enable the early detection of nutrient deficiencies, infections, and pests, allowing for timely intervention with effective treatments. This paper presents an ensemble deep learning model for image-based plant health monitoring, optimized for mobile agricultural equipment and autonomous systems. Preliminary results show significant improvements over existing plant monitoring solutions. Future research will focus on integrating this model into an AI-powered CPS, such as a drone or ground vehicle, for systematic field scanning, and crop health reporting. Unlike conventional CPS with multiple sensors, the final system will rely solely on computer vision, demonstrating that a high-resolution camera can serve as a powerful data acquisition tool in an AI-driven CPS.