Background: This study aimed to test the potential of nondestructive optical techniques for classifying sustainable-produced tomatoes according to a) growing practices; b) levels of water use (WUE) and partial factor productivity of nutrients (PFP). Three distinct hydroponic growing techniques for water and fertilizer use were applied over two cultivation cycles for two varieties (cv 'Carminio', and cv 'Mose'): i) free drain open cycle cultivation (OPEN); ii) open cycle cultivation with on-demand sensor-based fertigation (SMART); iii) closed cycle cultivation (CLOSED). Hyperspectral images (HSI) in the Vis-NIR and NIR range, as well as reflectance spectra obtained through Fourier Transform (FT)-NIR spectroscopy, were acquired throughout the harvesting period, with approximately 300 fully ripe tomatoes obtained per variety. For each variety, partial least squares discriminant analysis (PLS-DA) was initially employed to discriminate the three cultivation systems and subsequently distinguish two levels of WUE and PFP, per variety. Finally, the data obtained from both varieties were combined and PLS-DA utilized to categorize three levels of WUE: LOW, MEDIUM, and HIGH. Results: The PLS-DA models applied in external prediction for discriminating tomatoes of each variety according the three cultural practices achieved accuracy higher than 79.55 % for data obtained with FT-NIR and HSI in the Vis-NIR range, and lower for HSI (NIR). The performance increased when considering only two distinct classes (based on WUE and nutrients PFP) with accuracy, specificity, and sensitivity higher than 86 %. General models based on spectra obtained with HSI Vis-NIR, using data over the two varieties to classify tomato according to three levels of WUE, yielded accuracy and specificity of 89.8 % and 91.7 %, respectively, utilizing only 20 variables. Conclusions: Results of this study indicated the effectiveness of FT-NIR and HSI (Vis-NIR) techniques for discrimination of tomato fruits cultivated with varying levels of water and fertilizers. Further validation of these methods is needed before their widespread application to support the adoption of low input growing techniques.
Potential application of hyperspectral imaging and FT-NIR spectroscopy for discrimination of soilless tomato according to growing techniques, water use efficiency and fertilizer productivity
Fazayeli H.;Amodio M. L.;Fatchurrahman D.;Burud I.;Colelli G.
2024-01-01
Abstract
Background: This study aimed to test the potential of nondestructive optical techniques for classifying sustainable-produced tomatoes according to a) growing practices; b) levels of water use (WUE) and partial factor productivity of nutrients (PFP). Three distinct hydroponic growing techniques for water and fertilizer use were applied over two cultivation cycles for two varieties (cv 'Carminio', and cv 'Mose'): i) free drain open cycle cultivation (OPEN); ii) open cycle cultivation with on-demand sensor-based fertigation (SMART); iii) closed cycle cultivation (CLOSED). Hyperspectral images (HSI) in the Vis-NIR and NIR range, as well as reflectance spectra obtained through Fourier Transform (FT)-NIR spectroscopy, were acquired throughout the harvesting period, with approximately 300 fully ripe tomatoes obtained per variety. For each variety, partial least squares discriminant analysis (PLS-DA) was initially employed to discriminate the three cultivation systems and subsequently distinguish two levels of WUE and PFP, per variety. Finally, the data obtained from both varieties were combined and PLS-DA utilized to categorize three levels of WUE: LOW, MEDIUM, and HIGH. Results: The PLS-DA models applied in external prediction for discriminating tomatoes of each variety according the three cultural practices achieved accuracy higher than 79.55 % for data obtained with FT-NIR and HSI in the Vis-NIR range, and lower for HSI (NIR). The performance increased when considering only two distinct classes (based on WUE and nutrients PFP) with accuracy, specificity, and sensitivity higher than 86 %. General models based on spectra obtained with HSI Vis-NIR, using data over the two varieties to classify tomato according to three levels of WUE, yielded accuracy and specificity of 89.8 % and 91.7 %, respectively, utilizing only 20 variables. Conclusions: Results of this study indicated the effectiveness of FT-NIR and HSI (Vis-NIR) techniques for discrimination of tomato fruits cultivated with varying levels of water and fertilizers. Further validation of these methods is needed before their widespread application to support the adoption of low input growing techniques.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.