Hyperspectral characteristics of bruised tomatoes as affected by drop height and fruit size

Bruising is the most common type of damage occurring during all stages of postharvest handling, and it is difficult to detect by human inspection because of minor changes in appearance. Hyperspectral imaging technology was used to detect bruised tomatoes caused by falling damage, and the influence of fruit size, drop height and detection time point on hyperspectral characteristics was studied. The results showed that the spectra of bruised tomatoes were lower than those of healthy tomatoes, and the features corresponding to the spectra round 810 nm played a significant role. The fruit size had the greatest impact on the spectrum of damaged tissue due to falling damage, while, the detection time point had the smallest effect. Scanning electron microscopy (SEM) observations showed that the drop test under various influencing factors had an obvious effect on the internal cell structure of tomatoes, and the most serious damage to the cell structure occurred in the group with 1 m falling height of large tomatoes. Partial least squares discriminant analysis models were developed to classify bruised tomatoes with an overall classification accuracy of 90.93%. The SEM and simulated data confirmed the experimental results, indicating that hyperspectral technology can classify tomato bruise damage.

Automated summary

Bruising is the most common damage during postharvest handling and is difficult to detect visually. Hyperspectral imaging technology was used to detect bruised tomatoes caused by falls, with fruit size having the greatest impact on spectral features. Scanning electron microscopy showed that drop tests affected the internal cell structure of tomatoes, with the most severe damage in large tomatoes dropped from 1 meter. Partial least squares discriminant analysis models achieved an overall classification accuracy of 90.93%, confirming that hyperspectral technology can classify tomato bruise damage.