Innovar is an H2020 FP7 project which involves 21 different partners (universities, research centers and breeding companies) across Europe. The principal aim is to update and augment valuable information in bread and durum wheat DUS and VCU variety registration protocols by developing precise, rapid and automated methods for DUS testing, revising and developing VCU trial processes, and exploiting synergies between both testing methods.

Towards this goal, two panels of ca. 270 each bread and durum European varieties have been assembled for field trials that are being carried out for VCU and DUS across partners for 3 years (2020-2022 growing season) and data are bein collected using actual CPVO protocols. Different technologies will be used for creating a variety testing database with genomics, phenomics, weather and soil data, and machine learning. Databases were based on historical data at the beginning of the project and are being expanded with newly harmonized data generated from the trial series across Europe. High-throughput phenotyping is being carried out with drones. InnoVar trials are composed of 10 VCU and 3 DUS experimental fields located in Denmark, the United Kingdom, Italy, and Switzerland. On average we are collecting data points along of the wheat growth season at 10 key-developmental stages from establishment/beginning of tillering to complete senescence. DJI Matrice RTK 210 drone carrying a Micasense dual RedEdge multispectral sensor (Micasense, Seattle, WA, USA) has been used for all multispectral imagery with 10 spectral bands. All flights were conducted at 20 m altitude, with 80% forward and side overlaps among images to generate high resolution orthomosaics and dense point clouds. We used Agisoft Metashape, QGIS, and R programming software for preprocessing and acquisition of UAV-based multispectral traits derived from the visible and beyond-visible range of the electromagnetic spectrum, such as Normalized Difference Vegetation Index (NDVI, nitrogen use efficiency and biomass), green and red edge chlorophyll indexes, normalized difference red edge and plant senescence reflectance index  (PSRI). This dataset will be augmented with RGB images collected at the same key developmental stages on VCU field trials. Vegetation indexes in the visible spectrum will be extracted from images and compared with UAV data. Preliminary data were analyzed at UNIBO, extracting NDVI and PSRI from VCU and DUS trials. Results showed significant differences in vegetation indexes between varieties and also between different growing stages. Innovar partners are developing an R pipeline exploiting FieldImageR package to extract all possible vegetation indexes from Micasense dual RedEdge multispectral sensor camera and single wavelengths. This pipeline will represent an important source for high-throughput data analysis from phenomics input data.

By converting these digital measurements into useful biological knowledge (crop traits), we are aiming to extract more information on the diseases, height, photosynthesis, chlorophyll status, senescence, and other physio-chemical properties for setting informed DUS and VCU threshold and performing GWAS and genomic prediction analysis.