The first major obstacle is the need for robust hardware that can withstand harsh outdoor conditions – dirt, rain, extreme temperatures, varying lighting, and constant vibrations. Equally challenging is the scale of the application: agriculture is practiced worldwide, and developing a system that works reliably across such diverse environments without direct access to all possible data is a significant generalization problem.

However, things are starting to change. Advanced sensors like LiDAR are becoming more affordable, industrial-grade embedded boards are now powerful enough to run deep learning algorithms at the edge, and the industry has finally recognized the critical importance of data collection. These advancements are paving the way for a new era of AI-driven solutions in agriculture.