A novel deep learning model, DiffuCNN, is introduced in this paper, specifically designed for counting tobacco lesions in complex agricultural settings. By integrating advanced image processing techniques with deep learning methodologies, the model significantly enhances the accuracy of detecting tobacco lesions under low-resolution conditions. After detecting lesions, the grading of the disease severity is achieved through counting. The key features of DiffuCNN include a resolution enhancement module based on diffusion, an object detection network optimized through filter pruning, and the employment of the CentralSGD optimization algorithm. Experimental results demonstrate that DiffuCNN surpasses other models in precision, with respective values of 0.98 on precision, 0.96 on recall, 0.97 on accuracy, and 62 FPS. Particularly in counting tobacco lesions, DiffuCNN exhibits an exceptional performance, attributable to its efficient network architecture and advanced image processing techniques. The resolution enhancement module based on diffusion amplifies minute details and features in images, enabling the model to more effectively recognize and count tobacco lesions. Concurrently, filter pruning technology reduces the model’s parameter count and computational burden, enhancing the processing speed while retaining the capability to recognize key features. The application of the CentralSGD optimization algorithm further improves the model’s training efficiency and final performance. Moreover, an ablation study meticulously analyzes the contribution of each component within DiffuCNN. The results reveal that each component plays a crucial role in enhancing the model performance. The inclusion of the diffusion module significantly boosts the model’s precision and recall, highlighting the importance of optimizing at the model’s input end. The use of filter pruning and the CentralSGD optimization algorithm effectively elevates the model’s computational efficiency and detection accuracy.

This study proposes a model for leafy vegetable disease detection and segmentation based on a few-shot learning framework and a prototype attention mechanism, with the aim of addressing the challenges of complex backgrounds and few-shot problems. Experimental results show that the proposed method performs excellently in both object detection and semantic segmentation tasks. In the object detection task, the model achieves a precision of 0.93, recall of 0.90, accuracy of 0.91, mAP@50 of 0.91, and mAP@75 of 0.90. In the semantic segmentation task, the precision is 0.95, recall is 0.92, accuracy is 0.93, mAP@50 is 0.92, and mAP@75 is 0.92. These results show that the proposed method significantly outperforms the traditional methods, such as YOLOv10 and TinySegformer, validating the advantages of the prototype attention mechanism in enhancing model robustness and fine-grained feature expression. Furthermore, the prototype loss function, which optimizes the distance relationship between samples and category prototypes, significantly improves the model’s ability to discriminate between categories. The proposed method shows great potential in agricultural disease detection, particularly in scenarios with few samples and complex backgrounds, offering broad application prospects.

An innovative framework for peach tree disease recognition and segmentation is proposed in this paper, with the aim of significantly enhancing model performance in complex agricultural settings through deep learning techniques and data fusion strategies. The core innovations include a tiny feature attention mechanism backbone network, an aligned-head module, a Transformer-based semantic segmentation network, and a specially designed alignment loss function. The integration of these technologies not only optimizes the model’s ability to capture subtle disease features but also improves the efficiency of integrating sensor and image data, further enhancing the accuracy of the segmentation tasks. Experimental results demonstrate the superiority of this framework. For disease detection, the proposed method achieved a precision of 94%, a recall of 92%, and an accuracy of 92%, surpassing classical models like AlexNet, GoogLeNet, VGGNet, ResNet, and EfficientNet. In lesion segmentation tasks, the proposed method achieved a precision of 95%, a recall of 90%, and an mIoU of 94%, significantly outperforming models such as SegNet, UNet, and UNet++. The introduction of the aligned-head module and alignment loss function provides an effective solution for processing images lacking sensor data, significantly enhancing the model’s capability to process real agricultural image data. Through detailed ablation experiments, the study further validates the critical role of the aligned-head module and alignment loss function in enhancing model performance, particularly in the attention-head ablation experiment where the aligned-head configuration surpassed other configurations across all metrics, highlighting its key role in the overall framework. These experiments not only showcase the theoretical effectiveness of the proposed method but also confirm its practical value in agricultural disease management practices.

The development of smart agriculture has created an urgent demand for efficient and accurate weed recognition and detection technologies. However, the diverse and complex morphology of weeds, coupled with the scarcity of labeled data in agricultural scenarios, poses significant challenges to traditional supervised learning methods. To address these issues, a weed detection model based on a semi-supervised diffusion generative network is proposed. This model integrates a generative attention mechanism and semi-diffusion loss to enable the efficient utilization of both labeled and unlabeled data. Experimental results demonstrate that the proposed method outperforms existing approaches across multiple evaluation metrics, achieving a precision of 0.94, recall of 0.90, accuracy of 0.92, and mAP@50 and mAP@75 of 0.92 and 0.91, respectively. Compared to traditional methods such as DETR, precision and recall are improved by approximately 10% and 8%, respectively. Additionally, compared to the enhanced YOLOv10, mAP@50 and mAP@75 are increased by 1% and 2%, respectively. The proposed semi-supervised diffusion weed detection model provides an efficient and reliable solution for weed recognition and introduces new research perspectives for the application of semi-supervised learning in smart agriculture. This framework establishes both theoretical and practical foundations for addressing complex target detection challenges in the agricultural domain.

In facility horticultural production, intelligent disease recognition and precise intervention are vital for crop health and economic efficiency. We construct a multi-source dataset from Bayan Nur, Weifang, and Honghe that integrates handheld camera photos, drone field images, and laboratory-controlled samples. Handheld images capture fine lesion texture for close-up diagnosis common in greenhouses; drone images provide canopy-scale patterns and spatial context suited to open-field management; laboratory images offer controlled illumination and background for stable supervision and cross-crop feature learning. Our objective is robust cross-crop, cross-regional diagnosis and economically rational control. To this end, a model named CCGD-Net is proposed. It is designed as a multi-task framework. The framework incorporates a multi-scale perception module (MSFE) to produce hierarchical representations. It includes a cross-domain alignment module (CDAM) that reduces distribution shifts between greenhouse and open-field environments. The training follows an unsupervised domain adaptation setting that uses unlabeled target-region images. When such images are not available, the model functions in a pure generalization mode. The framework also integrates a regional economic strategy module (RESM) that transforms recognition outputs and local cost information into optimized intervention intensity. Experiments show an accuracy of 91.6%, an F1-score of 89.8%, and an mAP of 88.9%, outperforming Swin Transformer and ConvNeXt; removing RESM reduces F1 to 87.2%. In cross-regional testing (Weifang training → Honghe testing), the model attains an F1 of 88.0% and mAP of 86.5%. These results indicate that integrating complementary imaging modalities with domain alignment and economic optimization provides an effective solution for disease diagnosis across greenhouse and field systems.

Accurate estimation of dynamic environmental phenomena through intelligent sensing systems plays a critical role in enabling reliable monitoring and decision-making in complex real-world scenarios. With the rapid development of artificial intelligence-driven sensing technologies and Internet of Things systems, modern agricultural monitoring is evolving from isolated data acquisition toward intelligent, multimodal perception and decision-making. However, traditional approaches predominantly rely on single data sources, making it difficult to simultaneously capture plant phenotypic variations and environment-driven mechanisms, thereby limiting model applicability in complex field scenarios. To address this issue, a multimodal pest density estimation framework, namely the Pest Density Estimation Framework (PDEF), is proposed, which integrates UAV-based imagery, trap monitoring data, and environmental sensor measurements. In this framework, crop canopy damage features are extracted using convolutional neural networks, while temporal encoding is employed to model dynamic environmental variations. Cross-modal feature alignment and environment-aware enhancement mechanisms are further introduced to achieve deep integration of multi-source information, enabling the construction of a unified feature representation space and improving estimation accuracy. Extensive experiments conducted on a constructed multimodal agricultural dataset demonstrate that the proposed method achieves MAE, RMSE, and MAPE values of $5.47$, $7.62$, and $14.9\%$, respectively, significantly outperforming the Transformer-based fusion model (MAE $6.01$, RMSE $8.16$). Meanwhile, the coefficient of determination reaches $R^2=0.84$, indicating superior fitting capability and stability. In multimodal combination experiments, the three-modality fusion reduces error metrics by more than $20\%$ on average compared with single-modality models, validating the effectiveness of multi-source collaborative modeling. From the perspective of integrating plant phenotypic analysis and environmental perception, this study provides a novel AI-driven intelligent sensing framework for pest monitoring and crop management, contributing to improved pest prediction capability and enhanced intelligence in agricultural production systems. This study further provides practical implications for agricultural economics and supply chain optimization by enabling data-driven decision-making through intelligent sensing systems.