Hyperspectral image processing for automatic evaluation of Brassica rapa subsp. sylvestris (broccoli-raab/rabe) flower opening

In this work a methodological approach based on hyperspectral imaging (HSI) in the visible range (400-740 nm) was applied to monitor timing of Brassica rapa subsp. Sylvestris (broccoli-raab/rabe or “cime di rapa”). flower opening (anthesis) with the goal of developing tools useful to assess/predict optimal harvest times. A classification model based on partial least-squares-discriminant analysis (PLS-DA) was built to evaluate the anthesis of broccoli-raab plants. This preliminary lab-scale investigation showed that HSI is effective and reliable to monitor flower opening in an automatic and non-destructive way. Further studies will be devoted to the application of such strategy directly on the field.