Search

IntroductionBlackheart is one of the most common physiological diseases in potatoes during storage. In the initial stage, black spots only occur in tissues near the potato core and cannot be detected from an outward appearance. If not identified and removed in time, the disease will seriously undermine the quality and sale of theentire batch of potatoes. There is an urgent need to develop a method for early detection of blackheart in potatoes.MethodsThis paper used visible-near infrared (Vis/NIR) spectroscopy to conduct online discriminant analysis on potatoes with varying degrees of blackheart and healthy potatoes to achieve real-time detection. An efficient and lightweight detection model was developed for detecting different degrees of blackheart in potatoes by introducing the depthwise convolution, pointwise convolution, and efficient channel attention modules into the ResNet model. Two discriminative models, the support vector machine (SVM) and the ResNet model were compared with the modified ResNet model. Results and discussionThe prediction accuracy for blackheart and healthy potatoes test sets reached 0.971 using the original spectrum combined with a modified ResNet model. Moreover, the modified ResNet model significantly reduced the number of parameters to 1434052, achieving a substantial 62.71% reduction in model complexity. Meanwhile, its performance was evidenced by a 4.18% improvement in accuracy. The Grad-CAM++ visualizations provided a qualitative assessment of the model’s focus across different severity grades of blackheart condition, highlighting the importance of different wavelengths in the analysis. In these visualizations, the most significant features were predominantly found in the 650–750 nm range, with a notable peak near 700 nm. This peak was speculated to be associated with the vibrational activities of the C-H bond, specifically the fourth overtone of the C-H functional group, within the molecular structure of the potato components. This research demonstrated that the modified ResNet model combined with Vis/NIR could assist in the detection of different degrees of black in potatoes.

More efficient resource utilization and increased crop utilization rate are needed to address the growing demand for food. The efficient quality testing of key agricultural products such as potatoes, especially the rapid testing of key nutritional indicators, has become an important strategy for ensuring their quality and safety. In this study, visible and near infrared (Vis/NIR) transmittance spectroscopy (600–900 nm) was used for the online analysis of multiple quality parameters in potatoes. The study concentrated on comparing three one-dimensional convolutional neural network (1D-CNN) models, specifically, the fine-tuned DeepSpectra, the fine-tuned 1D-AlexNet, and classic CNN, with UVE-PLS (uninformative variable elimination–partial least squares) models. These models utilized spectral data for the real-time detection of dry matter (DM) content in potatoes. To address the challenges posed by limited data from Vis/NIR, this study strategically implemented data augmentation techniques. This approach significantly enhanced the robustness and generalization capabilities of the models. The 1D-AlexNet and DeepSpectra models achieved 0.934 and 0.913 R2P and 0.0603 and 0.0695 g/100 g RMSEP for DM, respectively. Compared to UVE-PLS, the R2P value improved by 21.31% (0.770 to 0.934) for the 1D-AlexNet model and 18.64% (0.770 to 0.913) for the DeepSpectra model. The RMSEP value was reduced by 47.31% (0.114 to 0.0603) for 1D-AlexNet, and 39.30% (0.114 to 0.0695) for the DeepSpectra model. As a result, this study would be helpful for researching the online Vis/NIR transmission determination of potato DM using deep learning. These results highlighted the immense potential of employing specific spectral features in deep-learning models for a more precise and efficient online assessment of agricultural quality. This advancement provided some insight and reference for further contributing to the evolution of more targeted and efficient quality assessment methods in agricultural products.

The accurate detection of external defects in kiwifruit is an important part of postharvest quality assessment. Previous studies have not considered the problems posed by the actual grading environment. In this study, we designed a novel approach based on improved Yolov5 to achieve real-time and efficient non-destructive detection of multiple defect categories in kiwifruit. First, a kiwifruit image acquisition device based on grading lines was developed to enhance the image acquisition. Subsequently, a kiwifruit dataset was constructed based on the external defect characteristics and a new data enhancement method was proposed to augment the kiwifruit samples. Thereafter, the SPD-Conv and DW-Conv modules were combined to improve Yolov5s, with EIOU as the loss calculation function. The results demonstrated that the improved model training loss value was 0.013 lower, the convergence was accelerated, the number of parameters was reduced, and the computational effort was increased. The detection accuracies of the samples in the test set, which included healthy, leaf-rubbing damaged, healed cuts or scarred, and sunburned samples, were 98.8%, 98.7%, 97.6%, and 95.9%, respectively, with an overall detection accuracy of 97.7%. The detection time was 8.0 ms, thereby meeting real-time sorting demands. The average detection accuracy and model size of SSD, Yolov5s, Yolov7, and Yolov5-Ours were compared. When the confidence threshold was 0.5, the detection accuracy of Yolov5-Ours was 10% and 6.4% higher than that of SSD and Yolov5s, respectively. In terms of the model size, Yolov5-Ours was approximately 6.5- and 4-fold smaller than SSD and Yolov7, respectively. Thus, Yolov5-Ours achieved the highest accuracy, adaptability, and robustness for the detection of all kiwifruit categories as well as a small volume and portability. These results can provide technical support for the non-destructive detection and grading of agricultural products in the future.

External defects of kiwifruit seriously affect its added commercialization. To address the existing problems, kiwifruit external defects detection has a few methods for detecting multi-category defects and weak adaptability to complex images. In this study, we proposed ResNet combined with CBAM for the automatic detection of external defects in kiwifruit. The experiment first built an acquisition device to obtain high-quality images. The optimal fusion scheme of ResNet and CBAM was investigated, the network training parameters were optimized, and Adam was used to accelerate the convergence speed of the model. It was found that the average recognition accuracy of ResNet34 + CBAM for kiwifruit was 99.6%, and all evaluation metrics were greater than 99%. Meanwhile, the experiment selected AlexNet, VGG16, InceptionV3, ResNet34, and ResNet34 + CBAM for comparison. The results showed that the recognition accuracy of ResNet34 + CBAM was 7.9%, 12.7%, 11.8%, and 4.3% higher than that of AlexNet, VGG16, InceptionV3, and ResNet34, respectively. Therefore, it can be concluded that ResNet34 + CBAM has the advantages of high recognition accuracy and good stability for kiwifruit external defect sample detection. It provides a technical guarantee for online detection and sorting of kiwifruit and other fruit defects.

Brassica rapa, commonly known as the rapeseed plant, is globally recognized for its nutrient-rich composition and oil-packed seeds, earning its distinction as a substantial oil-seed crop. The seed quality, particularly the germination rate, is instrumental in guaranteeing a high-yield rapeseed crop. Given this, the accurate, quantitative determination and selection of germination rates in seed batches prior to sowing is of paramount importance. However, conventional germination tests, employed to determine the average germination rate of seed batches, are marred by substantial time and cost inefficiencies. This study proposes the use of near-infrared spectral analysis as a proficient, non-invasive approach for assessing germination rates in rapeseed seed batches. The research involved artificial aging of seeds procured from a variety of rapeseed strains, resulting in 228 batches with a broad germination rate spectrum of 15.73% to 99.13%. We recorded near-infrared diffuse reflectance spectra and applied a range of strategies for spectral data preprocessing and feature variable selection. Furthermore, we leveraged support vector regression (SVR) modeling to augment the detection methodology. SVR training and detection were conducted using MATLAB, with selected feature wavelengths undergoing rigorous scrutiny and discussion. The results indicated that employing Savitzky–Golay convolution smoothing for spectral preprocessing, along with Synergy interval Partial Least Squares (SiPLS) in conjunction with Random Frog (RF) for the selection of 50 feature wavelength points, yielded optimal germination rate prediction performance within the SVR model. The coefficients of determination (R2c) for the training set and (R2p) for the testing set were observed to be 0.8559 and 0.8386, respectively, while the Root Mean Square Errors of Calibration (RMSEC) and Prediction (RMSEP) were calculated to be 13.76% and 17.04%. The mechanism of detecting seed vigor through near-infrared spectroscopy was analyzed based on joint variable screening and sensitive variable traceability. Consequently, the SG–SiPLS–RF–SVR model demonstrates its effectiveness in predicting the average germination rate of seed batches, offering a rapid, non-invasive detection method that can be universally applied to various rapeseed strains, thus significantly improving seed production efficiency.