Study on qualitative impact damage of yellow peaches using the combined hyperspectral and physicochemical indicators method

Abstract

Yellow peaches are soft and prone to impact damage, and the end-use and sales price of yellow peaches are directly affected by their damage degree. Hyperspectral technology has the advantages of fast, non-destructive, etc. It is feasible to directly detect the collision degree of fruit. However, the physicochemical indicators and storage time of fruit after bruise are different, which affect the robustness of the classification model. The geometric parameters are used to assess the damage degree of fruit, but the damage degree is not always positively correlated with geometric parameters. And the damaged fruit can directly lead to changing in its internal quality, and its physicochemical indexes are also affected, so, in order to quickly and accurately identify the impact damage degree of yellow peaches, a combined hyperspectral and physicochemical indicators method is proposed to classify and identify the different levels of impact damage of yellow peaches in this study. In the study process, firstly, the spectral data of samples is extracted and the physicochemical indexes of samples are determined. Then, the statistical analysis of each physicochemical index is performed and the models of them are established. Finally, the value of ripeness index (RPI) is used to characterize the impact damage degree of yellow peaches. The damaged yellow peaches are divided into three groups according to the value of RPI. The value of RPI of damaged yellow peaches greater than 2 is defined as lightly bruise, the value of RPI between 1.2 and 2 is defined as moderately bruise, and the value of RPI less than 1.2 is defined as severely bruise. The results show that the performance of extreme gradient boosting (XGBoost) model under the raw spectrum is best, and the correctness of the prediction set is 93.1%. In conclusion, this study shows that the combined hyperspectral and physicochemical indicators method is feasible to assess the impact damage degree of yellow peaches and it provides a new method to assess the mechanical damage of fruit.