Ceres Imaging Launches Index One Step Closer To Predicting Yield

Posted by Margy Eckelkamp on Tue, 01/22/2019 - 14:48

What if technology could predict yields months before harvest—down to the individual plant level? That’s one goal for the team at Ceres Imaging.

Originally designed for permanent tree crops, Ceres Imaging developed its Cumulative Stress Index to help producers and other agricultural stakeholders derive accurate yield predictions. This product combines multiple imagery layers from the growing season to assess pre-harvest yield potential and crop health for future management decisions.

“This is a next generation of imagery product for us,” says Ceres Imaging CEO Ash Madgavkar. “It combines not just our base type of images but also elements of computer vision, machine learning and plant science to drive conclusions and insights.”

He explains the cumulative stress score is a summation of the stress experienced at the plant level over the course of a season. The first round of development of the index focused on almond trees and the plant response to irrigation amounts and timings.

“We used in-season imagery and next generation analytics, to show what parts of the field may be 10% to 20% more productive, and we found an incredibly strong correlation to final yields,” Madgavkar says.

In developing the index, the Ceres team used 20 in-season images, but Madgavkar says at a minimum growers can use eight in-season images to correlate to the index. The company partnered with the University of California Cooperative Extension Service for the testing and development of this product, which included three University fields of 160 acres each plus commercial grower partners with thousands of acres.

Madgavkar also says the company is currently in development on similar products for use in dryland production and row crops.

“If we are able to predict yields ahead of harvest, that is something valuable to the grower for marketing and any remaining decisions in-season,” he says.

Ceres Imaging uses aerial images gathered via planes but can also overlay other remote sensing technologies such as satellite images and drone images.

The company is working on additional imagery development projects. This includes a new concept combining historical satellite data as a baseline for fields and then overlaying in-season aerial images to pinpoint potential stress areas with triggered alerts.


You can read more here.