How drone data significantly improves seed breeding and seed production

Seed companies around the world deploy drone solutions to collect high throughput plant data and unlock the phenotyping bottleneck with aerial drone data.
Collecting large-scale, field-based plant phenotypic data with sufficient resolution and accuracy in a reproducible manner, represents a challenge in plant science research. Seen as a real bottleneck for the effective use of genomic data for crop improvement, it can also be very labor intensive and costly.
Recent drone hardware and image processing innovations now enable frequent, fast, non-invasive and accurate measures of traits, and help streamline processes. Drone solutions also help accurate seed production monitoring and forecasting, thus lowering both logistics and markets risks. These solutions are currently successfully deployed globally by Delair – a leading supplier of professional UAV solutions – and their plant breeder and seed producer customers, allowing them to provide farmers with the most optimal crop varieties in terms of flowering time, grain yields or disease resistance.
By combining new-generation drone systems and sophisticatedaerial imageryanalytics, seed breeders and producers can get a complete, integrated and easy-to-use workflow to extract data. This complete workflow includes:
1. Build the digital twins of an experimental field with drones.
2. Assess and validate the trial.
3. Extract key traits phenotyping thank to aerial imagery analytics.
4. Link the extracted traits at microplot level with market software used by seed companies, so that they can value external data and algorithms.
Delair performed such an operation for Mas Seeds, a seed producer active in 40 countries across continental Europe and around the Mediterranean basin, for all hybrid field crops. Sunflowers, a majorcrop supplying 50% of Europe’s food oil market, were Mas Seeds’ main objective. They wanted to continuously improve sunflower seeds in order to decrease the breeding burden while enhancing the phenotyping capacities.
The seed producer collaborated with Delair on two key aspects of seed breeding: plant counting and yield forecasting. Based on the digital twins of more than 4,000 micro-plots in fields of around 10 hectares (25 acres), Delair was able to provide unprecedented insights on the biomass, the Leaf Area Index (LAI) and the chlorophyll concentration for each plant in the field. A benchmark of every microplot of the field has also allowed to quickly quantify traits and varieties performances.
There are many other traits that can be extracted from aerial imagery, including physical characteristics (plant height and color, emergence quality, flowering, stay-green or weed presence).
  • Emergence quality: select plants that grow very rapidly at the start of the growth cycle.
  • FCover: map the fraction of ground covered by green vegetation
  • Stay Green: enable frequent, accurate and cost-effective monitoring of this key trait, especially for corn crops
  • Flowering stage measurement: benchmark the crop behavior in between the microplots andprovide a flowering rate at microplot level.
Crop Inventory and Yield Forecasting Simplified
Accurate and reliable yield forecasting is a long-standing challenge for seed producers as uncertainty can have high impact across the value chain. To achieve better forecast accuracy, drone-based data isa real game changer to better estimate key parametersthat will have an impact on yields and drive precision practices.
Early in the season, analytics such as exact planted area measure, male and female qualification, plant population estimates and health monitoring are timely tools that greatly decrease the uncertainty areas and allow smart sampling in the field. They also enable precision herbicide spraying, with both an environmental (50% less volume applied, and no product waste) and an economical impacts benefits (savings of 20€/ha).