The green plum is rich in amino acids, lipids, inorganic salts, vitamins, and trace elements. It has high nutritional value and medicinal value and is very popular among Chinese people. However, green plums are susceptible to collisions and pests during growth, picking, storage, and transportation, causing surface defects, affecting the quality of green plums and their products and reducing their economic value. In China, defect detection and grading of green plum products are still performed manually. Traditional manual classification has low accuracy and high cost, which is far from meeting the production needs of green plum products. In order to improve the economic value of green plums and their products and improve the automation and intelligence level of the product production process, this study adopted deep learning methods based on a convolutional neural network and cost-effective computer vision technology to achieve efficient classification of green plum defects. First, a camera and LEDs were used to collect 1240 green plum images of RGB, and the green plum experimental classification standard was formulated and divided into five categories, namely, rot, spot, scar, crack, and normal. Images were randomly divided into a training set and test set, and the number of images of the training set was expanded. Then, the stochastic weight averaging (SWA) optimizer and w-softmax loss function were used to improve the VGG network, which was trained and tested to generate a green plum defect detection network model. The average recognition accuracy of green plum defects was 93.8%, the test time for each picture was 84.69 ms, the recognition rate of decay defect was 99.25%, and the recognition rate of normal green plum was 95.65%. The results were compared with the source VGG network, resnet18 network, and green lemon network. The results show that for the classification of green plum defects, the recognition accuracy of the green plum defect detection network increased by 9.8% and 16.6%, and the test speed is increased by 1.87 and 6.21 ms, respectively, which has certain advantages.

The acidity of green plum has an important influence on the fruit’s deep processing. Traditional physical and chemical analysis methods for green plum acidity detection are destructive, time-consuming, and unable to achieve online detection. In response, a rapid and non-destructive detection method based on hyperspectral imaging technology was studied in this paper. Research on prediction performance comparisons between supervised learning methods and unsupervised learning methods is currently popular. To further improve the accuracy of component prediction, a new hyperspectral imaging system was developed, and the kernel principle component analysis—linear discriminant analysis—extreme gradient boosting algorithm (KPCA-LDA-XGB) model was proposed to predict the acidity of green plum. The KPCA-LDA-XGB model is a supervised learning model combined with the extreme gradient boosting algorithm (XGBoost), kernel principal component analysis (KPCA), and linear discriminant analysis (LDA). The experimental results proved that the KPCA-LDA-XGB model offers good acidity predictions for green plum, with a correlation coefficient (R) of 0.829 and a root mean squared error (RMSE) of 0.107 for the prediction set. Compared with the basic XGBoost model, the KPCA-LDA-XGB model showed a 79.4% increase in R and a 31.2% decrease in RMSE. The use of linear, radial basis function (RBF), and polynomial (Poly) kernel functions were also compared and analyzed in this paper to further optimize the KPCA-LDA-XGB model.

Rain spots on green plum are superficial micro-defects. Defect detection based on a two-dimensional image is easily influenced by factors such as placement position and light and is prone to misjudgment and omission, which are the main problems affecting the accuracy of defect screening of green plum. In this paper, using computer vision technology, an improved structure from motion (SFM) and patch-based multi-view stereo (PMVS) algorithm based on similar graph clustering and graph matching is proposed to perform three-dimensional sparse and dense reconstruction of green plums. The results show that, compared with the traditional algorithm, the running time of this algorithm is lower, at only 26.55 s, and the mean values of camera optical center error and pose error are 0.019 and 0.631, respectively. This method obtains a higher reconstruction accuracy to meet the subsequent plum micro-defect detection requirements. Aiming at the dense point cloud model of green plums, through point cloud preprocessing, the improved adaptive segmentation algorithm based on the Lab color space realizes the effective segmentation of the point cloud of green plum micro-defects. The experimental results show that the average running time of the improved adaptive segmentation algorithm is 2.56 s, showing a faster segmentation speed and better effect than the traditional K-means and K-means++ algorithms. After clustering the micro-defect point cloud, the micro-defect information of green plums was extracted on the basis of random sample consensus (RANSAC) plane fitting, which provides a theoretical model for further improving the accuracy of sorting the appearance quality of green plums.

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.