Breeding companies aim to improve or introduce new traits and create new varieties. A large part of their activities focus on maize and winter barley breeding. To find new traits thousands of breeding plots need to be scored and analysed. Traditionally, this is done visually or manually, which is laborious, time consuming and expensive. Plant breeding trails would therefore benefit from fast, non-destructive techniques to score and evaluate trials. Remote sensing platforms and sensors, e.g. UAVs with hyperspectral, thermal and laser scanning cameras are now being used to gain spatial information at a very high resolution. However, the translation of these imagery into precise plant traits, such as emergence success, plant height, biomass and grain yield is the challenge.
This thesis topic is setup in close collaboration with a breeding company. Breeding trials contain a variety of genetic material and provide reference data for each experimental plot within the parcel. They are therefore interesting study areas for remote sensing research. UAVs can serve as a platform for small cameras or sensors. Nowadays, high spectral resolution sensors can be mounted on the UAV, which provides us with more spectral information. Several data products can be extracted from the acquired images: plant height can be calculated by Structure for Motion, but more precise by laser scanning; and spectral data is useful for analysing e.g. chlorophyll, plant diseases, and yield. These acquired products should be validated by ground reference data.
Over the season the company will provide a variety of reference data per experimental plot. Manual measurements during the season include plant count, plant height and visual scoring. At harvest they measure yield, dry matter and different quality parameters. Several UAV flights with hyperspectral and laser scanning cameras have been conducted over the season in 2017, but analysis to find the best correlation with yield and related plant traits of the maize and winter barley trials still have to be executed.
- Derive plant traits, such as plant height and biomass from LiDAR and hyperspectral imagery
- Validate derived plant traits from UAV imagery as vegetation height, canopy ground cover, logging areas, LAI, above-ground biomass, and yield for e.g. forage maize, corn maize and winter barley on basis of provided field measurements from Limagrain BV
- Jimenez-Berni et al., 2018. High Throughput Determination of Plant Height, Ground Cover, and Above-Ground Biomass in Wheat with LiDAR. Frontiers in Plant Science, Feb 2018, Volume 9, Article 237.
- Malambo et al., 2017. Multitemporal field-based plant height estimation using 3D point clouds generated from small unmanned aerial systems high-resolution imagery.
- Brede,B. et al., 2017. Comparing RIEGL RICOPTER UAV LiDAR Derived Canopy Height and DBH with Terrestrial LiDAR. SENSORS 2017, 17, 2371
Theme(s): Sensing & measuring, Integrated Land Monitoring