Green plums have produced significant economic benefits because of their nutritional and medicinal value. However, green plums are affected by factors such as plant diseases and insect pests during their growth, picking, transportation, and storage, which seriously affect the quality of green plums and their products, reducing their economic and nutritional value. At present, in the detection of green plum defects, some researchers have applied deep learning to identify their surface defects. However, the recognition rate is not high, the types of defects identified are singular, and the classification of green plum defects is not detailed enough. In the actual production process, green plums often have more than one defect, and the existing detection methods ignore minor defects. Therefore, this study used the vision transformer network model to identify all defects on the surfaces of green plums. The dataset was classified into multiple defects based on the four types of defects in green plums (scars, flaws, rain spots, and rot) and one type of feature (stem). After the permutation and combination of these defects, a total of 18 categories were obtained after the screening, combined with the actual situation. Based on the VIT model, a fine-grained defect detection link was added to the network for the analysis layer of the major defect hazard level and the detection of secondary defects. The improved network model has an average recognition accuracy rate of 96.21% for multiple defect detection of green plums, which is better than that of the VGG16 network, the Desnet121 network, the Resnet18 network, and the WideResNet50 network.

Most green plums need to be processed before consumption, and due to personal subjective factors, manual harvesting and sorting are difficult to achieve using standardized processing. Soluble solid content (SSC) of green plum was taken as the research object in this paper. Visible near-infrared (VIS-NIR) and shortwave near-infrared (SW-NIR) full-spectrum spectral information of green plums were collected, and the spectral data were corrected and pre-processed. Random forest algorithm based on induced random selection (IRS-RF) was proposed to screen four sets of characteristic wavebands. Bayesian optimization CatBoost model (BO-CatBoost) was constructed to predict SSC value of green plums. The experimental results showed that the preprocessing method of multiplicative scatter corrections (MSC) was obviously superior to Savitzky–Golay (S–G), the prediction effect of SSC based on VIS-NIR spectral waveband by partial least squares regression model (PLSR) was obviously superior to SW-NIR spectral waveband, MSC + IRS-RF was obviously superior to corresponding combination of correlation coefficient method (CCM), successive projections algorithm (SPA), competitive adaptive reweighted sampling (CARS), and random forest (RF). With the lowest dimensional selected feature waveband, the lowest VIS-NIR band group was only 53, and the SW-NIR band group was only 100. The model proposed in this paper based on MSC + IRS-RF + BO-CatBoost was superior to PLSR, XGBoost, and CatBoost in predicting SSC, with R²_P of 0.957, which was 3.1% higher than the traditional PLSR.