Virtual Reality (VR) in plant breeding

Graminor is one of the main partners in a new research project where the goal is to implement new technologies, such as Virtual Reality, in order to improve efficiency in developing new plant varieties.

The development of new plant varieties is essential to increase food production and to enable agriculture to cope with new plant diseases, production methods, product requirements and future climate changes. Plant breeding is a labour intensive and time consuming process in which large quantities of data are registered and analysed. The intention of the project is to use cutting edge innovative technological solutions such as Virtual Reality (VR), Genomic Selection (GS) and hyperspectral images to develop methods to bring the field to human evaluation through a VR image. The objective of this is to make the work more efficient and increase the level of precision in developing new plant varieties. The project title is “Reliable and efficient high-throughput phenotyping to accelerate genetic gains in Norwegian Plant breeding (Virtual Phenomics)” and is a four-year project financed by the Norwegian Research Funding for Agriculture and Food Industry. Participants are national and international entities working across different disciplines. The project is to commence in May 2017.

Plant breeding

Today’s traditional plant breeding starts with two “parents” with the desired traits that are genetically different – for example one wheat variety with a high yield and one with good disease resistance – that are crossed. In crossing, one plant pollinates the other plant. The progeny is then selected, based on the desired traits. The purpose of such crossing is to combine several valuable traits, e.g.  introduce disease resistance in a variety that otherwise has a good yield. Selection is a time-consuming process. The new lines (potential new varieties) are sown in the field and properties such as yield levels, disease resistance, height, length, earliness etc. are observed and registered. These observations and registrations are made in the field during the growth season by plant breeders and research technicians. In addition to being labour intensive and time-consuming such observations and registrations are “fixed” at the time made. There is little opportunity to go back in time in the field.

New technological solutions for faster plant development

In the “Virtual Phenomics” project the goal is to develop specially constructed drones and/or camera solutions that can film and photograph new potential plant varieties in the field. These photos will be converted to 3D images on a computer platform. The plant breeder can then use a virtual reality headset to go through the trial field when he/she needs to register and analyse the observations (phenotypical data). With the VR headset the plant breeder can return to the field several times, zoom in and out, and go back in time to check observations several times. It would also be possible for several people to “go into” the same field at the same time on the virtual platform. This will make observations and registrations more accurate and precise, and will accelerate progress on desired properties such as disease resistance.

Virtual Phenomics project partners

Morten Lillemo at the Norwegian University of Life Sciences (NMBU) is responsible for coordinating the project. For many years Morten Lillemo has carried out research related to the development of new wheat varieties and has participated in several international projects. With him from NMBU are Pål Johan From, who has developed the agricultural robot, Thorvald and Ingunn Burud, who is an expert in hyperspectral image analysis. From Graminor, R&D head Muath Al Sheik and wheat breeder Jon Arne will be the main contributors to the project, but the other cereal breeding programmes (barley and oats) will also be involved. The CEO of Graminor, Idun Christie, is head of the steering committe for the project. The development of drones, camera solutions and the content of the virtual software platform will be carried out by the Hamar company Making View AS. Making View has several years’ experience in the production of content, i.e. film, images with text and sound, for VR headsets.

CIMMYT is a Mexican non-profit research and training institution dedicated to both the development of improved varieties of wheat and maize, and introducing improved agricultural practices to farmers. CIMMYT is a world leader in combining gene technology methods with phenotypical data and developing statistical models for use in plant breeding. Boston University in the USA will contribute new technology for high definition 3D image models and “big data” processing.