The Promise of AI-enabled Precision Agriculture

Pete Weishaupt
2 min readApr 27, 2020

The application of Artificial Intelligence to precision agriculture is another exciting area where we can reduce environmental impact and simultaneously improve yields, thereby alleviating world hunger.

Through computer vision we’ll be able to cost-effectively detect, locate, and treat a variety of pests and diseases. Through agricultural robotics, we’ll be able to accelerate mechanical pruning and harvesting for a variety of crops.

An AI approach to identifying plants and detecting diseases is not an entirely new concept. Computer vision techniques used to identify plant disease have been around since the early 2000’s. What’s exciting is we now have the potential to automate the detection of plant disease.

Ampatzidis and Cruz developed a vision-based artificial intelligence disease detection system in 2018 to successfully identify grapevine Pierce’s disease (PD) and grapevine yellow (GY) and distinguish those diseases from other diseases. PD and GY symptoms are easily confused with other diseases that cause vine stress like black rot, esca, leaf spot, etc. I would argue these breakthroughs are particularly important to those of us who like wine.

Gus Selas asks two important questions which need considering: “Can a computer be better than man in making decisions related to other living organisms in a complex environment? Can an algorithm beat farmer’s gut instinct and experience?”

As it’s become easier to collect data, we need to be mindful of providing farmers with too much data so it becomes more noise than signal. That being said, AI can transform the farmer’s intuition into a data-driven, informed decision.

With technology and the proliferation of data already providing insights at the local farm level, the rapid adoption of precision agriculture techniques provide opportunities for the developing world to skip a few rungs on the ladder, much the way mobile telecommunications enabled large parts of the world to skip out on the landline telephone.

There are three things you should keep in mind when it comes to AI-enabled precision agriculture:

  • Accurate vision-based supervised learning requires extensive samples. (Training Data)
  • We have the technology in place to collect lots of agricultural data.
  • “Machines” can make much better decisions than humans.