Weather disturbances, difficult backgrounds, the shading of fruit and foliage, and other elements can significantly affect automated yield estimation and picking in small target apple orchards in natural settings. This study uses the MinneApple public dataset, which is processed to construct a dataset of 829 images with complex weather, including 232 images of fog scenarios and 236 images of rain scenarios, and proposes a lightweight detection algorithm based on the upgraded YOLOv7-tiny. In this study, a backbone network was constructed by adding skip connections to shallow features, using P2BiFPN for multi-scale feature fusion and feature reuse at the neck, and incorporating a lightweight ULSAM attention mechanism to reduce the loss of small target features, focusing on the correct target and discard redundant features, thereby improving detection accuracy. The experimental results demonstrate that the model has an mAP of 80.4% and a loss rate of 0.0316. The mAP is 5.5% higher than the original model, and the model size is reduced by 15.81%, reducing the requirement for equipment; In terms of counts, the MAE and RMSE are 2.737 and 4.220, respectively, which are 5.69% and 8.97% lower than the original model. Because of its improved performance and stronger robustness, this experimental model offers fresh perspectives on hardware deployment and orchard yield estimation.

Maize diseases are reported to occur often, and are complicated and difficult to control, which seriously affects the yield and quality of maize. This paper proposes an improved YOLOv5n model incorporating a CA (Coordinate Attention) mechanism and STR (Swin Transformer) detection head, CTR_YOLOv5n, to identify common maize leaf spot, gray spot, and rust diseases in mobile applications. Based on the lightweight model YOLOv5n, the accuracy of the model is improved by adding a CA attention module, and the global information acquisition capability is enhanced by using TR2 as the detection head. The average recognition accuracy of the algorithm model can reach 95.2%, which is 2.8 percent higher than the original model, and the memory size is reduced to 5.1MB compared to 92.9MB of YOLOv5l, which is 94.5% smaller and meets the requirement of being light weight. Compared with SE, CBAM, and ECA, which are the mainstream attention mechanisms, the recognition effect we used is better and the accuracy is higher, achieving fast and accurate recognition of maize leaf diseases with fewer computational resources, providing new ideas and methods for real-time recognition of maize and other crop spots in mobile applications.

Deep-learning-based methods for plant disease recognition pose challenges due to their high number of network parameters, extensive computational requirements, and overall complexity. To address this issue, we propose an improved residual-network-based multi-plant disease recognition method that combines the characteristics of plant diseases. Our approach introduces a lightweight technique called maximum grouping convolution to the ResNet18 model. We made three enhancements to adapt this method to the characteristics of plant diseases and ultimately reduced the convolution kernel requirements, resulting in the final model, Model_Lite. The experimental dataset comprises 20 types of plant diseases, including 13 selected from the publicly available Plant Village dataset and seven self-constructed images of apple leaves with complex backgrounds containing disease symptoms. The experimental results demonstrated that our improved network model, Model_Lite, contains only about 1/344th of the parameters and requires 1/35th of the computational effort compared to the original ResNet18 model, with a marginal decrease in the average accuracy of only 0.34%. Comparing Model_Lite with MobileNet, ShuffleNet, SqueezeNet, and GhostNet, our proposed Model_Lite model achieved a superior average recognition accuracy while maintaining a much smaller number of parameters and computational requirements than the above models. Thus, the Model_Lite model holds significant potential for widespread application in plant disease recognition and can serve as a valuable reference for future research on lightweight network model design.

Soil organic matter (SOM) is an important source of nutrients required during crop growth and is an important component of cultivated soil. In this paper, we studied the possibility of using deep learning methods to establish a multi-feature model to predict SOM content. Moreover, using Nong’an County of Changchun City as the study area, Sentinel-2A remote sensing images were taken as the data source to construct the dataset by using field sampling and image processing. The LeNet-5 convolutional neural network model was chosen as the deep learning model, which was improved based on the basic model. The evaluation metrics were selected as the root mean square error (RMSE) and the coefficient of determination R². Through comparison, the R² of the improved model was found to be higher than that of the linear regression method, Support Vector Machines (SVM) (RMSE = 2.471, R² = 0.4035), and Random Forest (RF) (RMSE = 2.577, R² = 0.4913). The result shows that: (1) It is feasible to use the multispectral data extracted from remote sensing images for soil organic matter content inversion based on the deep learning model with a minimum RMSE of 2.979 and with the R² reaching 0.89. (2) The choice of features has an impact on the prediction of the model to a certain extent. After ranking the importance of features, selecting the appropriate number of features for inversion provides better results than full feature inversion, and the computational speed is improved.

In the academic world, ginseng (Panax ginseng C. A. Meyer) has received much attention as the most representative element of Chinese medicine. To address the lack of traditional algorithms in the identification of ginseng appearance quality and further improve the manual identification on ginseng, we propose a grading method of ginseng appearance quality based on deep learning, taking advantage of the benefits of deep learning in the image identification. Firstly, we substituted LeakyReLU for the conventional activation function ReLU to enhance the predictive power of the model. Secondly, we added an ECA module to the residual block, which allowed attention to be focused on the input object to capture more precise and detailed features. Thirdly, we used the focal loss function to solve the problem of an imbalanced dataset. Then, the self-constructed dataset was processed with data enhancement and divided into four different classes of ginseng. The dataset was trained on a model with transfer learning to finally obtain the best model applicable to the identification of ginseng appearance quality. The experiments showed that, compared with the classical convolutional neural network models VGG16, GoogLeNet, ResNet50 and Densenet121, the proposed model reported the best performance, its accuracy in the test set was as high as 97.39%, and the loss value was 0.035. This method can efficiently classify the appearance quality of ginseng, and has a significant value in the field of ginseng appearance quality identification.

An improved whale optimization algorithm (SWEWOA) is presented for global optimization issues. Firstly, the sine mapping initialization strategy (SS) is used to generate the population. Secondly, the escape energy (EE) is introduced to balance the exploration and exploitation of WOA. Finally, the wormhole search (WS) strengthens the capacity for exploitation. The hybrid design effectively reinforces the optimization capability of SWEWOA. To prove the effectiveness of the design, SWEWOA is performed in two test sets, CEC 2017 and 2022, respectively. The advantage of SWEWOA is demonstrated in 26 superior comparison algorithms. Then a new feature selection method called BSWEWOA-KELM is developed based on the binary SWEWOA and kernel extreme learning machine (KELM). To verify its performance, 8 high-performance algorithms are selected and experimentally studied in 16 public datasets of different difficulty. The test results demonstrate that SWEWOA performs excellently in selecting the most valuable features for classification problems.