AI in Ag: What's that Weed?

Recognizing what something is, is so automatic to us, most of us probably do not even think about it anymore. Much like a computer, we either know what something is or we don’t. But have you ever seen something before, snapped and picture and let Google help you figure it out? Then you’ve used AI.

In agriculture, some really cool innovation is taking place to make that recognition of weeds easier. It can be as simple as ‘what is this weed’, or as complex as image recognition of the weed in real time while making a decision as to if to spray the identified plant or not. The binary, good plant/bad plant scenario.

Cameras and sensors have to perform AI on the edge in order to achieve the low latency response time to understand what application to perform as it travels across the field.

Cameras and sensors have to perform AI on the edge in order to achieve the low latency response time to understand what application to perform as it travels across the field.

Spraying of herbicides, insecticides and fungicides can be guided to specifically spray from only the nozzles, or passes of the field that require the treatment.  Innovation in equipment design enables real time IoT imagery to recognize the crop vs. weeds and spray very specifically only on the weeds.  Applying image recognition to agriculture in real time is innovation that improves the farmers ability to keep the land and their plants healthy.

Enablement of such precision applications, and the collection of data requires “smart equipment” and increasingly Data Science algorithms to optimize.  Precision prescriptions, variable rate seeding, fertilizer and spraying are only the start. Many of the interfaces still require refinement to be easy and intuitive to use, but it is the future of farming.