Predicting shelf life along horticultural supply chains: Evaluation of applicable quality parameters using near-infrared scanner

Visible-Near-Infrared Scanners enable a noninvasive prediction of quality properties of fruit and vegetable based on previously created models. A combination of NIR scanners and
machine learning methods can lead to economic improvements and reduction of food waste by strategies like "first expired, first out" and dynamic pricing. In order to identify parameters
capable of showing dynamic postharvest development, three horticultural products with different postharvest behavior (e. g. strawberry, table grape and mango) were chosen for
morphological and statictical analysis. According to the results, a graduation of spectra in correspondence to the day of measurement was noticeable for strawberry regarding the a-
value as well as presumingly mass loss for both mango and table grape. Furthermore, a PLS model for the a-values r2cv = 0.80 was developed for strawberries.