Modern agriculture and food production systems are facing increasing pressures from climate  change, land and water availability, and, more recently, a pandemic. Despite incredible advances in genetic tools over the past few decades, our ability to accurately assess crop status in the field,  at scale, has been severely lacking until recently. Thanks to recent advances in airborne Artificial  Intelligence and Remote Sensing, we are now looking to quantify field phenotypic information accurately and integrate the big data into useful information to the breeder. Machine Learning algorithms will be developed and used to train and test models based on hyperspectral imagery and  3D crop canopy data to estimate field crops traits that are relevant for breeders to select the most suitable genotypes to increase crop production.

Goal

The pipeline to be developed includes: (i) preprocessing: image registration, denoising and segmentation; (ii) training: learning the visual signature of plant traits; and (iii) validation: validation of the trained models.

Researchers

• Ittai Herrmann
• Matan Gavish

What we did

In progress