[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-678":3,"related-678":52},{"id":4,"title":5,"url":6,"summary":7,"summary_zh":8,"content":9,"source_name":10,"source_url":8,"published_at":11,"category":12,"cover_url":8,"hotness":13,"is_selected":14,"score":15,"score_detail":16,"sources":23,"tags":25,"primary_topic":8,"secondary_topics":30,"classification":31,"search_phrases":46,"slug":49,"view_count":50,"doi":8,"paper":8,"created_at":51},678,"Quantifying Visual Symptom Severity in Plants Using Deep Learning(MDPI Remote Sens. 2026-08-17)","https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776","利用深度学习量化植物视觉症状严重程度。提出基于二元分类器训练(仅使用无症状和严重症状图像)的症状严重程度量化方法。利用具有相似症状严重程度的图像组的分类准确率和类概率估计视觉症状严重程度。标签平滑可缓解中等症状的双峰类概率分布,提高方法对单个图像的适用性。与基于特征空间距离的方法不同,本方法无需额外特征提取或距离计算。该研究为远程感知领域提供了一种简单视觉性状量化方法,可降低标注成本。",null,"## 1. Introduction\n\nPests and diseases are a severe threat to food security, trade and livelihoods globally. Every year, up to 40 percent of crops are lost due to plant pests and diseases [[1](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B1-remotesensing-18-02776)]. Therefore, plant pest and disease protection has become an important research area, due to its high correlation with food security, climate change, and environmental sustainability [[2](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B2-remotesensing-18-02776)]. Reliable, precise and accurate estimates of disease severity are important for predicting yield loss, monitoring and forecasting epidemics, and effective disease resistance [[3](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B3-remotesensing-18-02776)]. Particularly, the estimation of early symptoms for plant disease severity is one of major milestones to improve agricultural production [[4](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B4-remotesensing-18-02776)]. Visual assessment of disease severity has long been used in plant pathology. Classical ordinal scales such as the Horsfall–Barratt scale categorize severity based on the proportion of symptomatic areas relative to the total leaf area [[5](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B5-remotesensing-18-02776)]. In parallel, especially for virus diseases, symptom severity is also often categorized into qualitative classes such as, asymptomatic, mild, moderate, severe based on overall visual symptom expression [[6](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B6-remotesensing-18-02776),[7](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B7-remotesensing-18-02776)]. Although these approaches are simple and widely applicable, they are inherently subjective and sensitive to scale design and evaluator experience. To enable automation of these tasks, numerous approaches have been explored, ranging from rule-based image recognition methods to machine learning-based techniques, such as support vector machines, as well as studies integrating hyperspectral imaging. A comprehensive overview of these approaches is provided in the review by Faye et al. Recently, with advances in deep learning-based image recognition techniques, several studies have sought to automate the quantitative assessment of symptom severity by calculating the proportion of symptomatic areas within images.\n\nThomas et al. applied a support vector machine (SVM) to images captured under outdoor conditions to detect symptomatic regions and subsequently evaluated disease severity based on the proportion of symptomatic areas [[8](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B8-remotesensing-18-02776)]. Garg et al. proposed a multi-stage segmentation approach for UAV images, where background regions were first removed, and this was then followed by segmentation of symptomatic areas [[9](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B9-remotesensing-18-02776)]. Zhao et al. utilized a deep learning-based segmentation model to extract lesion regions on leaves and quantified disease severity based on the proportion of lesion area relative to the total leaf area [[10](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B10-remotesensing-18-02776)]. Palma et al. demonstrated that, when symptomatic regions are clearly distinguishable in leaf images, segmentation of disease-affected areas can be achieved using a training-free method [[11](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B11-remotesensing-18-02776)]. In contrast, Bedi et al. addressed leaf images acquired under both indoor and outdoor conditions by first training an autoencoder to extract feature vectors, classifying images into healthy and diseased classes using a feature-based classifier, and finally segmenting symptomatic regions only in images classified as diseased to estimate disease severity based on the symptomatic area [[12](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B12-remotesensing-18-02776)].\n\nMoreover, Palma et al. demonstrated that, when symptomatic regions are clearly distinguishable in leaf images, disease-affected areas can be segmented and quantified using a training-free approach [[11](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B11-remotesensing-18-02776)]. However, these approaches are applicable only to disease symptoms that involve clearly visible necrotic lesions. In addition, attempts have been made to classify symptom severity into discrete categories using plant leaf images. Xiang et al. treated disease severity estimation as a classification task by categorizing leaf images into discrete severity levels [[4](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B4-remotesensing-18-02776)]. Yang et al. proposed a unified classification approach that simultaneously identified plant type, disease, and severity [[13](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B13-remotesensing-18-02776)]. In addition to individual methodological studies, several comprehensive reviews have summarized plant disease severity assessment from a deep learning perspective. Shi et al. reviewed plant disease severity assessment methods based on convolutional neural networks (CNNs), with a focus on deep learning-driven approaches [[14](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B14-remotesensing-18-02776)]. Faye et al. provided a broader overview of plant disease severity assessment, covering rule-based image analysis methods, conventional machine learning techniques, hyperspectral imaging-based approaches, and recent advances in deep learning [[2](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B2-remotesensing-18-02776)]. Furthermore, Demilie reviewed machine learning and deep learning techniques for plant disease detection and diagnosis, including methods for assessing disease severity [[15](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B15-remotesensing-18-02776)].\n\nHowever, the implementation of these approaches requires the construction of datasets in which symptom severity is predefined and categorized, followed by supervised model training based on such labeled data. Moreover, many image datasets used in previous disease severity estimation studies were acquired under controlled conditions, such as indoor environments with uniform backgrounds and lighting, thereby limiting their applicability to outdoor settings. Furthermore, most of these studies have focused primarily on severe disease symptoms that are easily recognizable even by non-experts.\n\nPotatoes are the world’s most important root and tuber crop. One seed potato can produce approximately 10 potato tubers, and disease in one seed potato can be propagated exponentially through seed potato production [[16](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B16-remotesensing-18-02776)]. To promote stable potato production, quality seed potatoes that are healthy and disease-free should be supplied. Experienced laborers are required in the fields for visual inspection and rogueing abnormal plants during seed potato production. Rogueing is an essential practice in seed potato production because it prevents diseased plants and tubers from entering subsequent generations of seed potato multiplication. In contrast, rogueing is generally not performed in commercial potato production fields, as diseased potatoes primarily reduce yield but do not pose a risk to human health. Therefore, efficient detection of abnormal plants is particularly important in seed potato production fields, which were the target application of this study. Our previous studies developed an automated abnormal potato plant detection system using deep learning models [[17](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B17-remotesensing-18-02776),[18](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B18-remotesensing-18-02776)]. The objective of these studies is to reduce the labor required for abnormal plant rogueing tasks conducted by experts. To achieve this goal, it is necessary to detect subtle disease symptoms that can be recognized only through expert experience. Early detection and rogueing of abnormal plants at the stage of subtle symptoms is essential, as failure to do so can lead to the spread of infection to neighboring plants. Therefore, existing visual symptom severity quantification methods that focus solely on clearly severe symptoms cannot be applied to this problem. Even if a custom dataset were to be constructed and existing deep learning-based classification approaches using qualitative classes—such as asymptomatic, mild, moderate, and severe—were adopted, it would still be necessary to prepare labeled training data for each visual symptom severity category. Moreover, since the manifestation of disease symptoms varies among varieties, a large amount of labeled data would have to be prepared for each variety, rendering this approach impractical. To address these challenges, we propose two simple methods for quantifying mild visual symptoms that can be recognized only by experts, using a dataset consisting solely of clearly healthy images and clearly symptomatic images: a class-probability-based method and a classification-accuracy-based method. Although the proposed framework enables quantitative assessment of visual symptom severity at the image-group level, its direct application to individual images remains challenging. This is because models trained with cross-entropy loss are optimized to assign high confidence to the predicted class, often resulting in overconfident and poorly calibrated probabilities for mildly symptomatic images. Such behavior has been widely recognized in deep learning-based classification [[19](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B19-remotesensing-18-02776)]. Consequently, mild symptoms tend to be assigned extreme probability values close to 0 or 1, which limits their usefulness for visual symptom severity quantification at the individual image level. To address this issue, we investigate the effectiveness of probability calibration and regularization techniques, specifically temperature scaling [[19](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B19-remotesensing-18-02776)] and label smoothing [[20](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B20-remotesensing-18-02776)], which has also been reported to improve probability calibration.\n\nThe dataset used in this study is described in [Section 2](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#sec2-remotesensing-18-02776). The two proposed methods are described in [Section 3](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#sec3-remotesensing-18-02776). The results of the proposed methods are shown in [Section 4](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#sec4-remotesensing-18-02776). Discussion and conclusions are presented in [Section 5](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#sec5-remotesensing-18-02776) and [Section 6](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#sec6-remotesensing-18-02776), respectively.\n\n## 2. Materials\n\nWe acquired images to prepare the training, validation, and test sets. During the daytime in the Sapporo and Obihiro areas (Hokkaido, Japan), shooting was performed in June–July 2021 and 2022 ([Figure 1](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#fig_body_display_remotesensing-18-02776-f001)).\n\n**Figure 1.** Research field locations. This map was generated from JAXA ALOS AW3D30.\n\nWe collected RGB images using portable cameras (GoPro, GoPro, Inc., San Mateo, CA, USA). To enable efficient image capture of multiple ridges simultaneously and to mitigate issues caused by direct sunlight—such as leaf shadows cast onto neighboring leaves and image overexposure due to specular reflection—we developed and employed a custom-designed imaging cart, as shown in [Figure 2](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#fig_body_display_remotesensing-18-02776-f002). The camera was mounted in a nadir-view configuration at a distance of approximately 60 cm above the plant canopy. Images were acquired at a resolution of 4000 × 3000 pixels, corresponding to a spatial resolution of approximately 0.3 mm pixel−1. The imaging cart was operated at a speed of approximately 2 km h−1. Images were acquired between 10:00 and 15:00 under sunny or cloudy weather conditions. Training data for the deep learning model should include a large volume of images of healthy and abnormal plants. However, the viral disease acceptance criteria in Japanese seed potato production can be as low as 0.1–0.3% [[21](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B21-remotesensing-18-02776)]. Thus, based on these criteria, very few abnormal plants are present in the seed potato production field. Therefore, acquisition of abnormal plant images was an issue. To address this problem, we planted healthy seed potatoes, potato virus Y (PVY)-infected seed potatoes, and seed potatoes immersed in a suspension of Pectobacterium brasiliense (approximately 1 × 10 6 cells\u002FmL), the causal agent of blackleg disease, in Sapporo and Obihiro. PVY belongs to the genus Potyvirus. Three potato cultivars, Toyoshiro, Konahime, and Kitaakari, were included in this study ([Table 1](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#table_body_display_remotesensing-18-02776-t001)). The research fields were located in areas where no potato production fields were present nearby, thereby reducing the risk of disease transmission to other potato fields.\n\n**Figure 2.** A custom-designed imaging cart for simultaneous imaging of multiple ridges and for mitigating issues caused by direct sunlight.\n\n**Table 1.** Summary of the research fields and numbers of asymptomatic and diseased plants used in this study.\n\nImages of abnormal plants with confirmed symptoms were then captured [[17](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B17-remotesensing-18-02776)]. The target disease symptoms in this study included mosaic and wilting. Each captured image was partitioned into non-overlapping grids of 530 × 530 pixels. All grid-based image patches were visually assessed by two experienced potato rogueing specialists and categorized into five classes: asymptomatic, ambiguous (asymptomatic but difficult to judge), mild, moderate, and severe ([Figure 3](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#fig_body_display_remotesensing-18-02776-f003)). In this study, the labels assigned by these specialists were treated as the reference standard because visual inspection by certified inspectors is the current standard practice in seed potato production and forms the basis of regulatory decision-making. While most existing studies rely on clearly manifested disease symptoms that can be identified without expert knowledge ([Figure 3](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#fig_body_display_remotesensing-18-02776-f003)b), such late-stage detection is inadequate for practical rogueing operations. Since effective disease control necessitates plant removal at the earliest possible stage, this study focuses on subtle symptoms that are difficult to assess without extensive experience. To this end, we independently constructed a dataset consisting of early-stage disease images that require expert-level judgment and used it for model development and evaluation ([Figure 3](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#fig_body_display_remotesensing-18-02776-f003)a).\n\n**Figure 3.** Representative image examples. (**a**) Each visual symptom severity level for mosaic and wilting symptoms as defined in this study. (**b**) A leaf image from the PlantVillage dataset.\n\n## 3. Methods\n\n### 3.1. Proposed Methods\n\nWe previously developed a technique to visualize the features learned by deep learning-based classification models, in which the distance between the input image coordinates in the feature space and the decision boundary is utilized [[22](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B22-remotesensing-18-02776)]. By applying the same principle, it is also possible to estimate relative visual symptom severity using simpler approaches. In the feature space, ambiguous and mild symptom images tend to be distributed in the vicinity of the decision boundary, while asymptomatic and severe symptom images are distributed at greater distances from the boundary ([Figure 4](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#fig_body_display_remotesensing-18-02776-f004)). As a result, the classification probabilities for the asymptomatic and severe classes are high, whereas those for the ambiguous and mild classes are relatively low. Similarly, classification accuracy is higher for the asymptomatic and severe classes and lower for the ambiguous and mild classes. Based on this property, visual symptom severity is determined by training a binary deep learning classification model using only asymptomatic and severe images and utilizing the resulting class probabilities. In addition, when a set of images with the same level of visual symptom severity is available, visual symptom severity is also determined using the classification accuracy obtained for the corresponding image set. Accordingly, in this study, a binary classification model was first trained using only images from the asymptomatic and severe classes. The trained model was then applied to images from all five classes (asymptomatic, ambiguous, mild, moderate, and severe), and both the class probabilities and classification accuracy were computed. Using these measures, the feasibility of assessing visual symptom severity was systematically evaluated.\n\n**Figure 4.** Schematic illustration of visual symptom severity distribution in the feature space.\n\nFor model comparison, five widely used deep learning classification architectures—ResNet-50 [[23](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B23-remotesensing-18-02776)], VGG-19 [[24](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B24-remotesensing-18-02776)], Inception V3 [[20](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B20-remotesensing-18-02776)], DenseNet-121 [[25](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B25-remotesensing-18-02776)], and Vision Transformer (ViT [[26](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#B26-remotesensing-18-02776)])—were implemented using the Torchvision (v0.14.1) library and systematically evaluated. The purpose of this comparison was not to identify the best-performing architecture, but to examine whether similar trends were observed across different architectures. The models were initialized using ImageNet-pretrained weights and trained on RGB images resized to 224 × 224 pixels. No data augmentation was applied. The input images were converted to tensors and normalized using the ImageNet mean and standard deviation ([0.485, 0.456, 0.406] and [0.229, 0.224, 0.225], respectively). Training was performed using stochastic gradient descent (SGD) with a momentum of 0.9 and a cross-entropy loss function. Learning rates of 0.001, 0.005, 0.01, and 0.05 were evaluated, and the learning rate yielding the highest validation accuracy was selected. A CyclicLR scheduler (mode = “triangular2”) was used. The batch size was 128, and training was performed for 50 epochs. The model achieving the highest validation accuracy on the validation dataset was selected for subsequent evaluation.\n\nAlthough the proposed framework enables visual symptom severity quantification at the image-group level, its direct application to individual images is limited. Models trained with cross-entropy loss tend to produce overconfident predictions, assigning extreme probabilities even to mildly symptomatic images. This behavior hinders accurate visual symptom severity quantification at the individual image level. In practice, the feature distribution does not follow the ideal separation assumed by the binary classification framework. As illustrated in [Figure 5](https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F16\u002F2776#fig_body_display_remotesensing-18-02776-f005), ambiguous samples are often located close to the asymptomatic cluster, while mild and moderate samples tend to be distributed near the severely affected cluster. This overlap leads to biased probability outputs and contributes to the overconfident predictions observed in intermediate symptom stages.\n\n**Figure 5.** Schematic illustratio","MDPI Remote Sensing 2026-08-17","2026-08-17T00:00:00Z","论文",10,false,77,{"impact":17,"substance":18,"depth":19,"authority":20,"freshness":13,"relevant":21,"comment":22},15,20,18,14,1,"论文提出深度学习量化植物病害症状严重度的方法，对智慧农业病害监测有参考价值，但属于细分领域进展。",[24],{"name":10,"url":6},[26,27,28,29],"深度学习","植物病害","症状量化","标签平滑",[],{"status":32,"locked":14,"version":33,"method":34,"evidence":35},"needs_review","2026-10-22.2","rules",{"candidates":36,"agricultural_evidence":14,"reason":45},[37],{"code":38,"score":39,"supported":40,"title":41,"abstract":43,"tags":44},"I2",12,true,[42],"deep learning",[26],[26],"农业相关性需要复核",[47,48],"标签平滑 植物病害 深度学习 症状量化","标签平滑 植物病害","标签平滑植物病害深度学习症状量化-678",7,"2026-08-17T23:54:56.095013Z",{"total":53,"page":21,"page_size":53,"items":54},6,[55,124,244,308,362,400],{"id":56,"title":57,"url":58,"summary":59,"summary_zh":60,"content":8,"source_name":61,"source_url":58,"published_at":62,"category":12,"cover_url":8,"hotness":13,"is_selected":40,"score":63,"score_detail":64,"sources":67,"tags":69,"primary_topic":72,"secondary_topics":77,"classification":78,"search_phrases":94,"slug":97,"view_count":83,"doi":98,"paper":99,"created_at":123},1760,"RAVA: a robust recurrent active vision agent for plant disease diagnosis under severe environmental noise","https:\u002F\u002Fdoi.org\u002F10.1038\u002Fs41598-026-67160-2","Abstract Deep learning models have achieved near-perfect accuracy in plant disease classification within controlled laboratory settings; however, their deployment in real-world agricultural environments is severely hindered by the “deployment gap”—a critical vulnerability to environmental corruptions such as sensor noise, motion blur, and occlusion. To bridge this gap, we propose the Recurrent Active Vision Agent (RAVA), formulating the disease detection task as a Partially Observable Markov Decision Process (POMDP). Unlike passive Convolutional Neural Networks (CNNs) that process images globally, RAVA mimics the active inspection behavior of human agronomists. Our architecture integrates a lightweight ResNet-18 backbone with a Recurrent Neural Network (RNN) and a Spatial Transformer Network (STN). Driven by Proximal Policy Optimization (PPO), the agent learns a sequential policy to intelligently navigate and zoom in on informative “glimpses,” effectively bypassing background clutter. To stabilize the reinforcement learning process and enforce noise-invariant feature representations, we introduce a hybrid objective incorporating Supervised Contrastive Learning (SupCon). Comprehensive experiments on a combined PlantVillage and PlantDoc dataset demonstrate RAVA’s overwhelming superiority under extreme conditions. In a “Severe Degradation” stress test, standard ResNet-50 accuracy collapses to 26.3%, whereas our active agent maintains a robust 77.0%. Under extreme noise and occlusion, RAVA preserves 54.6% accuracy compared to the baseline’s 17.2%. Notably, this resilience is achieved with merely ∼12M parameters—significantly fewer than large-scale Vision Transformers—proving that active visual attention, coupled with contrastive learning, offers a computationally efficient and highly robust pathway for field-ready precision agriculture.","深度学习模型在受控实验室环境中进行植物病害分类时已能达到近乎完美的准确率，然而，它们在真实农业环境中的部署却因“部署鸿沟”而严重受阻——即对传感器噪声、运动模糊和遮挡等环境干扰存在关键脆弱性。为弥合这一鸿沟，我们提出了循环主动视觉智能体（RAVA），将病害检测任务构建为部分可观测马尔可夫决策过程（POMDP）。与被动地全局处理图像的卷积神经网络（CNNs）不同，RAVA模拟了人类农学家的主动检查行为。我们的架构将轻量级ResNet-18骨干网络与循环神经网络（RNN）及空间变换网络（STN）相集成。在近端策略优化（PPO）的驱动下，该智能体学习一种序列策略，以智能地导航并放大信息丰富的“瞥视”区域，从而有效避开背景杂乱干扰。为稳定强化学习过程并强制实现噪声不变的特征表示，我们引入了一种融合监督对比学习（SupCon）的混合目标函数。在结合PlantVillage和PlantDoc数据集的综合实验中，RAVA在极端条件下展现出压倒性优势。在“严重退化”压力测试中，标准ResNet-50的准确率骤降至26.3%，而我们的主动智能体则保持了稳健的77.0%。在极端噪声和遮挡条件下，RAVA保持了54.6%的准确率，而基线模型仅为17.2%。值得注意的是，这种鲁棒性仅以约1200万参数即可实现——远少于大规模视觉Transformer——这证明了主动视觉注意力结合对比学习，为面向田间部署的精准农业提供了一条计算高效且高度稳健的路径。","Scientific Reports","2026-09-04T00:00:00Z",79,{"impact":19,"substance":65,"depth":18,"authority":39,"freshness":50,"relevant":21,"comment":66},22,"提出RAVA主动视觉模型，显著提升真实环境下植物病害诊断鲁棒性，方法新颖且实验扎实。",[68],{"name":61,"url":58},[70,71,26,27],"智慧农业","农业人工智能",{"code":38,"slug":73,"name":74,"path":75,"parent_code":76},"agricultural-ai","农业AI与大模型","\u002Ftopic\u002Fdigital-agriculture\u002Fagricultural-ai","I",[],{"status":79,"locked":14,"version":33,"method":34,"evidence":80},"classified",{"candidates":81,"agricultural_evidence":40},[82,87],{"code":38,"score":83,"supported":40,"title":84,"abstract":85,"tags":86},8,[],[26,42],[71,26],{"code":88,"score":50,"supported":40,"title":89,"abstract":90,"tags":93},"I1",[],[91,92],"精准农业","precision agriculture",[70],[95,96],"农业人工智能 智慧农业 植物病害 深度学习","农业人工智能 智慧农业","农业人工智能智慧农业植物病害深度学习-1760","10.1038\u002Fs41598-026-67160-2",{"doi":98,"openalex_id":100,"authors":101,"venue":61,"cited_by_count":113,"oa_url":114,"card":115,"direction":121,"ingested_from":122},"W7208713425",[102,104,106,109,111],{"name":103,"orcid":8},"Nima Saeedi",{"name":105,"orcid":8},"Sana Gholinavaz",{"name":107,"orcid":108},"Sina Samadi Gharehveran","https:\u002F\u002Forcid.org\u002F0000-0001-5442-2485",{"name":110,"orcid":8},"Kimia Shirini",{"name":112,"orcid":8},"Adel Taheri Hajivand",0,"https:\u002F\u002Fwww.nature.com\u002Farticles\u002Fs41598-026-67160-2_reference.pdf",{"tldr":116,"method":117,"finding":118,"direction":119,"opportunity":120},"提出RAVA，一种循环主动视觉智能体，通过主动观察应对环境噪声，提升植物病害诊断鲁棒性。","结合ResNet-18、RNN、STN，用PPO训练主动策略，并加入监督对比学习","在严重退化下，RAVA准确率77.0%，远超ResNet-50的26.3%，且参数仅约12M。","农业人工智能与决策模型","可探索将主动视觉与对比学习结合，应用于其他农业任务，如害虫检测或果实计数，在复杂田间环境下提升鲁棒性。","智慧农业 \u002F 农业物联网","openalex","2026-09-06T23:30:08.817010Z",{"id":125,"title":126,"url":127,"summary":128,"summary_zh":129,"content":8,"source_name":130,"source_url":127,"published_at":131,"category":12,"cover_url":8,"hotness":13,"is_selected":14,"score":132,"score_detail":133,"sources":136,"tags":138,"primary_topic":142,"secondary_topics":148,"classification":149,"search_phrases":206,"slug":209,"view_count":113,"doi":210,"paper":211,"created_at":243},5480,"Sustainable Plant Disease Management: Current Advances, Challenges, and Future Directions","https:\u002F\u002Fdoi.org\u002F10.3390\u002Fsu182010661","Sustainable plant disease management is essential for maintaining agricultural productivity, environmental health, and global food security. Climate change has increased crop vulnerability to emerging and re-emerging plant diseases, posing significant challenges for agricultural sustainability. Although synthetic chemical pesticides have long been effective for disease control, their extensive use is increasingly questioned because of environmental contamination, pathogen resistance, and potential risks to human and ecosystem health. Consequently, plant disease management (PDM) is shifting toward more sustainable and integrated approaches. Future strategies are expected to incorporate multiple disciplines, including: (i) biological control and microbiome engineering to enhance natural plant defense mechanisms; (ii) advanced molecular and genomic technologies, such as CRISPR\u002FCas systems, RNA interference (RNAi), and epigenome editing, to develop durable disease resistance; (iii) smart delivery systems for the precise application of antimicrobial compounds and defense elicitors; (iv) nanotechnology-based platforms for the targeted delivery of nutrients, pesticides, and biocontrol agents, improving efficacy while reducing chemical inputs; (v) digital and precision agriculture technologies, including remote sensing, unmanned aerial vehicles (UAVs), and artificial intelligence (AI)-based diagnostics, to facilitate real-time disease surveillance and targeted interventions; and (vi) agroecological practices such as crop diversification, soil health restoration, and climate-smart farming to enhance agroecosystem resilience. These innovations aim not only to suppress plant pathogens but also to build sustainable, adaptive, and resilient agricultural systems that address future environmental and biological challenges.","可持续的植物病害管理对于维持农业生产力、环境健康和全球粮食安全至关重要。气候变化加剧了作物对新发和再发植物病害的脆弱性，给农业可持续性带来了重大挑战。尽管合成化学农药长期以来在病害防控中行之有效，但其广泛使用正日益受到质疑，原因包括环境污染、病原菌抗药性以及对人类和生态系统健康的潜在风险。因此，植物病害管理（plant disease management, PDM）正转向更加可持续和综合的途径。未来的策略预计将融合多个学科，包括：(i) 生物防治和微生物组工程，以增强植物天然防御机制；(ii) 先进的分子和基因组技术，如CRISPR\u002FCas系统、RNA干扰（RNAi）和表观基因组编辑，以培育持久抗病性；(iii) 用于精准施用抗菌化合物和防御诱导子的智能递送系统；(iv) 基于纳米技术的平台，用于靶向递送养分、农药和生物防治剂，在提高功效的同时减少化学投入；(v) 数字和精准农业技术，包括遥感、无人机（UAVs）和基于人工智能（AI）的诊断，以促进实时病害监测和靶向干预；以及(vi) 农业生态实践，如作物多样化、土壤健康恢复和气候智慧型农业，以增强农业生态系统的韧性。这些创新不仅旨在抑制植物病原菌，还旨在构建可持续、适应性强且具有韧性的农业系统，以应对未来的环境和生物挑战。","Sustainability","2026-10-20T00:00:00Z",81,{"impact":65,"substance":18,"depth":19,"authority":134,"freshness":83,"relevant":21,"comment":135},13,"系统综述可持续植物病害管理六大技术方向，涵盖AI诊断、纳米递送与基因编辑，对智慧农业与绿色防控具参考价值。",[137],{"name":130,"url":127},[70,91,139,27,140,141],"基因编辑","生物防治","纳米技术",{"code":143,"slug":144,"name":145,"path":146,"parent_code":147},"F4","crop-protection","农药与绿色防控","\u002Ftopic\u002Finputs-equipment\u002Fcrop-protection","F",[],{"status":79,"locked":14,"version":33,"method":34,"evidence":150},{"candidates":151,"agricultural_evidence":40},[152,156,161,168,175,182,189,196,201],{"code":88,"score":83,"supported":40,"title":153,"abstract":154,"tags":155},[],[91,92],[70,91],{"code":143,"score":50,"supported":40,"title":157,"abstract":158,"tags":160},[],[159,140],"农药",[140],{"code":162,"score":53,"supported":40,"title":163,"abstract":164,"tags":167},"A8",[],[165,166],"粮食安全","food security",[],{"code":169,"score":53,"supported":40,"title":170,"abstract":171,"tags":174},"H2",[],[172,173],"土壤","soil",[],{"code":176,"score":53,"supported":40,"title":177,"abstract":178,"tags":181},"H7",[],[179,180],"气候变化","climate change",[],{"code":183,"score":53,"supported":40,"title":184,"abstract":185,"tags":188},"I3",[],[186,187],"遥感","remote sensing",[],{"code":190,"score":191,"supported":40,"title":192,"abstract":193,"tags":195},"H6",3,[],[194],"农业生态",[],{"code":38,"score":191,"supported":40,"title":197,"abstract":198,"tags":200},[],[199],"artificial intelligence",[],{"code":202,"score":21,"supported":14,"title":203,"abstract":204,"tags":205},"F2",[],[],[139],[207,208],"CRISPR RNAi 作物抗病","基因编辑 智慧农业 植物病害 生物防治","CRISPRRNAi作物抗病-5480","10.3390\u002Fsu182010661",{"doi":210,"openalex_id":212,"authors":213,"venue":130,"cited_by_count":113,"oa_url":127,"card":238,"direction":121,"ingested_from":122},"W7223044776",[214,217,220,223,226,229,232,235],{"name":215,"orcid":216},"Karim M. Hassan","https:\u002F\u002Forcid.org\u002F0000-0002-1241-2307",{"name":218,"orcid":219},"Basma T. Abd-Elhalim","https:\u002F\u002Forcid.org\u002F0000-0002-3009-332X",{"name":221,"orcid":222},"Mohamed K. Abou El-Nasr","https:\u002F\u002Forcid.org\u002F0000-0003-1451-833X",{"name":224,"orcid":225},"Essam Yosef Abdul-Hafeez","https:\u002F\u002Forcid.org\u002F0000-0003-4744-4672",{"name":227,"orcid":228},"Ahmed N. Abdelhamid","https:\u002F\u002Forcid.org\u002F0000-0002-1326-7885",{"name":230,"orcid":231},"Mostafa Abdelkader","https:\u002F\u002Forcid.org\u002F0000-0003-2339-5087",{"name":233,"orcid":234},"Mohamed Abd El-Hamed Nasser","https:\u002F\u002Forcid.org\u002F0000-0002-0328-2969",{"name":236,"orcid":237},"Mahmoud A. A. Ali","https:\u002F\u002Forcid.org\u002F0000-0002-9062-0048",{"tldr":239,"method":240,"finding":241,"direction":121,"opportunity":242},"综述植物病害可持续管理的多学科前沿进展、挑战与未来方向。","文献综述，整合生物防治、基因编辑、纳米递送、数字农业与生态农艺。","未来病害管理需融合生物、分子、纳米与数字技术构建韧性农业系统。","遥感、无人机与AI诊断的实时病害监测闭环仍缺田间验证与低成本落地研究。","2026-10-21T23:30:09.242484Z",{"id":245,"title":246,"url":247,"summary":248,"summary_zh":249,"content":8,"source_name":250,"source_url":247,"published_at":251,"category":12,"cover_url":8,"hotness":13,"is_selected":14,"score":252,"score_detail":253,"sources":256,"tags":258,"primary_topic":261,"secondary_topics":267,"classification":269,"search_phrases":288,"slug":291,"view_count":113,"doi":292,"paper":293,"created_at":307},5387,"LIGHTWEIGHT FEATURE FUSED ATTENTION DRIVEN MULTI LEVEL HYBRID MOBILENET V2 FRAMEWORK FOR EARLY TOMATO LEAF DISEASE DETECTION","https:\u002F\u002Fdoi.org\u002F10.62441\u002Factainnovations.vi.1918","It is crucial to find and treat tomato crop diseases early and accurately to improve crop efficiency and make sure there is no scarcity of crop. However, computerized analysis is provocative due to lighting differences, compound backgrounds, and graphic likenesses among infection regions. Research work proposed a lightweight Attention Driven Multi Level Hybrid MobileNet-V2 algorithm that fuses the handcrafted texture structures with Deep Learning (DL) features, then image preprocessing is incorporated using Gaussian, Bilateral, and Guided filters to minimize noise by preserving the disease lesions. LAB color model along with Contrast Limited Adaptive Histogram Equalization (CLAHE) are used to improve the contrast. Otsu thresholding is involved to effectively separate the leaf segment from diseased parts. Local Binary Pattern (LBP) and Gray Level Co-occurrence Matrix (GLCM) descriptors are used to obtain fine-grained features, while high level features are gained by pretrained MobileNet-V2 model. Features fusion is fed to the fully connected layers of deep learning model for training and validation purpose. Results have revealed that the suggested model attains better classification performance as compared to existing models, making it an effective explanation for real time tomato leaf disease detection for smart agriculture applications.","早期准确发现并治疗番茄作物病害，对于提高作物效率、确保作物不短缺至关重要。然而，由于光照差异、复杂背景以及感染区域之间的图形相似性，计算机化分析具有挑战性。研究提出了一种轻量级注意力驱动多层混合MobileNet-V2算法，该算法将手工纹理结构与深度学习（DL）特征相融合，然后结合高斯、双边和引导滤波器进行图像预处理，以在保留病害病变的同时最小化噪声。采用LAB颜色模型和对比度受限自适应直方图均衡化（CLAHE）来提高对比度。引入Otsu阈值分割以有效分离叶片片段与病害部分。使用局部二值模式（LBP）和灰度共生矩阵（GLCM）描述符获取细粒度特征，同时通过预训练的MobileNet-V2模型获取高层特征。特征融合后输入深度学习模型的全连接层进行训练和验证。结果表明，与现有模型相比，所提出的模型取得了更好的分类性能，使其成为智慧农业应用中实时番茄叶片病害检测的有效解决方案。","Acta Innovations","2026-10-18T00:00:00Z",64,{"impact":39,"substance":19,"depth":254,"authority":13,"freshness":83,"relevant":21,"comment":255},16,"提出融合手工纹理特征与MobileNet-V2的轻量番茄叶病检测方法，有一定方法新颖性，但属单篇论文、应用范围有限，未达每日精选门槛。",[257],{"name":250,"url":247},[70,71,26,259,260],"番茄病害","图像识别",{"code":262,"slug":263,"name":264,"path":265,"parent_code":266},"B1","vegetables","蔬菜","\u002Ftopic\u002Fhorticulture\u002Fvegetables","B",[268],{"code":38,"slug":73,"name":74,"path":75,"parent_code":76},{"status":79,"locked":14,"version":33,"method":34,"evidence":270},{"candidates":271,"agricultural_evidence":40},[272,278,283],{"code":262,"score":17,"supported":40,"title":273,"abstract":275,"tags":277},[274],"tomato",[276,274],"番茄",[259],{"code":38,"score":279,"supported":40,"title":280,"abstract":281,"tags":282},9,[],[26,42],[71,260,26],{"code":88,"score":50,"supported":40,"title":284,"abstract":285,"tags":287},[],[70,286],"smart agriculture",[70],[289,290],"MobileNet-V2 番茄叶片病害","番茄叶病 深度学习 检测","MobileNet-V2番茄叶片病害-5387","10.62441\u002Factainnovations.vi.1918",{"doi":292,"openalex_id":294,"authors":295,"venue":250,"cited_by_count":113,"oa_url":301,"card":302,"direction":121,"ingested_from":122},"W7222977034",[296,298],{"name":297,"orcid":8},"Mrs. Vijetalaxmi A Hegde",{"name":299,"orcid":300},"Abhay Ramakant Kasetwar","https:\u002F\u002Forcid.org\u002F0000-0002-5800-8764","https:\u002F\u002Factainnovations.com\u002Findex.php\u002Fpub\u002Farticle\u002Fdownload\u002F1918\u002F981",{"tldr":303,"method":304,"finding":305,"direction":119,"opportunity":306},"提出轻量注意力多级混合MobileNet-V2框架，融合手工纹理与深度学习特征实现番茄叶片病害早期检","高斯\u002F双边\u002F引导滤波去噪，LAB+CLAHE增强，Otsu分割，LBP与GLCM","所提模型分类性能优于现有模型，可有效用于智慧农业实时番茄叶片病害检测。","可探索多作物跨域泛化、田间复杂光照下轻量模型部署与可解释性，及病害早期分级检测。","2026-10-20T23:30:10.424511Z",{"id":309,"title":310,"url":311,"summary":312,"summary_zh":313,"content":8,"source_name":314,"source_url":311,"published_at":251,"category":12,"cover_url":8,"hotness":13,"is_selected":14,"score":315,"score_detail":316,"sources":318,"tags":320,"primary_topic":323,"secondary_topics":324,"classification":325,"search_phrases":347,"slug":350,"view_count":113,"doi":351,"paper":352,"created_at":361},5385,"A Comprehensive Review of Deep Learning and Computer Vision Techniques for Fruit Quality Assessment and Classification","https:\u002F\u002Fdoi.org\u002F10.25258\u002Fijddt.16.75s.151","Fruit image analysis has become an essential research area in precision agriculture, enabling automated fruit classification, quality grading, maturity estimation, disease diagnosis, yield prediction, and intelligent harvesting through advances in computer vision and artificial intelligence (AI).Recent progress in imaging technologies, large-scale benchmark datasets, and deep learning has substantially improved the accuracy, robustness, and scalability of automated fruit analysis systems.This survey presents a comprehensive and systematic review of image processing and AI techniques for fruit image analysis.It covers the complete image analysis pipeline, including image acquisition, preprocessing, enhancement, color space transformation, segmentation, feature extraction, and handcrafted texture descriptors.Furthermore, conventional machine learning algorithms are reviewed alongside state-of-the-art deep learning architectures, including Convolutional Neural Networks (CNNs), Vision Transformers (ViTs), and Vision Mamba models, for applications such as fruit recognition, quality assessment, disease detection, ripeness estimation, and object detection.Publicly available benchmark datasets, evaluation metrics, and performance comparison strategies are also summarized to facilitate objective assessment of existing approaches.In addition, the survey discusses key research challenges, including illumination variations, occlusion, complex orchard environments, limited annotated datasets, computational complexity, and domain adaptation, while highlighting emerging research trends such as self-supervised learning, multimodal learning, explainable AI, lightweight deep networks, foundation models, and edge intelligence for real-time deployment.By providing a unified review of methodologies, datasets, evaluation protocols, challenges, and future research directions, this survey offers a comprehensive reference for researchers and practitioners developing intelligent and scalable fruit image analysis systems for precision agriculture.","水果图像分析已成为精准农业中的一个重要研究领域，借助计算机视觉和人工智能（AI）的进步，实现了果实自动分类、品质分级、成熟度估计、病害诊断、产量预测和智能采收。成像技术、大规模基准数据集和深度学习的最新进展显著提升了自动化水果分析系统的准确性、鲁棒性和可扩展性。本综述对用于水果图像分析的图像处理和AI技术进行了全面而系统的回顾。内容涵盖完整的图像分析流程，包括图像采集、预处理、增强、颜色空间变换、分割、特征提取和手工纹理描述符。此外，综述了传统机器学习算法以及最先进的深度学习架构，包括卷积神经网络（CNN）、视觉Transformer（ViT）和视觉Mamba模型，用于果实识别、品质评估、病害检测、成熟度估计和目标检测等应用。同时总结了公开可用的基准数据集、评价指标和性能比较策略，以便对现有方法进行客观评估。此外，本综述讨论了关键研究挑战，包括光照变化、遮挡、复杂果园环境、标注数据集有限、计算复杂度和域适应，并重点介绍了新兴研究趋势，如自监督学习、多模态学习、可解释AI、轻量级深度网络、基础模型和面向实时部署的边缘智能。通过提供方法、数据集、评价协议、挑战和未来研究方向的统一综述，本综述为开发面向精准农业的智能且可扩展的水果图像分析系统的研究人员和实践者提供了全面的参考。","International Journal of Drug Delivery Technology",56,{"impact":83,"substance":19,"depth":254,"authority":53,"freshness":83,"relevant":21,"comment":317},"系统性综述覆盖CNN、ViT、Vision Mamba等前沿方法，信息量足，但期刊权威性一般且偏学术综述，公共价值有限，适合主题聚合而非每日精选头条。",[319],{"name":314,"url":311},[70,71,321,26,322],"计算机视觉","水果品质检测",{"code":38,"slug":73,"name":74,"path":75,"parent_code":76},[],{"status":79,"locked":14,"version":33,"method":34,"evidence":326},{"candidates":327,"agricultural_evidence":40},[328,337,341],{"code":38,"score":329,"supported":40,"title":330,"abstract":332,"tags":336},28,[42,331],"computer vision",[333,26,334,321,335,42,199,331],"目标检测","机器学习","machine learning",[71,26,321],{"code":88,"score":50,"supported":40,"title":338,"abstract":339,"tags":340},[],[91,92],[70],{"code":342,"score":191,"supported":40,"title":343,"abstract":344,"tags":346},"B2",[],[345],"orchard",[],[348,349],"水果品质 计算机视觉 深度学习","农业人工智能 水果品质检测 计算机视觉 智慧农业","水果品质计算机视觉深度学习-5385","10.25258\u002Fijddt.16.75s.151",{"doi":351,"openalex_id":353,"authors":354,"venue":314,"cited_by_count":113,"oa_url":311,"card":355,"direction":121,"ingested_from":122},"W7222892831",[],{"tldr":356,"method":357,"finding":358,"direction":359,"opportunity":360},"系统综述深度学习与计算机视觉在水果品质评估与分级中的全流程方法。","综述图像采集到分割、特征提取及CNN、ViT、Vision Mamba等模型与公","深度学习显著提升水果分析精度与鲁棒性，但光照、遮挡和小样本仍是主要瓶颈。","农业遥感与作物表型","可探索自监督与轻量化模型在复杂果园遮挡、小样本场景下的边缘实时部署。","2026-10-20T23:30:10.066042Z",{"id":363,"title":364,"url":365,"summary":366,"summary_zh":367,"content":8,"source_name":314,"source_url":365,"published_at":368,"category":12,"cover_url":8,"hotness":13,"is_selected":14,"score":113,"score_detail":369,"sources":371,"tags":373,"primary_topic":8,"secondary_topics":374,"classification":375,"search_phrases":382,"slug":385,"view_count":113,"doi":386,"paper":387,"created_at":399},5242,"Enhanced Convolutional Neural Network Architecture with Attention-Guided Feature Refinement for Robust Image Classification","https:\u002F\u002Fdoi.org\u002F10.25258\u002Fijddt.16.88s.43","Image classification remains a foundational problem in computer vision, underpinning applications that range from medical diagnostics to autonomous navigation. Although deep convolutional neural networks (CNNs) have substantially advanced classification accuracy over the last decade, standard architectures such as VGG, ResNet, and plain CNNs often struggle to jointly optimize accuracy, parameter efficiency, and robustness to intra-class variation. This paper proposes an AttentionGuided Refinement Network (AGR-Net), a CNN architecture that integrates a lightweight channel-and-spatial attention module with residual feature refinement blocks to improve discriminative feature learning while maintaining a compact parameter footprint. The proposed model is evaluated on the CIFAR-10 benchmark dataset using a rigorous experimental protocol that includes data augmentation, learning-rate scheduling, and five-fold cross-validation. Comparative experiments against baseline architectures, including a plain CNN, VGG-16, ResNet-18, and MobileNetV2, demonstrate that AGR-Net achieves a top-1 accuracy of 94.6%, outperforming the strongest baseline by 1.8 percentage points while using 27% fewer parameters than ResNet-18. Ablation studies confirm that the attention module and refinement blocks each contribute measurably to the accuracy gain. The results indicate that attention-guided refinement offers a favorable trade-off between accuracy, efficiency, and robustness, making the proposed architecture suitable for deployment in resource-constrained and real-time image classification systems. The paper further discusses the advantages, practical applications, limitations, and future research directions of the proposed approach.","图像分类仍然是计算机视觉中的一个基础性问题，支撑着从医学诊断到自主导航等诸多应用。尽管深度卷积神经网络（CNN）在过去十年中大幅提升了分类准确率，但VGG、ResNet和普通CNN等标准架构往往难以在准确率、参数效率和对类内变化的鲁棒性之间实现联合优化。本文提出了一种注意力引导细化网络（AttentionGuided Refinement Network，AGR-Net），这是一种CNN架构，它将轻量级的通道与空间注意力模块与残差特征细化模块相结合，以改进判别性特征学习，同时保持紧凑的参数规模。所提出的模型在CIFAR-10基准数据集上使用严格的实验方案进行评估，该方案包括数据增强、学习率调度和五折交叉验证。与基线架构（包括普通CNN、VGG-16、ResNet-18和MobileNetV2）的对比实验表明，AGR-Net实现了94.6%的top-1准确率，比最强基线高出1.8个百分点，同时使用的参数比ResNet-18少27%。消融研究证实，注意力模块和细化模块各自对准确率提升都有可测量的贡献。结果表明，注意力引导细化在准确率、效率和鲁棒性之间提供了有利的权衡，使所提出的架构适合部署在资源受限和实时的图像分类系统中。本文进一步讨论了所提出方法的优势、实际应用、局限性和未来研究方向。","2026-10-17T00:00:00Z",{"impact":113,"substance":113,"depth":113,"authority":113,"freshness":113,"relevant":113,"comment":370},"该文为通用计算机视觉论文，未涉及农业场景，与三农及农业信息化无关，不宜入选。",[372],{"name":314,"url":365},[71,26,260],[],{"status":32,"locked":14,"version":33,"method":34,"evidence":376},{"candidates":377,"agricultural_evidence":14,"reason":45},[378],{"code":38,"score":279,"supported":40,"title":379,"abstract":380,"tags":381},[],[321,331],[71,260,26],[383,384],"农业人工智能 图像识别 深度学习","农业人工智能 图像识别","农业人工智能图像识别深度学习-5242","10.25258\u002Fijddt.16.88s.43",{"doi":386,"openalex_id":388,"authors":389,"venue":314,"cited_by_count":113,"oa_url":365,"card":394,"direction":121,"ingested_from":122},"W7222796569",[390,392],{"name":391,"orcid":8},"KRISHNA SANDEEP JONNALA",{"name":393,"orcid":8},"K VENKAT RAJU",{"tldr":395,"method":396,"finding":397,"direction":119,"opportunity":398},"提出注意力引导细化网络AGR-Net，在CIFAR-10上兼顾精度与参数效率。","轻量通道-空间注意力模块+残差特征细化块，五折交叉验证对比基线。","AGR-Net达94.6%准确率，超最强基线1.8个百分点且参数比ResNet-18少27%。","可将该轻量注意力架构迁移至作物病害识别等农业图像任务，验证田间复杂场景鲁棒性。","2026-10-18T23:30:13.674128Z",{"id":401,"title":402,"url":403,"summary":404,"summary_zh":405,"content":8,"source_name":406,"source_url":403,"published_at":368,"category":12,"cover_url":8,"hotness":13,"is_selected":14,"score":407,"score_detail":408,"sources":410,"tags":412,"primary_topic":415,"secondary_topics":421,"classification":422,"search_phrases":439,"slug":442,"view_count":21,"doi":443,"paper":444,"created_at":458},5234,"A smart agriculture-oriented cross-attention multiscale CNN for sugarcane disease diagnosis","https:\u002F\u002Fdoi.org\u002F10.1080\u002F23080477.2026.2749739","Sugarcane leaf diseases can adversely affect crop health and yield, making timely disease identification important for smart agriculture. Manual inspection of leaf diseases is time-consuming and requires expert knowledge, particularly when large cultivation areas need to be monitored. To tackle this problem, this study proposes a Cross-Attention Multiscale Parallel Convolutional Neural Network for automatic classification of sugarcane leaf diseases. This framework comprises a Parallel Feature Extraction Stem, cross-attention mechanism, Dense Connection Blocks, and Inception Blocks to learn disease features at various spatial scales. The PFES extracts fine and broad disease patterns using parallel convolutional branches, while the cross-attention mechanism refines the extracted features by emphasizing disease-related information. The DCB supports feature propagation and reuse, and the IB further captures multi-scale feature representations. The proposed model was tested on two benchmark datasets of leaf diseases for sugarcane plants. The proposed framework obtained an accuracy (98.79%), precision (98.31%), recall (97.81%), specificity (98.60%), and F1-Score (98.05%). For assessing the classification accuracy and consistency of the framework, the comparative study, ablation studies, and five-fold cross-validation are carried out. The results show that CA-MPCNN provides accurate sugarcane leaf disease classification and can support automated disease diagnosis in smart agriculture applications.","甘蔗叶片病害会对作物健康与产量造成不利影响，因此及时识别病害对智慧农业具有重要意义。人工检查叶片病害耗时且需要专业知识，尤其是在需要监测大面积种植区域时。为解决该问题，本研究提出一种用于甘蔗叶片病害自动分类的交叉注意力多尺度并行卷积神经网络（Cross-Attention Multiscale Parallel Convolutional Neural Network）。该框架由并行特征提取主干（Parallel Feature Extraction Stem）、交叉注意力机制、密集连接块（Dense Connection Blocks）和Inception块组成，用于学习不同空间尺度上的病害特征。PFES通过并行卷积分支提取细粒度与广域病害模式，而交叉注意力机制通过强调病害相关信息对提取的特征进行细化。DCB支持特征传播与复用，IB进一步捕获多尺度特征表示。所提出的模型在两个甘蔗植株叶片病害基准数据集上进行了测试。该框架取得了准确率（98.79%）、精确率（98.31%）、召回率（97.81%）、特异性（98.60%）和F1分数（98.05%）。为评估该框架的分类准确性和一致性，开展了对比研究、消融研究和五折交叉验证。结果表明，CA-MPCNN能够实现准确的甘蔗叶片病害分类，并可支持智慧农业应用中的自动化病害诊断。","Smart Science",69,{"impact":39,"substance":18,"depth":254,"authority":39,"freshness":279,"relevant":21,"comment":409},"该研究提出一种面向甘蔗叶部病害自动分类的交叉注意力多尺度CNN，在两个基准数据集上准确率达98.79%，方法新颖、实验充分，但属于细分作物应用研究，产业影响有限，时效性强，值得进入每日精选。",[411],{"name":406,"url":403},[70,71,413,26,414],"甘蔗","病害诊断",{"code":416,"slug":417,"name":418,"path":419,"parent_code":420},"A6","cash-crops","棉麻糖与特色经作","\u002Ftopic\u002Ffield-crops\u002Fcash-crops","A",[],{"status":79,"locked":14,"version":33,"method":34,"evidence":423},{"candidates":424,"agricultural_evidence":40},[425,430,434],{"code":416,"score":17,"supported":40,"title":426,"abstract":428,"tags":429},[427],"sugarcane",[413,427],[413],{"code":88,"score":17,"supported":40,"title":431,"abstract":432,"tags":433},[286],[70,286],[70],{"code":38,"score":435,"supported":14,"title":436,"abstract":437,"tags":438},2,[],[],[71,26],[440,441],"甘蔗 叶部病害 智能诊断","CA-MPCNN 甘蔗 病害","甘蔗叶部病害智能诊断-5234","10.1080\u002F23080477.2026.2749739",{"doi":443,"openalex_id":445,"authors":446,"venue":406,"cited_by_count":113,"oa_url":8,"card":453,"direction":121,"ingested_from":122},"W7222832245",[447,449,451],{"name":448,"orcid":8},"Bala Subramanian Chokkalingam",{"name":450,"orcid":8},"Uvarani Meenrajan",{"name":452,"orcid":8},"Shanmugapriya Sankaran",{"tldr":454,"method":455,"finding":456,"direction":119,"opportunity":457},"提出交叉注意力多尺度并行CNN，实现甘蔗叶病自动分类。","并行卷积分支、交叉注意力、密集连接与Inception块，两个基准数据集验证。","模型准确率达98.79%，消融与五折交叉验证证明其有效稳定。","可探索田间复杂背景与轻量化部署，并扩展到多作物病害实时诊断。","2026-10-18T23:30:07.777036Z"]