Non-destructive assessment of kiwifruit physico-chemical parameters to optimise the osmotic dehydration process: A study on FT-NIR spectroscopy

• OD process allows to treat unripe fruits to obtain final acceptable products.
• OD promoted changes dependent on kiwifruit ripening stage.
• NIRS was evaluated to simultaneously predict four physico-chemical parameters.
• Predictive models were successfully built by means of PLSR analysis.
• This study potentially allows the on-line optimisation of OD process-parameters.

Non-destructive rapid method based on FT-NIR spectroscopy is assessed to predict the processing response of raw materials at different ripening stages. During osmotic dehydration (61.5% sucrose solution, 5 h) ripe and unripe kiwifruits were analysed with FT-NIR spectroscopy and the most representative physico-chemical parameters to osmotic dehydration (dry matter, soluble solids content, water self-diffusion coefficient and firmness) were assessed by destructive measurements. Predictive models were successfully built by means of partial least square regression (PLSR) analysis (R2 > 0.772, test set validations) for all the four parameters destructively measured. The application of vector normalisation pre-processing was critical to eliminate spectral information that did not relate to the OD process. FT-NIR spectroscopy can successfully predict the evolution of kiwifruit physico-chemical parameters during osmotic dehydration. Thus it can be used as a tool to tune online the process parameters (e.g. time and temperature) to obtain a standardised final product starting from non-homogeneous raw materials.