Rice diseases are extremely harmful to rice growth, and achieving the identification and rapid classification of rice disease spots is an essential means to promote intelligent rice production. However, due to the large variety of rice diseases and the similar appearance of some rice diseases, the existing deep learning methods are less effective at classification and detection. Aiming at such problems, this paper took the spot images of five common rice diseases as the research object and constructed a rice disease data set containing 2500 images of rice bacterial blight, sheath blight, flax leaf spot, leaf streak and rice blast, including 500 images of each disease. An improved lightweight deep learning network model was proposed to realize the accurate identification of disease types and disease spots. A rice disease image classification network was designed based on the RlpNet (rice leaf plaque net) network model, Which is the underlying network, in addition to the YOLOv3 target detection network model in order to achieve the optimization of the feature extraction link, i.e., upsampling by transposed convolution and downsampling by dilated convolution. The improved YOLOv3 model was compared with traditional convolutional neural network models, including the AlexNet, GoogLeNet, VGG-16 and ResNet-34 models, for disease recognition, and the results showed that the average recall, average precision, average F1-score and overall accuracy of the network model for rice disease classification were 91.84%, 92.14%, 91.87% and 91.84%, respectively, which were all improved compared with the traditional algorithms. The improved YOLOv3 network model was compared with FSSD, Faster-RCNN, YOLOv3 and YOLOv4 for spot detection studies, and the results showed that it could achieve a mean average precision (mAP) of 86.72%, a detection rate (DR) of 93.92%, a frames per second (FPS) rate of 63.4 and a false alarm rate (FAR) of only 5.12%. In summary, the comprehensive performance of the proposed model was better than that of the traditional YOLOv3 algorithm, so this study provides a new method for rice disease identification and disease spot detection. It also had good performance in terms of the common detection and classification of multiple rice diseases, which provides some support for the common differentiation of multiple rice diseases and has some practical application value.

Obtaining the number of plants is the key to evaluating the effect of maize mechanical sowing, and is also a reference for subsequent statistics on the number of missing seedlings. When the existing model is used for plant number detection, the recognition accuracy is low, the model parameters are large, and the single recognition area is small. This study proposes a method for detecting the number of maize seedlings based on an improved You Only Look Once version 4 (YOLOv4) lightweight neural network. First, the method uses the improved Ghostnet as the model feature extraction network, and successively introduces the attention mechanism and k-means clustering algorithm into the model, thereby improving the detection accuracy of the number of maize seedlings. Second, using depthwise separable convolutions instead of ordinary convolutions makes the network more lightweight. Finally, the multi-scale feature fusion network structure is improved to further reduce the total number of model parameters, pre-training with transfer learning to obtain the optimal model for prediction on the test set. The experimental results show that the harmonic mean, recall rate, average precision and accuracy rate of the model on all test sets are 0.95%, 94.02%, 97.03% and 96.25%, respectively, the model network parameters are 18.793 M, the model size is 71.690 MB, and frames per second (FPS) is 22.92. The research results show that the model has high recognition accuracy, fast recognition speed, and low model complexity, which can provide technical support for corn management at the seedling stage.

To ameliorate soil oxygen deficiencies around subsurface drip irrigation (SDI) drippers, aerated irrigation (AI) was introduced to supply aerated water to the soil through venturi installed in the SDI pipeline. The objectives of this study were to investigate the effect of AI on the soil respiration rate and the mechanism of regulation. The Daejeon experiment included two treatments: AI and unaerated SDI as a control check (CK), and used the National Soil Quality Zhanjiang Observation and Experiment Station as a platform to carry out a 2-year (2020–2021) positioning experiment. The effects on the soil respiration rate, soil temperature, soil water content, oxygen content, soil bacterial biomass and root biomass of the two treatments were established. The oxygen content, soil bacterial biomass and root biomass regression equation, using the partial least squares regression analysis (PLSR) algorithm and structural equation modeling (SEM), screened out the influence of soil respiration under AI treatment as the main soil environmental factor and driving mechanism of rate change. The results showed that compared with the control CK, the soil respiration rate, soil oxygen content, root biomass and soil bacterial biomass were significantly enhanced under AI treatment, the soil water content had a decreasing trend, and there was no significant difference in the effect on soil temperature between the different treatments. The regression fitting results showed that the soil respiration rate under both treatments was negatively correlated with soil temperature using a quadratic polynomial correlation, linearly correlated with the soil oxygen content, positively correlated with root biomass and soil bacterial biomass using power function and positively correlated with the soil water content using a cubic polynomial correlation. The PLSR and SEM results demonstrated that aerated irrigation technology could drive the increase in the soil respiration rate by changing the soil oxygen content, root biomass and bacterial biomass.

Introduction: Intelligent rice disease prevention and control are crucial components in the development of smart agriculture. In recent years, with the rapid advancement of computer vision technologies, a variety of deep learning-based methods for rice disease identification have been proposed, and some models have already surpassed the diagnostic performance of agricultural technicians. However, the existing models generally suffer from high computational complexity and limited generalization capabilities, rendering them difficult to deploy on edge devices for real-time and accurate disease recognition under offline field conditions. Methods: To promote engineering applications of related technologies, this study investigated leaf disease identification methods for mountain-grown rice oriented toward edge intelligence. Based on a self-constructed image dataset of mountain rice leaf diseases and following the design principles of lightweight convolutional neural networks, a novel lightweight mountain rice disease recognition model architecture suitable for edge intelligent devices was constructed. Furthermore, a mountain rice leaf disease recognition application was developed for smartphones on the Android platform. Results: Field validation experiments demonstrated that the application achieves an average accuracy of 92.41% across multiple disease categories and an average inference speed of approximately 22.47 frames per second on various smartphone models, indicating high real-time performance and recognition accuracy. Discussion: The research outcomes will provide a reliable theoretical foundation and technical support for the intelligent prevention and control of mountain rice diseases.

Accurate crop pest recognition is critical for intelligent agriculture, yet it remains constrained by the inherent trade-off between complex field variability and the stringent resource limitations of edge devices. In this study, MOIR-Net, a lightweight framework, is proposed to address this bottleneck by integrating Multi-scale Omnikernel Inverted Residuals (MOIR) with Parameter-free SimAM Attention. The MOIR module employs a structural re-parameterization strategy that transforms multi-branch training topologies into single-path inference architectures, thereby enabling lossless feature aggregation with zero inference-time memory overhead. Extensive evaluations conducted on the IP102, PlantVillage, and PlantDoc datasets demonstrate thatMOIR-Net achieves a strong accuracy–efficiency balance while maintaining minimal computational cost (0.967 M parameters, 2.30 ms latency). On the challenging IP102 benchmark, a Top-1 accuracy of 64.18% and mAP of 58.32% are achieved, significantly surpassing state-of-the-art baselines such as MobileViT-v2. Furthermore, the model exhibits superior cross-domain generalization (41.9% Macro-F1 gain) and highly competitive prediction reliability (ECE = 0.0220) through label smoothing. MOIR-Net successfully reconciles extreme computational efficiency with high trustworthiness, thereby offering a robust solution for real-time crop monitoring on resource-constrained agricultural edge platforms.

Introduction: To achieve high-precision and high-efficiency recognition of maize seedlings and weeds in complex field environments while meeting the deployment requirements of resource-constrained edge devices, this study constructed a lightweight object detection model, DCC-YOLOv8n. Methods: Based on YOLOv8n, the model incorporates three key improvements: a dynamic convolution module to enhance feature diversity, a context-guided module to improve target identification in complex backgrounds, and a content-aware reassembly of features (CARAFE) module to improve small-object detection. A self-built dataset of maize seedlings and weeds containing 1,200 images was utilized, incorporating multidimensional data augmentation. Results: Experimental results demonstrated that DCC-YOLOv8n achieved a precision of 90.1% and a recall of 95.5%, with mAP@0.5 and mAP@0.5:0.95 reaching 93.7% and 73.9%, respectively, outperforming YOLOv8n and mainstream comparison models, including Faster R-CNN and YOLOv5n. Deployment on the NVIDIA Jetson Nano edge-computing platform achieved a real-time inference speed of 18.6 FPS with mAP@0.5 of 90.8%, verifying the feasibility of its application and deployment on actual farmland edge devices. Discussion: The proposed DCC-YOLOv8n model achieved an optimal balance between accuracy, lightweight architecture, and real-time performance. The constructed field recognition system effectively addresses the challenges of complex farmland scenarios and provides a viable technical solution for intelligent operations in precision agriculture.