An effective method to inspect and classify the bruising degree of apples based on the optical properties

Detection and classification of the bruising degree for fruit can provide useful information for appropriate postharvest handling and storage operations. However, the detection of early small bruises, which are hardly to detect directly by vision alone, remains a challenge. In this study, we investigated the optical properties of apple flesh with different bruising degree and combined principal component analysis and support vector machine to discriminate and classify the degree of bruising of apples. The apples were divided into three groups and dropped at heights of 0 cm, 10 cm and 20 cm. The optical properties of each group were measured in 400–1050 nm. The measurements of optical properties were performed using an integrating sphere setup and the results were assessed using the inverse adding-doubling method. Based on the measured optical properties, a discriminant model that combines PCA with SVM to classify and predict the bruise degree of apples was established. The accuracy of the model for bruise degree classification was as high as 92.5%. Overall, this study demonstrated it is feasible to detect the early small bruising degree of apples based on optical properties. It also laid a foundation for future studies about detecting early small bruising with non-destructive measurement of optical properties, which is a promising method for rapid and convenient classification of bruise degree of apples.