Project Details:

Title:
A Public-Private Partnership to Use Drone-acquired Metrics to Increase Accuracy of Yield Estimation in Multi-environment Yield Trials of Soybeans

Parent Project: This is the first year of this project.
Checkoff Organization:Indiana Soybean Alliance
Categories:Crop management systems, Analytical standards & measurements
Organization Project Code:
Project Year:2017
Lead Principal Investigator:Katy M Rainey (Purdue University)
Co-Principal Investigators:
Keywords:

Contributing Organizations

Funding Institutions

Information and Results

Click a section heading to display its contents.

Project Summary

Purdue will collaborate with Beck's Hybrids to use drone based canopy measurements to increase the accuracy of yield estimation in Beck's replicated soybean yield trials. This project is distinct from previous activity because they will focus on measuring canopy development in advanced lines that are being considered for sale to farmers as varieties. Purdue and Beck's will also explore development of custom seeding rates for soybean varieties based on canopy development so that farmers can minimize plant populations.

Project Objectives

1. Test the efficacy of image based parameters acquired with drones to increase the accuracy of yield estimation in multi-environment yield trials of soybeans.

2. Develop methods to determine custom plant populations for all soybean varieties using UAS imagery.

Project Deliverables

Project outputs will be published and will advance adoption of phenomic approaches by the private sector. The project enhances capabilities in Indiana for precision agriculture and incorporation of new technologies into breeding and agronomy.

Progress of Work

Updated February 13, 2019:
While farmers and agronomists are well aware that canopy development is critical to yield in soybean, it had never before been quantified for selection of new soybean varieties until the Purdue soybean breeding lab, led by Dr. Katy Martin Rainey, began doing so in 2015. Quantifying soybean canopy development has been made possible with recent advances in drone technology (or unmanned aerial systems; UAS). To accelerate the impact of precision phenotyping technology on farmers, we sought to extend our capabilities to seed companies, so for this project we established a public-private partnership with Beck’s Hybrids. We started in 2017 and we are continuing the project in 2018-19 to obtain two seasons of data to test the efficacy of image-based parameters from UAS.
One our goals is to use UAS canopy measurements to increase the accuracy of yield estimation in Beck’s replicated soybean yield trials, with a focus on measuring canopy development in advanced lines that are being considered for sale to farmers as varieties. The idea is that any yield trial has a margin of error that we can improve with UAS data. In 2017 and 2018, we have conducted over 35 UAS flights over Beck’s on-farm research trials (MG 2-4.2) in Remington and Lafayette, IN, acquiring hundreds of RBG images each flight. We completed automatic extraction of plot images of equal pixel dimensions from drone imagery of each plot, including quantification of canopy coverage and row length. Beck’s has shared 2017 yield data for both locations, and the experiment is ongoing for 2018. Purdue implemented different statistical models that include yield covariates, canopy coverage and estimated row length from UAS, and spatial statistics to calculate yield predictions. As expected, there were changes in the lines rankings between the different models in all the trials. We will compare Purdue rankings from two years and two locations to Beck’s overall yield rankings from about a dozen environments.
The Rainey lab at Purdue and Beck’s are also exploring development of custom seeding rates for soybean varieties based on canopy development, so that farmers can minimize plant populations on a variety basis. A fast canopy development promoted by medium (110,000 seed/acre) and high (160,000 seed/acre) seeding rates planted early in the season increased the yield in around 14% (10 bu/acre) compared with the low seeding rate (60,000 seed/acre). However, the change in yield between the medium and high seeding rate were marginal with just 5% (4 bu/acre) of difference. Increased seeding rates in late planting date (third week of June 2017) did not show any advantage on final yield even when canopy differences were observed. The differences among the same three seeding rates were lower than 3% (1.7 bu/acre) in all the cases. These preliminary results suggest that improving the canopy development through seeding rates increases the yield when the planting date occurs early in the season but has limited effect in late planting dates. Seeding rates greater than ~ 110,000 seed/acre might not imply important yield increases in early season while late season could be managed with seeding rates around ~ 60,000-100,000 seed/acre. These results also provide evidence for a so-called ‘genotype-by-environment interaction’ that warrants developing custom seeding rates on a variety basis that optimize canopy development.

Updated July 17, 2019:

View uploaded report PDF file

Final Project Results

View Attachment for final report

Benefit to Soybean Farmers

The project enhances capabilities in Indiana for precision agriculture and incorporation of new technologies into breeding and agronomy. The project introduces higher yielding varieties to farmers.

Performance Metrics


Project Years