KeyGene: Speed breeding of crops through AI-integrated 4D Phenotyping Technologies

KeyGene is an international research company for the development and application of breakthrough technology innovation for crop improvement, with a Dutch headquarters in Wageningen. In the PRECIS project KeyGene and consortium partner JB Hyperspectral Devices were awarded a grant from the Eurostars program to develop a revolutionary AI-driven 4D phenotyping technology (Point cloud (3D) + hyperspectral imaging (1D)) for faster and targeted breeding of crop plants to improve productivity and desired phenotypic traits. Catalyze is proud to have supported KeyGene’s successful subsidy application. Read on to learn more about KeyGene and the thoughts of Digital Phenotyping Scientist, Marco van Schriek, on the PRECIS project.

Overcoming current plant phenotyping bottlenecks

The phenotyping of plants is a crucial step to guide breeding towards productive plants most suitable for their environment. Applying effective phenotyping in the breeding process improves the yield and quality of any crop, thereby increasing crop harvest and better supporting food security. However, the use of existing phenotyping methods leads to challenges in getting reliable results, as they are limited to characterizing single-plant traits in controlled conditions.

The technology developed in PRECIS will overcome this bottleneck by enabling accurate, AI-driven high-throughput acquisition and analysis of multi-dimensional phenotypes. In turn, this will greatly improve the speed and precision of the plant breeding processes, enabling improved crop-yield and quality for the breeder.

AI-driven 4D phenotyping

The project furthers existing plant phenotyping technologies by advancing the state-of-the-art in 3D (point cloud) and hyperspectral imaging to develop next-generation 4D phenotyping. PRECIS will generate 3D point clouds, merging this data with hyperspectral imaging and deep learning algorithms built to provide the AI-interface for 4D phenotyping.

Marco describes how their project will lead to a breakthrough in the market for crop breeding activities:

“The PRECIS project will merge the 3D pointcloud data with hyperspectral imaging to enable within-plant precision phenotyping. This research aims to exploit the quality for stress detection that a hyperspectral camera can deliver and to unleash it on a full plant in a non-destructive way.”

The 3D point cloud images will allow crop breeders to extract structural information, such as leaf angles and internode length, while physiological information from hyperspectral data – used to assess physiological traits such as plant stress, water status – and will be merged with the 3D images to facilitate even deeper insights.

Joining KeyGene in the project consortium is JB Hyperspectral Devices, a company founded in 2016 and based in Düsseldorf, Germany, focused on design and production of hyperspectral field instruments.

Collaborating with Catalyze on the Eurostars submissions

Eurostars is a highly competitive program. Aside from having an excellent project idea to begin with, shaping your project proposal, framing it in the correct light, and finding complimentary expertise are all essential to a successful application.

Marco comments on the support received from Catalyze:

“Catalyze converted the existing scientific ideas into a concise proposal and assisted in finding the right partners for this project.”

Visit the KeyGene website to learn more