We will develop a data analysis pipeline to process hyperspectral (HSL) data for the receptors NRCan and MFFP researchers.

We will transform the raw image data into phenotype traits, e.g. derive spatio-temporal and spectro-temporal proxies for physiological and phenological parameters.

Data preprocessing steps involve using AgiSoft software for point cloud and orthomosaic processing. Subsequent image processing steps use open-source scripts available for Python such as Shapely and Rasterio and our own algorithms for accurate and automated crown detection and delineation for automated tree-level HSL data processing of all trees in each field site.