Funding tool: Basic research projects
Project-ID: FTI18-005
Project start: 01. März 2020
Project end: will follow
Runtime: 36 months / ongoing
Funding amount: € 199.800,00

Brief summary:

Imaging technologies are increasingly accessible to plant sciences and agriculture. These new sources of digital data are expected to enhance in-vivo plant diagnostics, particularly in relation to abiotic and biotic stresses, and thereby support sustainable decisions for “smart-farming”.

This project primarily targets FTI areas "Ecosystems and ecosystem services" and "Sustainable land management and production optimization" by introducing a novel approach to detect plant physiological functioning in spectral imaging data. This is expected to improve monitoring capacities for plant stress and contribute to better phenotyping and precision management solutions.

Currently progress is constraint by a lack of suitable plant physiological reference datasets that allow for development of robust feature detection methods and reliable prediction models for in-vivo plant diagnostics. Therefore the project also has a focus on the FTI area "Data". It develops an innovative approach for physiological trait digitalization by co-acquisition of physiological time series and hyperspectral imaging data. Thereby datasets will be created with sufficient size and controlled modification of ambient conditions to explore spectral features for encoding physiological functioning.

The approach combines spectral imaging with a closed chamber setup for gas exchange, as well as leaf water potential, chlorophyll fluorescence, atmospheric and soil measurements. A comprehensive time series dataset will be acquired for two different grapevine rootstocks (drought tolerant vs. susceptible) subject to dehydration stress via dry down of soil moisture and increase of vapour pressure deficit.

Based on these datasets, novel feature detection approaches for physiological traits and predictive analysis models for abiotic stress will be developed. Application is demonstrated for prediction of grapevine stress response using multimodal models based on spectral features combined with ambient parameters. It is expected that the novel analysis techniques will improve diagnostics of stress influence on the plant compared to simple vegetation indices and soil moisture-based thresholds commonly used in irrigation scheduling.

Results from this project will provide an adapted solution for generating high-quality physiological reference datasets for advanced spectral data usage and enhance the detectability of high-level plant features. The obtained datasets, feature detection methods and prediction models will be compiled into a publicly available library to foster comparability and reproducibility and support different stakeholders involved in smart farming to optimize exploitation of spectral sensor data. Furthermore, the availability of/access to this unique plant trait digitalization pipeline will strengthen involvement of Lower Austrian research institutions into national and European plant phenotyping and smart farming networks for future research initiatives.