[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-5558":3,"related-5558":123},{"id":4,"title":5,"url":6,"summary":7,"summary_zh":8,"content":9,"source_name":10,"source_url":6,"published_at":11,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":15,"score_detail":16,"sources":23,"tags":25,"primary_topic":31,"secondary_topics":37,"classification":50,"search_phrases":96,"slug":99,"view_count":100,"doi":101,"paper":102,"created_at":122},5558,"Estimation of Soil Water Content in Winter Wheat Field at Jointing Stage Using Multi-Source Features and Machine Learning","https:\u002F\u002Fdoi.org\u002F10.3390\u002Fagronomy16202071","Accurate and timely monitoring of soil water content (SWC) plays an essential supporting role in crop irrigation management and smart agricultural practice. Remote sensing technology provides rapid and convenient approaches to large-scale farmland SWC estimation or inversion. However, existing mechanistic modeling methods are limited by parameter localization and land surface heterogeneity, while machine learning data-driven methods lack quite good physical interpretability. To address deficiencies in single data-driven methods and single vegetation microwave scattering models, and explore the estimation effect of different feature variables in machine learning. This study proposes an innovative estimation model that integrates machine learning strategies and multi-source features, including spectral indices and features derived from the classic water cloud model (WCM). The novelty of the proposed model lies in balancing physical interpretability and data-driven accuracy, namely extracting relevant soil backscattering parameters through WCM, which are then used as input features in the machine learning. A combination of field-measured data and unmanned aerial vehicle (UAV) multi-source remote sensing data, including multispectral and synthetic aperture radar (SAR) data, was acquired synchronously at the jointing stage of winter wheat. Through extracting variables such as the SAR microwave backscattering coefficient and the optical vegetation index, and conducting feature screening and parameter sensitivity analysis, we optimized the input variable sets of models based on back-propagation (BP) neural network, support vector regression (SVR) and random forest (RF) algorithms. The study results showed that the established estimation model had excellent estimation performance, especially for the RF algorithm, with the highest R2 of 0.855, lowest MAE of 0.0212 cm3\u002Fcm3, and lowest RMSE of 0.0253 cm3\u002Fcm3. It was also shown that different feature input schemes have obvious effects on SWC estimation performance. On the whole, RF had the best estimation accuracy, followed by SVR and then BPNN. This study demonstrated that models integrating multi-source remote sensing data and machine learning could enhance SWC estimation ability for winter wheat fields, and could also serve as a promising approach for SWC monitoring in winter wheat fields in the North China Plain.","准确、及时地监测土壤含水量（SWC）对作物灌溉管理和智慧农业实践具有重要的支撑作用。遥感技术为大规模农田土壤含水量估算或反演提供了快速便捷的手段。然而，现有机理建模方法受参数本地化和地表异质性限制，机器学习数据驱动方法则缺乏良好的物理解释性。针对单一数据驱动方法和单一植被微波散射模型的不足，并探究不同特征变量在机器学习中的估算效果，本研究提出了一种融合机器学习策略与多源特征的创新估算模型，所采用特征包括光谱指数和经典水云模型（WCM）衍生的特征。该模型的新颖性在于平衡物理解释性与数据驱动精度，即通过WCM提取相关土壤后向散射参数，再将其作为机器学习的输入特征。研究在冬小麦拔节期同步获取了田间实测数据与无人机（UAV）多源遥感数据，包括多光谱和合成孔径雷达（SAR）数据。通过提取SAR微波后向散射系数和光学植被指数等变量，并进行特征筛选和参数敏感性分析，优化了基于反向传播（BP）神经网络、支持向量回归（SVR）和随机森林（RF）算法的模型输入变量集。研究结果表明，所建立的估算模型具有优良的估算性能，尤其是RF算法，最高R²为0.855，最低MAE为0.0212 cm³\u002Fcm³，最低RMSE为0.0253 cm³\u002Fcm³。研究还表明，不同特征输入方案对土壤含水量估算性能有明显影响。总体而言，RF估算精度最佳，SVR次之，BPNN再次之。本研究证明，融合多源遥感数据与机器学习的模型能够提升冬小麦田土壤含水量估算能力，也可为华北平原冬小麦田土壤含水量监测提供一种有前景的方法。",null,"Agronomy","2026-10-21T00:00:00Z","论文",10,false,77,{"impact":17,"substance":18,"depth":19,"authority":20,"freshness":13,"relevant":21,"comment":22},16,21,17,13,1,"该研究融合多源遥感与机器学习估算冬小麦拔节期土壤含水量，方法有创新、结论可靠，对智慧灌溉有参考价值，但属细分领域论文，影响范围有限。",[24],{"name":10,"url":6},[26,27,28,29,30],"智慧农业","机器学习","遥感监测","土壤墒情","冬小麦",{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},"A2","wheat","小麦","\u002Ftopic\u002Ffield-crops\u002Fwheat","A",[38,44],{"code":39,"slug":40,"name":41,"path":42,"parent_code":43},"H2","soil","土壤健康与改良","\u002Ftopic\u002Fenvironment\u002Fsoil","H",{"code":45,"slug":46,"name":47,"path":48,"parent_code":49},"I2","agricultural-ai","农业AI与大模型","\u002Ftopic\u002Fdigital-agriculture\u002Fagricultural-ai","I",{"status":51,"locked":14,"version":52,"method":53,"evidence":54},"classified","2026-10-22.2","rules",{"candidates":55,"agricultural_evidence":58},[56,63,69,74,82,90],{"code":32,"score":57,"supported":58,"title":59,"abstract":61,"tags":62},26,true,[60,33],"winter wheat",[34,60,33],[30],{"code":39,"score":64,"supported":58,"title":65,"abstract":66,"tags":68},15,[40],[67,40],"土壤",[29],{"code":45,"score":64,"supported":58,"title":70,"abstract":72,"tags":73},[71],"machine learning",[27,71],[27],{"code":75,"score":13,"supported":58,"title":76,"abstract":77,"tags":81},"I3",[],[78,79,80],"遥感","remote sensing","multispectral",[28],{"code":83,"score":84,"supported":58,"title":85,"abstract":86,"tags":89},"H3",6,[],[87,88],"灌溉","irrigation",[],{"code":91,"score":92,"supported":58,"title":93,"abstract":94,"tags":95},"I1",4,[],[26],[26],[97,98],"冬小麦 土壤含水量 遥感","多源遥感 机器学习 墒情","冬小麦土壤含水量遥感-5558",0,"10.3390\u002Fagronomy16202071",{"doi":101,"openalex_id":103,"authors":104,"venue":10,"cited_by_count":100,"oa_url":6,"card":114,"direction":120,"ingested_from":121},"W7223156204",[105,108,110,112],{"name":106,"orcid":107},"Mingliang Jiang","https:\u002F\u002Forcid.org\u002F0000-0002-6486-6087",{"name":109,"orcid":9},"Yi Zhang",{"name":111,"orcid":9},"WEI Lei",{"name":113,"orcid":9},"Chengcai Zhang",{"tldr":115,"method":116,"finding":117,"direction":118,"opportunity":119},"融合WCM物理特征与多源遥感，用机器学习估算冬小麦拔节期土壤含水量。","无人机多光谱与SAR数据，WCM参数+光谱指数，BP\u002FSVR\u002FRF建模。","RF最优，R²达0.855，RMSE为0.0253 cm³\u002Fcm³，特征方案影响显著。","农业遥感与作物表型","可探索物理约束与深度学习耦合，并迁移至多生育期、多作物及不同土壤质地验证。","智慧农业 \u002F 农业物联网","openalex","2026-10-22T11:20:18.887301Z",{"total":84,"page":21,"page_size":84,"items":124},[125,206,270,325,395,445],{"id":126,"title":127,"url":128,"summary":129,"summary_zh":130,"content":9,"source_name":131,"source_url":128,"published_at":132,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":133,"score_detail":134,"sources":140,"tags":142,"primary_topic":145,"secondary_topics":148,"classification":151,"search_phrases":185,"slug":188,"view_count":189,"doi":190,"paper":191,"created_at":205},3804,"Evaluating Reference Evapotranspiration and Soil Moisture as Predictors for Machine Learning-Based Actual Evapotranspiration Estimation: Model Performance and Explainability","https:\u002F\u002Fdoi.org\u002F10.3390\u002Fatmos17100941","Accurate estimation of actual evapotranspiration (ETa) is essential for sustainable water resources management and agricultural planning. This study investigated how the availability of reference evapotranspiration (ET0) and soil moisture (SM) affects ETa prediction performance and predictor importance using Random Forest (RF), Bagging Trees (BT), Least Squares Boosting (LSBoost), Generalized Additive Models (GAM), and Multiple Linear Regression (MLR). Global solar radiation (Rs), land surface temperature (LST), normalized difference vegetation index (NDVI), and SM were evaluated under two primary scenarios, with and without ET0. Models were developed using 80% of the data, with five-fold cross-validation conducted within the training subset, and final performance was evaluated on an independent 20% test subset. Model interpretability was assessed using SHAP, Permutation Feature Importance (PFI), Partial Dependence Plot (PDP), and temporal signed-SHAP analyses, while an additional SM-ablation experiment quantified the incremental predictive contribution of soil moisture. In Scenario I, ET0 was the dominant predictor, and RF and BT achieved the highest independent-test performance (R2 = 0.875; RMSE = 0.398–0.399 mm day−1). Excluding ET0 reduced predictive performance, with RF, LSBoost, and BT achieving R2 values of approximately 0.80 in Scenario II, while the predictor-importance structure shifted primarily toward Rs, followed by SM, LST, and NDVI. Removal of SM further reduced performance in both scenarios, with substantially greater deterioration when ET0 was unavailable: R2 decreased by 0.025–0.044 in Scenario I and by 0.051–0.099 in Scenario II. Under the SM-excluded Scenario I configuration, which more closely aligned the predictor information with previous ANN- and SAFER-based studies, R2 values ranged from 0.820 to 0.832, indicating that differences in ETa prediction performance cannot be attributed solely to model structure but also depend strongly on predictor information content. Overall, ET0 provides substantial atmospheric-demand information for ETa prediction, whereas Rs assumes the dominant predictive role, and SM provides particularly important complementary information when ET0 is unavailable. The combined performance, ablation, and explainability analyses demonstrate the importance of considering predictor composition together with model structure when developing and comparing ETa estimation approaches.","准确估算实际蒸散发（ETa）对于可持续水资源管理和农业规划至关重要。本研究探讨了参考蒸散发（ET0）和土壤水分（SM）的可用性如何影响ETa预测性能及预测因子重要性，采用了随机森林（RF）、Bagging Trees（BT）、最小二乘提升（LSBoost）、广义加性模型（GAM）和多元线性回归（MLR）方法。在全球太阳辐射（Rs）、地表温度（LST）、归一化差异植被指数（NDVI）和SM的条件下，评估了包含和不包含ET0两种主要情景。模型使用80%的数据进行开发，在训练子集内进行五折交叉验证，最终性能在独立的20%测试子集上评估。模型可解释性通过SHAP、排列特征重要性（PFI）、偏依赖图（PDP）和时间符号SHAP分析进行评估，同时通过额外的SM消融实验量化了土壤水分的增量预测贡献。在情景I中，ET0是主导预测因子，RF和BT取得了最高的独立测试性能（R2 = 0.875；RMSE = 0.398–0.399 mm day−1）。排除ET0降低了预测性能，在情景II中RF、LSBoost和BT的R2值约为0.80，同时预测因子重要性结构主要转向Rs，其次是SM、LST和NDVI。去除SM进一步降低了两种情景下的性能，且在ET0不可用时性能下降幅度更大：情景I中R2下降了0.025–0.044，情景II中下降了0.051–0.099。在排除SM的情景I配置下（该配置使预测因子信息更接近此前的ANN和SAFER研究），R2值范围为0.820至0.832，表明ETa预测性能的差异不能仅归因于模型结构，还强烈依赖于预测因子信息含量。总体而言，ET0为ETa预测提供了大量大气需求信息，而Rs承担了主导预测作用，SM在ET0不可用时提供了特别重要的补充信息。综合性能、消融和可解释性分析表明，在开发和比较ETa估算方法时，需将预测因子组成与模型结构一并考虑。","Atmosphere","2026-09-27T00:00:00Z",73,{"impact":135,"substance":136,"depth":137,"authority":20,"freshness":138,"relevant":21,"comment":139},12,22,18,8,"方法对比与可解释性分析扎实，对农业水资源管理和灌溉决策有参考价值，但属专业论文，公共影响有限。",[141],{"name":131,"url":128},[26,143,27,29,144],"农业遥感","蒸散发估算",{"code":83,"slug":88,"name":146,"path":147,"parent_code":43},"水资源与灌溉","\u002Ftopic\u002Fenvironment\u002Firrigation",[149,150],{"code":39,"slug":40,"name":41,"path":42,"parent_code":43},{"code":45,"slug":46,"name":47,"path":48,"parent_code":49},{"status":51,"locked":14,"version":52,"method":53,"evidence":152},{"candidates":153,"agricultural_evidence":58},[154,161,165,170,177,181],{"code":83,"score":19,"supported":58,"title":155,"abstract":157,"tags":160},[156],"evapotranspiration",[158,159,156],"水资源","蒸散发",[],{"code":39,"score":64,"supported":58,"title":162,"abstract":163,"tags":164},[40],[67,40],[29],{"code":45,"score":166,"supported":58,"title":167,"abstract":168,"tags":169},9,[71],[],[27],{"code":171,"score":172,"supported":58,"title":173,"abstract":174,"tags":176},"J1",3,[],[175],"农业规划",[],{"code":91,"score":21,"supported":14,"title":178,"abstract":179,"tags":180},[],[],[26],{"code":75,"score":21,"supported":14,"title":182,"abstract":183,"tags":184},[],[],[143],[186,187],"参考蒸散发 土壤水分 机器学习","ET0 实际蒸散发 随机森林","参考蒸散发土壤水分机器学习-3804",2,"10.3390\u002Fatmos17100941",{"doi":190,"openalex_id":192,"authors":193,"venue":131,"cited_by_count":100,"oa_url":128,"card":199,"direction":203,"ingested_from":121},"W7214588993",[194,197],{"name":195,"orcid":196},"Halil Karahan","https:\u002F\u002Forcid.org\u002F0000-0001-5346-5686",{"name":198,"orcid":9},"Devrim Alkaya",{"tldr":200,"method":201,"finding":202,"direction":203,"opportunity":204},"评估ET0与土壤湿度对机器学习估算实际蒸散的影响及可解释性。","用RF、BT、LSBoost、GAM、MLR结合SHAP、PFI、PDP及SM消","ET0主导预测，缺失时Rs主导且SM补充价值显著增大。","农业人工智能与决策模型","可探索缺ET0条件下融合遥感与土壤湿度的可解释ETa估算框架。","2026-09-29T23:30:57.729880Z",{"id":207,"title":208,"url":209,"summary":210,"summary_zh":211,"content":9,"source_name":212,"source_url":209,"published_at":213,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":214,"score_detail":215,"sources":217,"tags":219,"primary_topic":222,"secondary_topics":223,"classification":226,"search_phrases":246,"slug":249,"view_count":21,"doi":250,"paper":251,"created_at":269},2543,"Effects of Noise on the Accuracy of Estimating Rootzone Total Soil Moisture Using a Non‐Linear Autoregressive Exogenous Machine Learning Model for Precision Agriculture","https:\u002F\u002Fdoi.org\u002F10.1002\u002Fird.70226","ABSTRACT Widely used evapotranspiration (ET)–based irrigation methods contain many uncertainties that can affect irrigation quantity. To reduce these uncertainties, an alternative approach is the use of soil moisture (SM) data to estimate plant water uptake (PWU). If rootzone SM dynamics are understood, a surrogate machine learning (ML) model to predict the behaviour of SM dynamics can be developed to estimate PWU. Given the time and external variable dependency of SM, the non‐linear autoregressive exogenous (NARX) algorithm can be a better ML model for this purpose. However, the effect of measurement errors can affect prediction quality and hence the full deployment of data collection technology and computational algorithms. This paper presents a methodology consisting of analysing real‐world data, developing a generalized hypothetical SM curve, simulating measurement errors to develop noisy datasets and developing an ML model. The results show that the prediction accuracy is inversely proportional to the noise level. Up to a 10% measurement error, the NARX model can capture the SM dynamics with relatively high accuracy. However, the prediction quality decreases significantly as the noise level increases to 20%. For noise levels within 5%, the model prediction accuracy is significantly high, with correlation coefficients higher than 0.90 for all sets.","摘要 广泛使用的基于蒸散发（ET）的灌溉方法存在许多不确定性，可能影响灌溉量。为减少这些不确定性，一种替代方法是利用土壤水分（SM）数据估算植物吸水量（PWU）。如果理解了根区土壤水分动态，就可以开发一个替代性机器学习（ML）模型来预测土壤水分动态行为，从而估算植物吸水量。鉴于土壤水分对时间和外部变量的依赖性，非线性自回归外生（NARX）算法可能是更适合此目的的机器学习模型。然而，测量误差的影响可能影响预测质量，进而影响数据采集技术和计算算法的全面部署。本文提出了一套方法，包括分析真实世界数据、建立广义假设土壤水分曲线、模拟测量误差以生成含噪数据集，以及开发机器学习模型。结果表明，预测精度与噪声水平成反比。在测量误差不超过10%时，NARX模型能够以较高精度捕捉土壤水分动态。然而，当噪声水平增至20%时，预测质量显著下降。对于5%以内的噪声水平，模型预测精度显著较高，所有数据集的决定系数均高于0.90。","Irrigation and Drainage","2026-09-14T00:00:00Z",72,{"impact":20,"substance":18,"depth":19,"authority":20,"freshness":138,"relevant":21,"comment":216},"核心期刊论文，量化了测量噪声对NARX土壤水分预测精度的影响，为精准灌溉传感器布设与数据质量控制提供参考，但属细分方法研究，公共影响有限。",[218],{"name":212,"url":209},[26,220,27,221,29],"农业人工智能","精准灌溉",{"code":45,"slug":46,"name":47,"path":48,"parent_code":49},[224,225],{"code":39,"slug":40,"name":41,"path":42,"parent_code":43},{"code":83,"slug":88,"name":146,"path":147,"parent_code":43},{"status":51,"locked":14,"version":52,"method":53,"evidence":227},{"candidates":228,"agricultural_evidence":58},[229,233,237,241],{"code":45,"score":17,"supported":58,"title":230,"abstract":231,"tags":232},[71],[27,71],[220,27],{"code":39,"score":64,"supported":58,"title":234,"abstract":235,"tags":236},[40],[67,40],[29],{"code":83,"score":13,"supported":58,"title":238,"abstract":239,"tags":240},[],[87,159,88,156],[221],{"code":91,"score":166,"supported":58,"title":242,"abstract":244,"tags":245},[243],"precision agriculture",[],[26],[247,248],"农业人工智能 土壤墒情 智慧农业 机器学习","农业人工智能 土壤墒情","农业人工智能土壤墒情智慧农业机器学习-2543","10.1002\u002Fird.70226",{"doi":250,"openalex_id":252,"authors":253,"venue":212,"cited_by_count":100,"oa_url":9,"card":264,"direction":203,"ingested_from":121},"W7213181547",[254,257,260,262],{"name":255,"orcid":256},"Fayzul Pasha","https:\u002F\u002Forcid.org\u002F0000-0002-8295-0602",{"name":258,"orcid":259},"Ashok Inturi","https:\u002F\u002Forcid.org\u002F0009-0004-1934-4310",{"name":261,"orcid":9},"Kinnoree R. Pasha",{"name":263,"orcid":9},"Dilruba Yeasmin",{"tldr":265,"method":266,"finding":267,"direction":203,"opportunity":268},"该论文评估了测量噪声对NARX机器学习模型估算根区土壤水分动态精度的影响。","使用NARX模型，基于真实数据、广义假设曲线和模拟噪声数据集。","预测精度与噪声水平成反比；噪声≤10%时精度较高，≤5%时相关系数>0.90，20%时显著下降。","可研究自适应去噪或鲁棒NARX模型，以在20%以上噪声下维持土壤水分预测精度。","2026-09-15T23:30:38.023164Z",{"id":271,"title":272,"url":273,"summary":274,"summary_zh":275,"content":9,"source_name":276,"source_url":273,"published_at":277,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":278,"score_detail":279,"sources":282,"tags":284,"primary_topic":287,"secondary_topics":288,"classification":290,"search_phrases":307,"slug":310,"view_count":100,"doi":311,"paper":312,"created_at":324},4498,"SMART CROP SELECTION MODEL USING SOIL PH VALUE, TEMPERATURE, HUMIDITY, AND RAINFALL PREDICTION","https:\u002F\u002Fdoi.org\u002F10.5281\u002Fzenodo.23211381","Crop suitability with soil and climatic conditions plays a vital role in increasing productivity of agriculture. This study proposes a smart crop selection model which recommends the suitable crop based on the machine learning techniques with the input parameters of soil pH, soil temperature, soil humidity and rainfall. The data set consisted of 1000 observations and 8 different crop classes and missing values were handled, crop labels were encoded and numerical features were selected. Five-fold cross validation was used to compare the performance of the three classifiers: Random Forest, XGBoost and Support Vector Machine (SVM). Finally, a Gradient-Boosted classification model was trained and evaluated with test-set accuracy, classification metrics, Multiclass ROC analysis and feature importance. The results of the application were 93.60% of the final test accuracy and the model comparison was used to compare the three classification methods. The contribution of four environmental parameters to the crop classification was analyzed by feature-importance analysis. To facilitate practical accessibility, the entire process was carried out using an interactive dashboard developed with Streamlit, allowing for the upload of datasets, evaluation of models, selection of parameters and the generation of predictions in real time for the crops of interest. The developed system is an example of how machine learning can be integrated with an interactive decision support interface to develop a data-driven, crop selection system, relying on simple soil and weather parameters.","作物适宜性与土壤及气候条件对提高农业生产率至关重要。本研究提出了一种智能作物选择模型，该模型基于机器学习技术，以土壤pH值、土壤温度、土壤湿度和降雨量为输入参数，推荐适宜的作物。数据集包含1000个观测值和8个不同作物类别，对缺失值进行了处理，对作物标签进行了编码，并筛选了数值特征。采用五折交叉验证比较了三种分类器——随机森林（Random Forest）、XGBoost和支持向量机（Support Vector Machine, SVM）的性能。最终，训练了一个梯度提升分类模型，并通过测试集准确率、分类指标、多类ROC分析和特征重要性进行了评估。应用结果为最终测试准确率93.60%，并通过模型比较对三种分类方法进行了对比。通过特征重要性分析，探讨了四个环境参数对作物分类的贡献。为便于实际应用，整个流程通过使用Streamlit开发的交互式仪表板实现，支持数据集上传、模型评估、参数选择以及针对感兴趣作物的实时预测生成。所开发的系统展示了如何将机器学习与交互式决策支持界面相结合，构建一个数据驱动的作物选择系统，且仅依赖于简单的土壤和气象参数。","Zenodo (CERN European Organization for Nuclear Research)","2026-10-07T00:00:00Z",69,{"impact":135,"substance":280,"depth":17,"authority":20,"freshness":138,"relevant":21,"comment":281},20,"基于土壤pH、温湿度与降雨的机器学习选种模型，含93.6%测试精度与Streamlit交互看板，方法完整但属常规应用型研究，产业影响有限。",[283],{"name":276,"url":273},[26,220,285,29,286],"决策支持系统","作物选择",{"code":39,"slug":40,"name":41,"path":42,"parent_code":43},[289],{"code":45,"slug":46,"name":47,"path":48,"parent_code":49},{"status":51,"locked":14,"version":52,"method":53,"evidence":291},{"candidates":292,"agricultural_evidence":58},[293,297,303],{"code":39,"score":64,"supported":58,"title":294,"abstract":295,"tags":296},[40],[67,40],[29],{"code":45,"score":298,"supported":58,"title":299,"abstract":300,"tags":302},11,[],[27,71,301],"decision support",[220,285],{"code":91,"score":21,"supported":14,"title":304,"abstract":305,"tags":306},[],[],[26],[308,309],"土壤pH 降雨量 作物选择模型","机器学习 作物推荐 智慧农业","土壤pH降雨量作物选择模型-4498","10.5281\u002Fzenodo.23211381",{"doi":311,"openalex_id":313,"authors":314,"venue":276,"cited_by_count":100,"oa_url":273,"card":319,"direction":120,"ingested_from":121},"W7220991946",[315,317],{"name":316,"orcid":9},"C. SAI KUMAR",{"name":318,"orcid":9},"Prof. G SREENIVASULU",{"tldr":320,"method":321,"finding":322,"direction":203,"opportunity":323},"基于土壤pH、温湿度和降雨，用机器学习构建智能选种模型并开发交互式仪表板。","1000条8类作物数据，五折交叉验证比较RF、XGBoost、SVM，最终用梯度","梯度提升模型测试准确率达93.60%，特征重要性揭示四参数对选种的贡献。","可引入时序气象与多源土壤数据，提升模型泛化性并适配不同区域作物推荐。","2026-10-09T23:30:12.285129Z",{"id":326,"title":327,"url":328,"summary":329,"summary_zh":330,"content":9,"source_name":10,"source_url":328,"published_at":277,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":331,"score_detail":332,"sources":335,"tags":337,"primary_topic":340,"secondary_topics":341,"classification":347,"search_phrases":371,"slug":374,"view_count":375,"doi":376,"paper":377,"created_at":394},4433,"PROSAIL-DNN-TL: A Hyperspectral Inversion Framework for Winter Wheat Leaf Chlorophyll Content Based on Fine-Tuned Transfer Learning","https:\u002F\u002Fdoi.org\u002F10.3390\u002Fagronomy16191971","Accurate and efficient estimation of leaf chlorophyll content (LCC) is essential for crop physiological monitoring and precision agriculture. Current remote sensing approaches for LCC estimation mainly rely on machine learning and deep learning methods, which exhibit strong nonlinear modelling capabilities but require sufficient and representative samples. Radiative transfer models (RTMs) can simulate canopy spectral response based on physical mechanisms and generate large-scale datasets, thereby overcoming the limitations of purely data-driven approaches. However, their applications under diverse environmental conditions remain challenging due to uncertainties caused by model parameterization and discrepancies between simulated and observed data. This study proposed a hybrid framework, PROSAIL-DNN-TL, integrating the PROSAIL RTM with full-network fine-tuned transfer learning for winter wheat LCC estimation. First, 85,680 simulated samples generated by PROSAIL were used to pre-train a deep neural network (DNN), and field observations (n = 270) were subsequently used for full-network fine-tuning. A conventional DNN model trained only with field data and machine learning models based on different feature selection strategies were established for comparison. The results showed that: (1) UVE-selected bands were distributed across visible, red-edge, and near-infrared regions; CARS retained more bands in the visible region, whereas SPA mainly selected near-infrared bands; (2) RR and SVR achieved higher accuracy than KNN, with the full-spectrum and CARS feature sets showing superior performance and SPA showing the lowest accuracy; (3) PROSAIL-DNN-TL achieved reliable LCC estimation performance, with an average test R2 of 0.79 across 500 repeated experiments, and showed improved performance compared with conventional machine learning models; (4) PROSAIL-DNN-TL achieved reliable predictive performance when evaluated using datasets collected during the 2024 and 2025 growing seasons, with R2 values of 0.75 and 0.92, respectively. Overall, the proposed hybrid framework integrates physical knowledge from RTMs with data-driven modelling through transfer learning, providing a potential approach for crop chlorophyll monitoring. However, the SPAD-to-LCC conversion used in this study was not specifically calibrated for the instrument and crop, and therefore, the absolute LCC values should be interpreted with caution.","准确、高效地估算叶片叶绿素含量（leaf chlorophyll content，LCC）对作物生理监测和精准农业至关重要。当前遥感估算LCC的方法主要依赖机器学习和深度学习方法，这类方法具有较强的非线性建模能力，但需要充足且具有代表性的样本。辐射传输模型（radiative transfer models，RTMs）能够基于物理机制模拟冠层光谱响应并生成大规模数据集，从而克服纯数据驱动方法的局限性。然而，由于模型参数化带来的不确定性以及模拟数据与观测数据之间的差异，其在不同环境条件下的应用仍面临挑战。本研究提出了一种混合框架PROSAIL-DNN-TL，将PROSAIL辐射传输模型与全网络微调迁移学习相结合，用于冬小麦LCC估算。首先，利用PROSAIL生成的85 680个模拟样本预训练深度神经网络（deep neural network，DNN），随后利用田间观测数据（n = 270）进行全网络微调。同时建立了仅使用田间数据训练的传统DNN模型以及基于不同特征选择策略的机器学习模型作为对比。结果表明：（1）UVE选择的波段分布在可见光、红边和近红外区域；CARS在可见光区域保留了更多波段，而SPA主要选择近红外波段；（2）RR和SVR的精度高于KNN，全光谱和CARS特征集表现较优，SPA精度最低；（3）PROSAIL-DNN-TL取得了可靠的LCC估算性能，在500次重复实验中平均测试R²为0.79，且相比传统机器学习模型表现出更优的性能；（4）利用2024年和2025年生长季采集的数据集进行评估时，PROSAIL-DNN-TL均表现出可靠的预测性能，R²分别为0.75和0.92。总体而言，所提出的混合框架通过迁移学习将辐射传输模型的物理知识与数据驱动建模相结合，为作物叶绿素监测提供了一种潜在途径。然而，本研究中使用的SPAD到LCC的转换并未针对仪器和作物进行专门校准，因此绝对LCC值应谨慎解读。",81,{"impact":137,"substance":136,"depth":137,"authority":333,"freshness":166,"relevant":21,"comment":334},14,"将PROSAIL辐射传输模型与深度神经网络迁移学习结合，用8万余模拟样本预训练、270个田间样本微调，冬小麦叶绿素反演精度可靠，方法新颖且数据规模扎实，对作物生理监测与精准农业有参考价值。",[336],{"name":10,"url":328},[26,143,30,338,339],"迁移学习","叶绿素监测",{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},[342,346],{"code":75,"slug":343,"name":344,"path":345,"parent_code":49},"remote-sensing","遥感与农情监测","\u002Ftopic\u002Fdigital-agriculture\u002Fremote-sensing",{"code":45,"slug":46,"name":47,"path":48,"parent_code":49},{"status":51,"locked":14,"version":52,"method":53,"evidence":348},{"candidates":349,"agricultural_evidence":58},[350,354,359,365],{"code":32,"score":57,"supported":58,"title":351,"abstract":352,"tags":353},[60,33],[34,60,33],[30],{"code":75,"score":64,"supported":58,"title":355,"abstract":357,"tags":358},[356],"hyperspectral",[78,79],[143],{"code":45,"score":13,"supported":58,"title":360,"abstract":361,"tags":364},[],[362,27,71,363],"深度学习","deep learning",[338],{"code":91,"score":366,"supported":58,"title":367,"abstract":368,"tags":370},7,[],[369,243],"精准农业",[26],[372,373],"PROSAIL 冬小麦 叶绿素","高光谱 叶绿素 反演","PROSAIL冬小麦叶绿素-4433",5,"10.3390\u002Fagronomy16191971",{"doi":376,"openalex_id":378,"authors":379,"venue":10,"cited_by_count":100,"oa_url":328,"card":389,"direction":118,"ingested_from":121},"W7220844556",[380,382,384,387],{"name":381,"orcid":9},"Xueqing Zhu",{"name":383,"orcid":9},"Ying Nian",{"name":385,"orcid":386},"Jikai Liu","https:\u002F\u002Forcid.org\u002F0000-0003-2365-5221",{"name":388,"orcid":9},"Xinwei Li",{"tldr":390,"method":391,"finding":392,"direction":118,"opportunity":393},"提出PROSAIL-DNN-TL框架，用PROSAIL模拟数据预训练DNN并全网络微调，反演冬小麦叶","PROSAIL辐射传输模型生成85680样本预训练DNN，270个田间样本全网络","500次重复实验平均测试R²为0.79，2024和2025年生长季验证R²分别为0.75和0.92，","SPAD到LCC转换未针对仪器和作物标定，可研究跨传感器、跨作物、跨生长季的迁移学习泛化与绝对叶绿素标定。","2026-10-08T23:30:29.383957Z",{"id":396,"title":397,"url":398,"summary":399,"summary_zh":9,"content":9,"source_name":400,"source_url":9,"published_at":401,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":402,"score_detail":403,"sources":405,"tags":407,"primary_topic":409,"secondary_topics":410,"classification":412,"search_phrases":432,"slug":435,"view_count":92,"doi":9,"paper":436,"created_at":444},4286,"Applications of Artificial Intelligence in Soil Characterization and Agriculture: A Systematic Review","https:\u002F\u002Fm2.mtmt.hu\u002Fapi\u002Fpublication\u002F37386408","Navarro Rubio等系统综述2018-2026年196项AI土壤特性与农业应用研究，遵循PRISMA 2020框架。综述发现研究明显向数据驱动方法转变，机器学习和深度学习模型主导近期研究；随机森林、支持向量机、梯度提升方法、人工神经网络、卷积神经网络、长短期记忆架构最常被报告；主要数据源包括原位传感、实验室测量、遥感影像和环境协变量，常通过多源数据融合框架集成；树集成模型在不同土壤特性上提供稳健性能。","Agronomy 2026 16(13):1241 DOI:10.3390\u002Fagronomy16131241","2026-10-02T00:00:00Z",78,{"impact":137,"substance":136,"depth":137,"authority":333,"freshness":84,"relevant":21,"comment":404},"PRISMA框架下196项研究的系统综述，方法学扎实、结论可靠，对农业AI土壤应用具较高参考价值。",[406],{"name":400,"url":398},[26,220,27,78,408],"土壤监测",{"code":45,"slug":46,"name":47,"path":48,"parent_code":49},[411],{"code":39,"slug":40,"name":41,"path":42,"parent_code":43},{"status":51,"locked":14,"version":52,"method":53,"evidence":413},{"candidates":414,"agricultural_evidence":58},[415,420,424,428],{"code":45,"score":17,"supported":58,"title":416,"abstract":418,"tags":419},[417],"artificial intelligence",[362,27],[220,27],{"code":39,"score":298,"supported":58,"title":421,"abstract":422,"tags":423},[40],[67],[],{"code":75,"score":92,"supported":58,"title":425,"abstract":426,"tags":427},[],[78],[78],{"code":91,"score":21,"supported":14,"title":429,"abstract":430,"tags":431},[],[],[26],[433,434],"AI 土壤特性 系统综述","机器学习 土壤 遥感","AI土壤特性系统综述-4286",{"doi":9,"openalex_id":9,"authors":437,"venue":9,"cited_by_count":100,"oa_url":9,"card":438,"direction":203,"ingested_from":443},[],{"tldr":439,"method":440,"finding":441,"direction":203,"opportunity":442},"系统综述2018-2026年196项AI用于土壤特性与农业的研究。","PRISMA 2020框架，分析机器学习与深度学习模型及多源数据融合。","研究转向数据驱动，树集成模型在不同土壤特性上表现稳健。","可探索可解释AI与多源数据融合在土壤特性预测中的泛化能力。","agent","2026-10-06T00:05:04.976398Z",{"id":446,"title":447,"url":448,"summary":449,"summary_zh":450,"content":9,"source_name":451,"source_url":448,"published_at":452,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":453,"score_detail":454,"sources":456,"tags":458,"primary_topic":462,"secondary_topics":463,"classification":470,"search_phrases":497,"slug":500,"view_count":189,"doi":501,"paper":502,"created_at":529},4172,"Machine learning reveals distinct drivers of soil organic carbon across the mountain-oasis-desert system of northwest China","https:\u002F\u002Fdoi.org\u002F10.1186\u002Fs13021-026-00522-5","The mountain-oasis-desert system (MODS) is a distinctive landscape pattern shaped by interactions between geomorphological configurations and hydrological processes in arid regions. However, most soil organic carbon (SOC) mapping studies treat arid landscapes as homogeneous units, leaving limited quantitative evidence on how SOC patterns and environmental associations differ among MODS subsystems. As a critical ecosystem highly sensitive to climate change and human activities, understanding these responses is essential for regional soil carbon management and ecological restoration. In this study, we developed an ecosystem-based modeling framework in the Hexi Region (HXR) to construct separate SOC prediction models for different subsystems, aiming to investigate the spatiotemporal dynamics of SOC stocks and identify their key driving factors across MODS. The results show that incorporating remote sensing (RS) variables significantly improved model performance, increasing the mean R 2 from 0.51 to 0.58. Among the tested models, XGBoost achieved the best performance, providing an effective tool for SOC mapping in arid regions. Shapley additive explanations (SHAP) analysis of the best-performing XGBoost models identified distinct leading predictors across the MODS, including surface shortwave radiation (SRAD) in mountain and desert ecosystems and the difference vegetation index (DVI) in oasis ecosystems. Furthermore, the partial least squares path model (PLS-PM) revealed contrasting influences of climate change and human activities across ecological subsystems. Climatic factors exerted the strongest influence in the mountainous subsystem, whereas in the oasis subsystem intensive human management became the dominant driver of soil carbon cycling. From 2001 to 2023, SOC stocks in oasis regions increased significantly at a rate of 0.17 t C ha⁻ 1 year⁻ 1 , potentially associated with land-use change and agricultural management. Approximately 19.3% of the study area experienced significant SOC stocks declines. By resolving subsystem-specific SOC patterns and environmental associations, this study advances ecosystem-aware SOC assessment in arid regions. The resulting framework also provides a basis for differentiated carbon management across MODS.","山地-绿洲-荒漠系统（MODS）是干旱区地貌格局与水文过程相互作用形成的独特景观格局。然而，大多数土壤有机碳（SOC）制图研究将干旱区景观视为均质单元，关于MODS各子系统间SOC格局及其环境关联差异的定量证据仍然有限。作为对气候变化和人类活动高度敏感的关键生态系统，理解这些响应对于区域土壤碳管理和生态恢复至关重要。本研究在河西地区（HXR）构建了基于生态系统的建模框架，针对不同子系统分别建立SOC预测模型，旨在探究MODS中SOC储量的时空动态并识别其关键驱动因子。结果表明，引入遥感（RS）变量显著提升了模型性能，平均R²从0.51提高至0.58。在测试的模型中，XGBoost表现最佳，为干旱区SOC制图提供了有效工具。对最优XGBoost模型的SHAP（Shapley加法解释）分析识别出MODS各子系统的不同主导预测因子，其中山地和荒漠生态系统为地表短波辐射（SRAD），绿洲生态系统为差值植被指数（DVI）。此外，偏最小二乘路径模型（PLS-PM）揭示了气候变化和人类活动在各生态子系统中的差异化影响。气候因子在山地子系统中影响最强，而在绿洲子系统中，集约化人类管理成为土壤碳循环的主导驱动因素。2001年至2023年，绿洲区域SOC储量以0.17 t C ha⁻¹ year⁻¹的速率显著增加，这可能与土地利用变化和农业管理有关。约19.3%的研究区域经历了SOC储量的显著下降。通过解析子系统特定的SOC格局及其环境关联，本研究推进了干旱区生态系统感知型SOC评估，所构建的框架也为MODS差异化碳管理提供了依据。","Carbon Balance and Management","2026-10-03T00:00:00Z",79,{"impact":17,"substance":136,"depth":137,"authority":333,"freshness":166,"relevant":21,"comment":455},"基于XGBoost与SHAP的干旱区SOC分异驱动研究，方法新颖、数据翔实，对区域碳管理与生态修复有参考价值。",[457],{"name":451,"url":448},[27,459,460,28,461],"土壤有机碳","生态修复","干旱区农业",{"code":39,"slug":40,"name":41,"path":42,"parent_code":43},[464,469],{"code":465,"slug":466,"name":467,"path":468,"parent_code":43},"H7","climate","气候适应与防灾减灾","\u002Ftopic\u002Fenvironment\u002Fclimate",{"code":45,"slug":46,"name":47,"path":48,"parent_code":49},{"status":51,"locked":14,"version":52,"method":53,"evidence":471},{"candidates":472,"agricultural_evidence":58},[473,477,484,488,492],{"code":39,"score":64,"supported":58,"title":474,"abstract":475,"tags":476},[40],[67,40],[459],{"code":465,"score":13,"supported":58,"title":478,"abstract":479,"tags":483},[],[480,481,482],"气候变化","干旱","climate change",[461],{"code":45,"score":166,"supported":58,"title":485,"abstract":486,"tags":487},[71],[],[27],{"code":75,"score":366,"supported":58,"title":489,"abstract":490,"tags":491},[],[78,79],[28],{"code":493,"score":21,"supported":14,"title":494,"abstract":495,"tags":496},"H6",[],[],[460],[498,499],"河西地区 土壤有机碳 机器学习","山地-绿洲-荒漠 遥感 碳储量","河西地区土壤有机碳机器学习-4172","10.1186\u002Fs13021-026-00522-5",{"doi":501,"openalex_id":503,"authors":504,"venue":451,"cited_by_count":100,"oa_url":448,"card":524,"direction":118,"ingested_from":121},"W7219075005",[505,507,510,512,515,517,520,522],{"name":506,"orcid":9},"Y. Chen",{"name":508,"orcid":509},"Haiyang Xi","https:\u002F\u002Forcid.org\u002F0000-0003-3158-8161",{"name":511,"orcid":9},"Meng Zhu",{"name":513,"orcid":514},"Wenju Cheng","https:\u002F\u002Forcid.org\u002F0000-0002-1822-5668",{"name":516,"orcid":9},"Yulian Hao",{"name":518,"orcid":519},"Linbo Qu","https:\u002F\u002Forcid.org\u002F0009-0007-4669-8407",{"name":521,"orcid":9},"Zhehui Wang",{"name":523,"orcid":9},"Bin Wang",{"tldr":525,"method":526,"finding":527,"direction":118,"opportunity":528},"构建基于生态系统的机器学习框架，预测西北荒漠-绿洲-山地系统土壤有机碳并识别各子系统驱动因子。","XGBoost、SHAP、PLS-PM，结合遥感变量与2001-2023年多源数","遥感变量提升模型R²至0.58；山地和沙漠由辐射主导，绿洲由人类管理主导，绿洲碳储量年增0.17 t","可延伸研究绿洲农业管理措施对土壤碳的长期影响，并开发子系统自适应的碳监测与决策模型。","2026-10-04T23:30:24.631460Z"]