From Almanac to AI

An industry that has long depended on an annual almanac is agriculture. Far from that quaint past, the field is now positioned to make quantum leaps driven by 5G and edge processing. While simple environmental sensors like temperature, humidity, and soil monitoring add improvements, disruptions will come from drones mounted with cameras as images and videos fill up the data pipe. Drone-mounted cameras bring about many possibilities, from crop growth monitoring, coverage, crop infestation, anomaly detection, efficient resource utilization, and estimation (yield and fertilizers).

Data created by agricultural drones has two distinct flows:

i. Real-time or near real-time processing for immediate corrections (watering adjustments, soil control, alerts, asset estimation).

1. Wi-Fi drones will send data directly to an on-site (barn) controller for actions.

2. 5G drones will send data offsite to a 5G telco tower with controller function for actions.

ii. Store and forward video data over 5G for model training.

1. Data flows over 5G to the near edge on telco tower, which may in some cases have full-on micro modular data center for first-order deep learning.

2. In cases where the near edge only does preprocessing, the data flows to the cloud edge, which will be a micro modular data center.

3. Data then flows to the core (centralized private or public cloud) for second-order deeper and wider analytics.

4. Finally, data flows to the active archive tier in the core for long-term archiving with faster access for validation, verification, replay, and retraining.

5. A variation is that the edge could be the barn with IP connectivity to the core.

6. Another variation is to transfer data using shuttles from the edge to the core.