Discover how cutting-edge AI technologies are transforming the agricultural landscape, empowering scientists to predict crop yields with unprecedented accuracy.
Harnessing Remote Sensing
In this constantly changing world of agriculture, scientists are turning to remote sensing technologies to change how data is collected and analyzed.
Conventional plant phenotyping; such as through manual irrigation and chemical analysis can be laborious, time-consuming, and expensive. The difference now is the integration of uncrewed aerial vehicles (UAVs) and satellites allows a wealth of information to be gathered from a distance, like plant height, light reflecting off the ground, and a whole bunch more.
Not only does this remote sensing approach lighten the research load, it also reveals information that humans on their own cannot perceive. For example, Hyperspectral cameras capture fine-grain light reflectance measurements outside of the visible spectrum, and LiDAR (Light Detection and Ranging) instruments produce dense geometric point clouds of plant architectures.
Unlocking the Potential of AI
With the industry facing a wide array of challenges from climate change, and scientists are turning to AI in their ongoing efforts for answers.
A study published in Frontiers in Plant Science highlights the exceptional power of a model that trains computers to analyze data via long short-term memory: the recurrent neural network. The new deep learning model takes into account remote sensing data, environmental conditions, and genetic markers to accurately predict maize yield.
At Purdue University, a research team — led by Claudia Aviles Toledo, who was studying geomatics at the time of their work on the project — created an online tool called Benthoscope. The models, together with our Ph. D students Nicolas Blanchet & Deepak Kumar, have implemented a neural network that not only classifies healthy and stressed crops but also gives us some insight of how well different management practices can endure on those situations. That kind of insight could change how plant breeders and growers think about optimizing crops and deciding what to farm.
Conclusion
New fields are exposing, new horizons in agriculture innovation remote sensing + deep learning algorithm based genetic field as an era of agriculture upcoming soon. Thanks to this special AI, more accurate crop yield prediction is now possible; this significantly reduces labor costs and allows farmers to make better decisions. As the impacts of climate change mount up in front of us, this is a new technology with the potential to deliver a more secure food system.
