The accurate identification of weeds in peanut fields can significantly reduce the use of herbicides in the weed control process. To address the identification difficulties caused by the cross-growth of peanuts and weeds and by the variety of weed species, this paper proposes a weed identification model named EM-YOLOv4-Tiny incorporating multiscale detection and attention mechanisms based on YOLOv4-Tiny. Firstly, an Efficient Channel Attention (ECA) module is added to the Feature Pyramid Network (FPN) of YOLOv4-Tiny to improve the recognition of small target weeds by using the detailed information of shallow features. Secondly, the soft Non-Maximum Suppression (soft-NMS) is used in the output prediction layer to filter the best prediction frames to avoid the problem of missed weed detection caused by overlapping anchor frames. Finally, the Complete Intersection over Union (CIoU) loss is used to replace the original Intersection over Union (IoU) loss so that the model can reach the convergence state faster. The experimental results show that the EM-YOLOv4-Tiny network is 28.7 M in size and takes 10.4 ms to detect a single image, which meets the requirement of real-time weed detection. Meanwhile, the mAP on the test dataset reached 94.54%, which is 6.83%, 4.78%, 6.76%, 4.84%, and 9.64% higher compared with YOLOv4-Tiny, YOLOv4, YOLOv5s, Swin-Transformer, and Faster-RCNN, respectively. The method has much reference value for solving the problem of fast and accurate weed identification in peanut fields.

The intelligent transformation of crop leaf disease detection has driven the use of deep neural network algorithms to develop more accurate disease detection models. In resource-constrained environments, the deployment of crop leaf disease detection models on the cloud introduces challenges such as communication latency and privacy concerns. Edge AI devices offer lower communication latency and enhanced scalability. To achieve the efficient deployment of crop leaf disease detection models on edge AI devices, a dataset of 700 images depicting peanut leaf spot, scorch spot, and rust diseases was collected. The YOLOX-Tiny network was utilized to conduct deployment experiments with the peanut leaf disease detection model on the Jetson Nano B01. The experiments initially focused on three aspects of efficient deployment optimization: the fusion of rectified linear unit (ReLU) and convolution operations, the integration of Efficient Non-Maximum Suppression for TensorRT (EfficientNMS_TRT) to accelerate post-processing within the TensorRT model, and the conversion of model formats from number of samples, channels, height, width (NCHW) to number of samples, height, width, and channels (NHWC) in the TensorFlow Lite model. Additionally, experiments were conducted to compare the memory usage, power consumption, and inference latency between the two inference frameworks, as well as to evaluate the real-time video detection performance using DeepStream. The results demonstrate that the fusion of ReLU activation functions with convolution operations reduced the inference latency by 55.5% compared to the use of the Sigmoid linear unit (SiLU) activation alone. In the TensorRT model, the integration of the EfficientNMS_TRT module accelerated post-processing, leading to a reduction in the inference latency of 19.6% and an increase in the frames per second (FPS) of 20.4%. In the TensorFlow Lite model, conversion to the NHWC format decreased the model conversion time by 88.7% and reduced the inference latency by 32.3%. These three efficient deployment optimization methods effectively decreased the inference latency and enhanced the inference efficiency. Moreover, a comparison between the two frameworks revealed that TensorFlow Lite exhibited memory usage reductions of 15% to 20% and power consumption decreases of 15% to 25% compared to TensorRT. Additionally, TensorRT achieved inference latency reductions of 53.2% to 55.2% relative to TensorFlow Lite. Consequently, TensorRT is deemed suitable for tasks requiring strong real-time performance and low latency, whereas TensorFlow Lite is more appropriate for scenarios with constrained memory and power resources. Additionally, the integration of DeepStream and EfficientNMS_TRT was found to optimize memory and power utilization, thereby enhancing the speed of real-time video detection. A detection rate of 28.7 FPS was achieved at a resolution of 1280 × 720. These experiments validate the feasibility and advantages of deploying crop leaf disease detection models on edge AI devices.

As an important economic crop, peanut is frequently affected by leaf diseases during its growth period, which severely threaten its yield and quality. Therefore, early and accurate disease detection is critical. However, existing lightweight deep learning methods often struggle to balance model size, real-time detection accuracy, and edge device deployment, limiting their widespread application in large-scale agricultural scenarios. This study proposes a lightweight real-time detection model, YOLO-PLNet, designed for edge deployment. The model is based on YOLO11n, with lightweight improvements to the backbone network and Neck structure. It introduces a Lightweight Attention-Enhanced (LAE) convolution module to reduce computational overhead and incorporates a Channel-Spatial Attention Mechanism (CBAM) to enhance feature representation for small lesions and edge-blurred targets. Additionally, the detection head adopts an Asymptotic Feature Pyramid Network (AFPN), leveraging staged cross-level fusion to improve detection performance across multiple scales. These improvements significantly enhance the detection accuracy of peanut leaf diseases under complex backgrounds while improving adaptability for edge device deployment. Experimental results show that YOLO-PLNet achieves a parameter count, computational complexity, and model size of 2.13M, 5.4G, and 4.51MB, respectively, representing reductions of 18.07%, 16.92%, and 15.70% compared to the baseline YOLO11n. The mAP@0.5 and mAP@0.5:0.95 reach 98.1% and 94.7%, respectively, improving by 1.4% and 1.7% over YOLO11n. When deployed on the Jetson Orin NX platform with real-time video input from a CSI camera, the model achieves a latency of 19.1 ms and 28.2 FPS at FP16 precision. At INT8 precision, latency is reduced to 11.8 ms, with real-time detection speed increasing to 41.3 FPS, while GPU usage and power consumption are significantly reduced with only a slight decrease in detection accuracy. In summary, YOLO-PLNet achieves high detection accuracy and robust edge deployment performance, providing an efficient and feasible solution for intelligent monitoring of multiple categories of peanut leaf diseases.

Peanut southern blight has a severe impact on peanut production and is one of the most devastating soil-borne fungal diseases. We conducted a hyperspectral analysis of the spectral responses of plants to peanut southern blight to provide theoretical support for detecting the severity of the disease via remote sensing. In this study, we collected leaf-level spectral data during the winter of 2021 and the spring of 2022 in a greenhouse laboratory. We explored the spectral response mechanisms of diseased peanut leaves and developed a method for assessing the severity of peanut southern blight disease by comparing the continuous wavelet transform (CWT) with traditional spectral indices and incorporating machine learning techniques. The results showed that the SVM model performed best and was able to effectively detect the severity of peanut southern blight when using CWT (WF_770~780, 5) as an input feature. The overall accuracy (OA) of the modeling dataset was 91.8% and the kappa coefficient was 0.88. For the validation dataset, the OA was 90.5% and the kappa coefficient was 0.87. These findings highlight the potential of this CWT-based method for accurately assessing the severity of peanut southern blight.

Crop disease identification is a pivotal research area in smart agriculture, forming the foundation for disease mapping and targeted prevention strategies. Among the most prevalent global wheat diseases, powdery mildew—caused by fungal infection—poses a significant threat to crop yield and quality, making early and accurate detection crucial for effective management. In this study, we present QY-SE-MResNet34, a deep learning-based classification model that builds upon ResNet34 to perform multi-class classification of wheat leaf images and assess powdery mildew severity at the single-leaf level. The proposed methodology begins with dataset construction following the GBT 17980.22-2000 national standard for powdery mildew severity grading, resulting in a curated collection of 4248 wheat leaf images at the grain-filling stage across six severity levels. To enhance model performance, we integrated transfer learning with ResNet34, leveraging pretrained weights to improve feature extraction and accelerate convergence. Further refinements included embedding a Squeeze-and-Excitation (SE) block to strengthen feature representation while maintaining computational efficiency. The model architecture was also optimized by modifying the first convolutional layer (conv1)—replacing the original 7 × 7 kernel with a 3 × 3 kernel, adjusting the stride to 1, and setting padding to 1—to better capture fine-grained leaf textures and edge features. Subsequently, the optimal training strategy was determined through hyperparameter tuning experiments, and GrabCut-based background processing along with data augmentation were introduced to enhance model robustness. In addition, interpretability techniques such as channel masking and Grad-CAM were employed to visualize the model’s decision-making process. Experimental validation demonstrated that QY-SE-MResNet34 achieved an 89% classification accuracy, outperforming established models such as ResNet50, VGG16, and MobileNetV2 and surpassing the original ResNet34 by 11%. This study delivers a high-performance solution for single-leaf wheat powdery mildew severity assessment, offering practical value for intelligent disease monitoring and early warning systems in precision agriculture.

Bridge health assessment and deterioration prediction are essential for traffic safety and maintenance planning. However, conventional bridge evaluation still relies heavily on expert judgment and heuristic rules, limiting objectivity and long-term forecasting capability. This study proposes a physics-constrained multi-task learning–semi-Markov framework for bridge condition assessment, multi-year deterioration forecasting, and risk-aware maintenance prioritization. Guided by the Highway Bridge Technical Condition Rating Code, the proposed model jointly predicts health grade, defect type, and severity from inspection item/defect entry records, improving assessment robustness through cross-task information sharing. The predicted health states are then incorporated into a physics-constrained semi-Markov model to forecast bridge deterioration over a three-year horizon, and the current and predicted states are further used for maintenance prioritization. Experiments on real bridge inspection data show that the proposed model achieves test accuracies of 95.20%, 99.67%, and 99.72% for health grade, defect type, and severity, respectively, with a Cohen’s kappa of 0.872, while the proposed semi-Markov model achieves a three-year prediction accuracy of 95.4% and a weighted accuracy of 99.1%. Comparative and ablation studies further demonstrate the effectiveness of the proposed framework for bridge lifecycle management.