Above-ground biomass and plant nitrogen content play a crucial role in crop growth, development, and yield formation. Therefore, dynamic monitoring of crop growth and nutritional status is of considerable importance. The study used unmanned aerial vehicles to obtain hyperspectral data and above-ground biomass during the budding stage, tuber formation stage, tuber growth and starch accumulation stage, to analyze the correlation and the importance of variable projection between vegetation indices and biomass and plant nitrogen content, and to screen out vegetation indices that are sensitive to biomass and plant nitrogen content combining deep neural network (DNN), partial least squares (PLSR), clastic network regression (ENR), ridge regression (RR) and support vector machine (SVR) to estimate biomass and plant nitrogen content and comparing the effectiveness of different models in estimating biomass and plant nitrogen content, The results showed that (1) the correlation between vegetation indices and both biomass and plant nitrogen content reached 0, 01 significant level, and the importance of the variable projection was used to screen out the vegetation indices that were sensitive to biomass and plant nitrogen content; (2) Comparing the remote sensing estimation models for the five growth stages, the best model for biomass and plant nitrogen content was constructed at the tuber formation stage, the worst model for biomass was estimated at the present bud stage, and the worst model for plant nitrogen content was estimated at the tuber growth stage. (3) The optimum biomass model constructed in the tuber formation stage using the PLSR method was modelled with R, RMSE and NRMSE was 0.60, 235.65 kg center dot hm(-2 )and 0. 15 kg hm respectively, and validated with R-2, RMSE and NRMSE was 0.58, 344.72 kg center dot hm(-2) and 0.26 kg center dot hm(-2), The optimum plant nitrogen content model constructed during tuber formation stage using RR method was modelled with R-2, RMSE and NRMSE was 0.74, 0.31% and 0.15%, validated R-2, RMSE and NRMSE was 0.77, 0.58% and 0.28%. Comprehensively comparing the DNN, PLSR, ENR, RR, and SVR algorithms for estimating biomass and plant nitrogen content models, the accuracy of the estimated plant nitrogen content model is found to be better than that of the estimated biomass model. The plant's nitrogen content can be used to more effectively monitor crop growth and nutritional characteristics, providing a reference for informed agricultural management.

Above-ground biomass (AGB) is a measure used to assess crop growth and estimate yield. Accurate, non-destructive and rapid AGB estimation is crucial for advancing modern agriculture. Unlike traditional destructive survey methods, the unmanned aerial vehicle (UAV) remote sensing provides a highly flexible and low-cost way. This study provides a systematic review in terms of UAV-based AGB estimation, including the UAV monitoring mechanisms, UAV platforms, sensors, features, methods, and difficulties and future development directions in AGB estimation. Also, this study subjectively evaluates functions and limitations of different sensors, features, estimation methods, and explores the application and prospects for the development of crop AGB estimation models. Given the current challenges of UAV-based AGB estimation, some important directions for future research are also provided. This review aims to provide practical strategies for future research on estimating AGB and enhancing the efficiency of this technique in assessing crop growth and yield prediction, thereby promoting the development of precision agriculture.

Timely and accurate potato yield estimation is of great significance for large-scale agricultural production and early government decision-making. This study employed unmanned aerial vehicles to acquire hyperspectral imagery during critical growth stages of potatoes, concurrently collecting ground-based yield and morphological parameters. Extreme Learning Machine (ELLM), Partial Least Squares (PLS), Lasso regression (Least Absolute Shrinkage and Selection Operator), and ridge regression (RR) were utilized to construct yield estimation models for different growth stages. Theresults indicated that when estimating potato yields using univariate morphological parameters, vegetation indices, or a combination of vegetation indices and morphological parameters at different growth stages, the highest predictive accuracy is achieved during the tuber formation stage. The yield estimation model constructed by combining vegetation index and morphological parameters is the best (with determination coefficients R & sup2; of 0.87, 0.79, and 0.86 for the three growth stages), followed by the vegetation index yield estimation model (with determination coefficients R & sup2; of 0.86, 0.75, and 0.78 for the three growth stages), and the univariate morphological parameter yield estimation model is the worst (with determination coefficients R degrees of 0.77, 0.60, and 0.81 for each growth stage). When estimating potato yields using vegetation indices as input variables, the PLS model demonstrated superior accuracy compared to the other three models. When estimating potato yield using vegetation indices combined with morphological parameters, the ELM method yielded the optimal model. The potato yield model integrating vegetation indices and plant height during tuber formation achieved the highest modelling R & sup2; of 0.87. The optimal potato yield model during the tuber growth stage was constructed by integrating vegetation indices and canopy cover, yielding an R of 0.79. During the starch accumulation stage, the optimal potato yield model was constructed by integrating vegetation indices with canopy cover and canopy volume, yielding an R of 0.86. The results indicate that the tuber formation period can effectively estimate potato yield and provide scientific basis for early monitoring of potato yield.

Timely and accurate acquisition of crop biomass and height information is of great significance for monitoring crop growth conditions and improving crop yields. To enhance the accuracy and stability of potato yield forecasting, this study constructed three-dimensional growth information by multiplying vegetation indices with plant height, and it incorporated biomass an agronomic parameter reflecting potato growth status. Ridge Regression (RR), Random Forest (RF), K-Nearest Neighbors (KNN), Partial Least Squares Regression (PLSR) and Support Vector Regression (SVR) were used to establish the potato yield estimation a potato yield estimation model using "remote sensing information (remote sensing information fused with plant height) biomass yield". The results indicate: The plant height extracted by UAVs exhibited and the measured plant height had high fitting precision (R & sup2; values of 0.88, 0.87, and 0.76 respectively); (2) The correlation between vegetation indices, vegetation index-fused plant height parameters, and biomass is highly significant (p<0.01). Compared to single vegetation indices, the correlation between vegetation index-fused plant height parameters and biomass improves significantly across different growth stages; (3) The biomass model constructed by integrating vegetation index with plant height parameters using RR, RF, KNN, PLSR, and SVR has better accuracy and stability than the biomass model constructed solely based on vegetation index; (4) The remote sensing estimation model for potato yield, constructed based on the "vegetation index fused with plant height parameters biomass yield" chain relationship, demonstrated superior accuracy and stability across different growth stages compared to the model built on the "vegetation index biomass yield" chain relationship. During the tuber formation stage, the yield model constructed using PLSR was optimal (R(2)0.86, RMSE-3988.08 kg hm(-2 )tuber growth stage, the yield model constructed using SVR was optimal (R-0.73, RMSE-5 601.89 kg the starch accumulation stage, the yield model constructed using SVR was optimal (R-0.70, RMSE-5 888.33 kg hm(-2)). This study provides a reference for efficiently and rapidly predicting potato yields using drone remote sensing. hm), during the hm), and during

Rapid, real-time and accurate acquisition of the nitrogen nutrition status of winter wheat is crucial for evaluating winter wheat growth, estimating yield, and guiding agricultural modernization and production. This study utilized hyperspectral cameras and digital cameras mounted on drones to collect canopy spectral data during the three critical growth periods, Simultaneous ground experiments were conducted to determine the physical and chemical properties of biomass and plant nitrogen content. Four characteristic parameters, including the vegetation index, the red edge index, the red edge parameter, and the three-band parameter of the hyperspectral image, as well as the color index of the digital camera and its fusion parameters, were selected. Partial Least Squares Regression (PLSR). Stepwise Regression (SWR). Random Forest (RF) and Back Propagation. (BP) algorithms were used to establish a winter wheat nitrogen nutrition index monitoring model. The accuracy of the model was evaluated, and the optimal estimation model was selected. The results showed that (1) In univariate modeling, the nitrogen nutrition index model constructed with the red-edge parameter DIDRInid was the best, achieving a modeling R-2 of 0. 66. RMSE of 0.11% and a validation R-2 of 0.55. RMSE of 0.13%. (2) In the multivariate modeling the nitrogen nutrient index model constructed with the red edge parameter as the independent variable was superior to the nitrogen nutrient index model constructed with the vegetation index, the red edge index, the three-band parameter and the color index as the independent variables. where the nitrogen nutrition index model constructed by the BP algorithm based on red edge parameters was optimal (modeling R-2 - 0.75, RMSE - 0.10%, validation R-2 - 0.60. RMSE - 0.12%). (3) In fusing hyperspectral parameters and digital index variable modelling, the nitrogen nutrient index model constructed with the multimodal variable red edge parameter + color index was superior to the nitrogen nutrient index models constructed with the red edge parameter + vegetation index. red edge parameter + red edge index and red edge parameter triple band parameter, where the nitrogen nutrient index model constructed using PLSR with the fusion of the red edge parameters and color index was optimal (modeling R-2-0.77. RMSE-0.09%, validation R-2-0.65, RMSE-0.11%). Its model accuracy was better than that of univariate modeling and multivariate modeling. This study can provide an important reference for estimating the nitrogen nutrition status of winter wheat, Keywords

Timely and accurate monitoring of potato crop growth and estimating yields are essential to improve agricultural production. Unmanned aerial vehicle (UAV)-based hyperspectral remote sensing is a non-destructive method for crop growth monitoring (CGM) and yield estimation, which plays a vital role in the agricultural application. However, CGM and yield estimation are typically achieved through quantitative inversion of specific crop traits, which lacks consideration for the interactive impacts among traits. Thus, this study aimed to integrate multiple agronomic traits using a fuzzy comprehensive evaluation (FCE) method to construct a new crop growth monitoring indicator (CGMI) for CGM and yield estimation. In 2018 and 2019, UAV hyperspectral images and ground parameters were acquired during three growth stages of potatoes. Compared to single agronomic traits, CGMI could be better described by vegetation indices (VIs). The accuracy and stability of the CGMI estimation model were effectively validated, while the single trait estimation model performed poorly on the validation set. The coefficient of determination (R2) values of CGMI estimation for three stages were in the range of 0.56-0.72 and 0.56-0.66 for calibration and validation sets. The CGMI at different stages was closely correlated with potato yield, reaching a highly significant level. The VIs selected based on CGMI and Akaike information criterion (AIC) were input into the PLSR model to estimate potato yields. The R2 values of yield estimation for three stages were in the range of 0.63-0.69 and 0.54-0.60 for calibration and validation sets. The study demonstrated that integrating multiple crop traits could enhance the relationship with yield and provided a comprehensive reflection of crop growth. The CGMI constructed in this study can provide decision-making services for crop production management in the field.

Aboveground biomass (AGB) reflects the accumulation of crop photosynthesis, and AGB data guide agricultural production and field management practices. AGB can be estimated using UAV hyperspectral data; however, external factors and high-dimensional data lead to uncertainties. To address these issues, a cascading spectral preprocessing and band-optimized AGB estimation framework are proposed. We collected canopy hyperspectral reflectance and potato AGB data across two varieties, three planting densities, four nitrogen levels, and two potassium treatments during three growth stages. Then, we systematically compared the performance of Savitzky-Golay (SG) smoothing, multiplicative scatter correction (MSC), first-order differentiation (FOD) and their cascaded combinations. We also rigorously evaluated the ability of competitive adaptive reweighted sampling (CARS), successive projection algorithm (SPA) and their cascaded combination (CARS-SPA) to identify sensitive bands. The results indicated that cascaded spectral preprocessing methods significantly enhance the accuracy of potato AGB estimation. Among these approaches, the SG-MSC-FOD cascade performed most effectively. The combination of CARS and SPA yielded the fewest model variables while achieving the highest estimation accuracy. Furthermore, the integration of SG-MSC-FOD and CARS-SPA with partial least squares regression achieved the highest accuracy in AGB estimation across multiple growth stages, with a coefficient of determination (R2) of 0.73, root mean square error (RMSE) of 256.09 kg/hm2, and normalized root mean square error (NRMSE) of 21.51 %. We validated the proposed method under different varieties, planting densities, and nitrogen and potassium treatments. This approach effectively reduces noise, lowers dimensionality, and enhances AGB estimation accuracy, providing a reliable solution for monitoring potato crop growth using hyperspectral remote sensing.

Timely and accurate crop yield estimation is crucial for making informed decisions regarding crop management and assessing food security. This study aims to develop a method that combines continuous wavelet transform (CWT) with machine learning to predict wheat yield accurately. This research is based on the spectral data of canopy height and yield data obtained from two-year field trials conducted during wheat growth's flowering and filling stages in 2020 2021, Initially. CWT is employed to extract three wavelet features (WFs), namely Hortus- WFs based on the Bortua method, 1% R-2-WFs representing WFs along with the top 1% determination coefficient for wheat yield, and SS-WFs encompassing all WFs under a single decomposition scale, Subsequently, three machine learning algorithms Random Forest (RF). K-nearest neighbor (KNN). and extreme gradient Lift (XGPoost) are utilized to construct the yield prediction model, Finally, optimal spectral features are selected using the same methodology for modeling and comparison purposes. The results demonstrate that: (1) all three WFs models combined with machine learning methods perform well, with higher accuracy and stability observed in the model built based on Boruta-WFs (2) Compared to the spectral characteristic model, improved accuracy was achieved by utilizing Bortua WFs at cach growth stager specifically, an increase in R' accuracy by 17.5%, 4% and 39.6% during flowering stage. well as an increase by 8.4%. 5.6%, and 16.9% during filling stage respectively were observed across different models, (3) The estimation model at the grouting stage outperformed that at the flowering stages particularly noteworthy was the performance of XGBoost when combined with Bortua-WFs, which yielded an R-2 value of 0. 83 accompanied by an RMSE value of 0.78 t. ha(-1). This study compared the performance of different characteristics and methods. It determined the best model accuracy under different schemes, which can provide technical references for the accurate wheat yield prediction by spectral technology.

The rapid and accurate estimation of crop yield is of great importance for large-scale agricultural production and national food security. Using winter wheat as the research object, the effects of color indexes, texture feature and fusion index on yield estimation were investigated based on unmanned aerial vehicle (UAV) high-definition digital images, which can provide a reliable technical means for the high-precision yield estimation of winter wheat. In total, 22 visible color indexes were extracted using UAV high-resolution digital images, and a total of 24 texture features in red, green, and blue bands extracted by ENVI 5.3 were correlated with yield, while color indexes and texture features with high correlation and fusion indexes were selected to establish yield estimation models for flagging, flowering and filling stages using partial least squares regression (PLSR) and random forest (RF). The yield estimation model constructed with color indexes at the flagging and flowering stages, along with texture characteristics and fusion indexes at the filling stage, had the best accuracy, with R2 values of 0.70, 0.71 and 0.76 and RMSE values of 808.95 kg/hm2, 794.77 kg/hm2 and 728.85 kg/hm2, respectively. The accuracy of winter wheat yield estimation using PLSR at the flagging, flowering, and filling stages was better than that of RF winter wheat estimation, and the accuracy of winter wheat yield estimation using the fusion feature index was better than that of color and texture feature indexes; the distribution maps of yield results are in good agreement with those of the actual test fields. Thus, this study can provide a scientific reference for estimating winter wheat yield based on UAV digital images and provide a reference for agricultural farm management.

Current techniques to estimate crop aboveground biomass (AGB) across the multiple growth stages mainly used optical remote-sensing techniques. However, this technology was limited by saturation of the canopy spectrum. To meet this problem, this study used digital images obtained by an unmanned aerial vehicle to extract the spectral and structural indicators of the crop canopy in three key potato growth stages. We took the color parameters (CP) of assorted color space transformations as the canopy spectral information, and crop height (CH), crop coverage (CC), and crop canopy volume (CCV) as the canopy structural indicators. Based on the complementary advantages of CP and CCV, we proposed a new metric: the color parameter-weighted crop-canopy volume (CCVCP). Results showed that the CH, CCV, and CCVCP correlated more strongly with potato AGB during the multi-growth stages than do CP and CC. The hue-weighted crop-canopy volume (CCVH) correlated most strongly with the potato AGB among all structural indicators. Using CH was more accurate in estimating potato AGB compared to CP and CC. Combining indicators (CP + CC/CH, CP + CC + CH) improved the accuracy of potato AGB estimation over the multi-growth stages. Except for the CP + CC + CH model, other AGB estimation models produced inaccurate AGB estimation than the models based on CCV and CCVH. The AGB estimation accuracy produced by the univariate-based CCVH model (R2 = 0.65, RMSE = 281 kg/hm2, and NRMSE = 23.61 %) was comparable to that of the complex model [CP + CC + CH using random forest (RF) or multiple stepwise regression (MSR)]. Compared with CP + CC + CH using RF and MSR, the RMSE decreased and increased by 0.35 % and 4.24 %, respectively. Compared with CP, CP + CC, CP + CH, and CCV, the use of CCVH to estimate AGB decreased the RMSE by 10.24 %, 7.42 %, 6.36 %, and 6.33 %, respectively. Meanwhile, the performance of CCVH was verified at different stages and among varieties. Thus, this indicator can be used for monitoring potato growth to help guide field production management.

快速、准确地监测冬小麦生物量,对于冬小麦田间管理、产量预测等具有重要意义.使用 2015 年开花期的冬小麦无人机数码影像及相应的生物量数据,将相关系数(|r|)、灰色关联分析(GRA)、投影变量重要性(VIP)与遗传算法(GA)-BP神经网络进行整合,构建了3种开花期冬小麦生物量估算模型,并对这 3 种模型进行可视化空间分析.结果表明:|r|-GA-BP,GRA-GA-BP,VIP-GA-BP模型的决定系数R2 分别为0.753 9,0.689 8,0.704 4,RMSE分别为763.3,908.8,836.9 kg·hm-2,MAPE分别为 10.31%,15.65%和 12.55%,|r|-GA-BP比GRA-GA-BP和VIP-GA-BP对冬小麦生物量有更好的预测能力.经可视化处理后能较为直观地反映冬小麦生物量的空间分布状况,为冬小麦的生长监测提供技术支持.

Nitrogen nutritional status is an important parameter for crop growth,and how to accurately monitor it is particularly important.In this paper,the winter wheat leaf reflection spectrum data and the corresponding leaf nitrogen content and leaf nitrogen accumulation data at the experimental base of Beijing Academy of Agricultural and Forestry Sciences in 2013-2014 were used to screen out the wave bands which are sensitive to leaf nitrogen content and leaf nitrogen accumulation by the combination of bands,and sensitive wave bands were used to establish nitrogen nutrition diagnosis model of winter wheat in each growth period and whole growth period.The results show that:（1） The sensitive spectral parameters of leaf nitrogen content and leaf nitrogen accumulation screened by the combination of bands are NDSI （564,728） ,NDSI （543,728） ,RSI （564,728） and RSI （543,728） ;（2） Among the nitrogen nutrition diagnosis models constructed in each growth period and the whole growth period,the stability and reliability of LNC in each growth period is better than LNA,and the leave-one-out cross validation method also shows that the accuracy of LNC is relatively higher.This study has shown that the nitrogen content of leaves can be used to monitor the nitrogen nutrition status of winter wheat,so as to achieve precise management of nitrogen fertilizer.

黄河流域是我国重要的矿产资源分布区域.长期以来,矿区的高强度开采引发了水土流失、植被退化等一系列问题,而植被作为矿区生态系统的能源动力,有着巨大的固碳速率和潜力,精准获取及监测矿区植被生长状态对黄河流域生态保护与高质量发展具有重要意义.提出一种无人机密集匹配点云矿区植被自动提取方法,通过无人机搭载的数码相机获取矿区序列影像,经特征提取、空三测量、多视影像密集匹配,重建矿区三维点云,利用匹配点云具有丰富的地物光谱特性,构建点云的差异植被指数DEVI,通过Otsu阈值法自动求取全局阈值,从而实现矿区植被点云的自动提取.选取黄河流域河南段某矿区进行实验,实验结果表明:植被提取的总体精度为96.40％,Kappa系数为0.9271,可实现无人机密集匹配点云中的植被立体信息有效提取,为基于低成本无人机摄影测量进行矿区植被立体监测研究提供一种可行方法.

The nitrogen content of crops affects the growth status of crops. A suitable nitrogen content can greatly improve the growth and yield of crops. Therefore, it is very important to monitor nitrogen content quickly. This study aimed to explore the potential of combining vegetation indices and spectral feature parameters acquired by UAV imaging hyperspectral to improve the accuracy of nitrogen content estimation during key growth stages of winter wheat. Firstly, the UAV was used as a remote sensing platform with hyperspectral sensors to acquire hyperspectral remote sensing images of four major growth stages of winter wheat: plucking, flag picking, flowering, and filling stages, and the nitrogen content data of each growth stage were measured. Secondly, based on pre-processed hyperspectral images, we extracted the canopy reflectance data of winter wheat at each growth stage. As a result, we constructed 12 vegetation indices and 12 spectral feature parameters that can better reflect the nitrogen nutrient status of the crop. Then, the correlation between the spectral parameters and the nitrogen content of winter wheat was calculated, and vegetation indices and spectral feature parameters with a strong correlation with the nitrogen content in each growth period were screened out. Finally, a nitrogen content estimation model based on vegetation indices and vegetation indices combined with spectral feature parameters was constructed using Stepwise Regression (SWR) analysis. The results showed that (1) most of the selected vegetation indices and spectral feature parameters were highly correlated with the N content of winter wheat. Among them, the correlation of vegetation indices was higher than that of spectral feature parameters; (2) although it is feasible to estimate winter wheat based on individual vegetation indices or spectral feature parameters, the accuracy needs to be further improved. (3) compared with a single vegetation index or spectral feature parameter, the accuracy and stability of the nitrogen content estimation model constructed by vegetation index combined with spectral feature variables using the SWR method were higher (at the plucking stage: modeling R-2 = 0.64, RMSE = 24.68% NRMSE= 7.96% validation R-2 = 0.77, RMSE =23.13% NRMSE= 7.81%; flag picking phase: modeling R-2 = 0.81, RMSE= 15.79% NRMSE= 7.41%, validation R-2 = 0.84, RMSE= 15.10%, NRMSE= 7.08%; flowering phase: modeling R-2 = 0.78, RMSE= 9.88% NRMSE= 5.66%, validation R-2 = 0.85 RMSE = 9.12% NRMSE = 4.76%; filling stage: modeling R-2 = 0.49 RMSE = 13.68% NRMSE = 9.85% validation R-2 = 0.40, RMSE= 18.29% NRMSE= 14.73%). The results showed high accuracy and stability of the winter wheat N content estimation model constructed by combining vegetation indices and spectral feature parameters obtained by UAV imaging hyperspectral. The research results can provide a reference for the spatial distribution and precise management of winter wheat N content.

为了寻求高效的马铃薯生物量估算方法,该研究利用2017年北京小汤山地区幼苗期、块茎形成期、块茎增长期、块茎增长后期和淀粉积累期的马铃薯生物量和对应的地面非成像高光谱数据,构建了植被指数NDVI和RVI与生物量的决定系数R2等高线图,利用重心公式分析了不同区域对生物量的敏感波段,最后用经验回归分析方法构建了马铃薯干生物量遥感估算模型.结果表明抛物线模型构建的NDVI(382,669)和RVI(385,668)可以很好地估算地上部干生物量,其建模的R2均为0.545,验证精度RMSE均为0.054 kg/m2,MAE分别为0.040、0.039 kg/m2.该方法可提高模型的稳定性和准确性,为快速无损诊断马铃薯地上部干生物量提供新的波段选择方法.

以冬小麦为研究对象,利用无人机搭载UHD185相机获取了挑旗期和开花期的高光谱影像,并同步采集了各小区的植株氮含量信息,结合相关性和方差膨胀因子,筛选了对植株氮含量敏感、植被指数之间共线性弱的植被指数,最后用多元线性回归、偏最小二乘回归和逐步回归算法3种方法探究冬小麦氮含量的较优高光谱反演模型.结果表明,无论是挑旗期还是开花期,偏最小二乘回归构建的植株氮含量模型估算精度高于逐步回归和多元线性回归构建的植株氮含量估算模型,验证结果同样表明偏最小二乘回归的均方根误差为最小,将该模型应用于无人机高光谱影像上,可以为田块尺度的精准施肥提供参考.

针对当前氮营养指数研究中对植被指数相关性考虑不足的问题,以挑旗期冬小麦为研究对象,提出品种、水和氮耦合实验对农作物氮素进行精准监测与反演的方法.该方法根据相关性程度和膨胀系数,选取相关性好和共线性强的植被指数,结合偏最小二乘法和BP神经网络构建氮营养指数模型,以决定系数和均方根误差为评价指标对模型精度评价.实验结果表明:偏最小二乘建模和验证的R2分别为0.6815和0.6815,RM SE分别为0.2840和0.2125;BP神经网络建模和验证的R2分别为0.9352和0.7484,RMSE分别为0.2677和0.2163.因此,BP神经网络可以更好地进行冬小麦氮素营养状况估算,反演后影像能较为直观地反映冬小麦氮素营养状况.该研究可为冬小麦氮素无损检测、掌握作物长势情况提供参考.

Chlorophyll content (SPAD) is a vital index for crop growth evaluation, which can monitor the growth of crops and is crucial for agricultural management, so it is important to estimate SPAD quickly and accurately. In this study, the remote sensing images of the jointing, flagging, and flowering stages were acquired using UAV hyperspectral for winter wheat. The vegetation indices and red edge parameters were extracted to explore the ability of vegetation indices and red edge parameters to estimate SPAD. Firstly, the vegetation indices and red edge parameters were correlated with the SPAD of different fertility stages. Then, the SPAD was estimated based on the vegetation indices, vegetation indices combined with red edge parameters , and using partial least square regression (PLSR) method. Finally, the SPAD distribution map was produced to verify the validity of the model. The results showed that (1) most of the vegetation indices and red edge parameters were correlated with SPAD at highly significant levels (0. 01 significant) in all three major reproductive stages; (2) the SPAD estimation model constructed from individual vegetation index had the best performance for LCI among vegetation indexes (best R-2 = 0. 56 , RMSE= 2. 96, NRMSE=8. 14%) and Dr/Dr min performed best (best R-2 = 0. 49 , RMSE= 3. 18, NRMSE= 8. 76%) ; (3) SPAD estimation model based on vegetation indices combined with red edge parameters was the best and better than SPAD estimation model based on vegetation indices only. Meanwhile, both models reached the highest accuracy at the flowering stage as the fertility stage progressed, with R-2 of 0. 73 and 0. 78, RMSE of 2. 49 and 2. 22, and NRMSE of 5. 57% and 4. 95% , respectively. Therefore, based on the vegetation indices combined with the red edge parameters, using the PLSR method can improve the estimation effect of SPAD, which can provide a new method for SPAD monitoring based on UAV remote sensing, and also provide a reference for agricultural management.

获取冬小麦挑旗期的无人机数码影像数据及生物量数据,利用相关系数与灰色关联度分析了植被指数与生物量的关联程度,利用方差膨胀因子结合相关系数及灰色关联分析筛选出最佳植被指数,最后利用多元线性回归和主成分分析对冬小麦生物量进行遥感估算反演及可视化分析.结果表明:利用方差膨胀因子结合相关系数分析筛选出的最佳数码影像指数为b、(r-g-b)/(r+g)、EXGR和g/b;利用灰色关联分析筛选出的最佳数码影像指数为(r-g-b)/(r+g)、EXG、b和RGBVI.利用无人机数码影像数据进行生物量估算时,结合灰色关联分析的多元线性回归建模方法精度最高,建模的R2和验证的RMSE分别为0.57和1.773 t/hm2.该方法可以有效提高冬小麦生物量的反演精度,为监测冬小麦长势提供了一种有效的思路.

[目的]准确、快速地获取作物的氮素信息,对监测作物氮素营养状况、指导变量施肥具有重要意义.[方法]获取了冬小麦挑旗期及开花期数码影像和相应的冬小麦地面农学参数,首先分析了数码图像指数与氮营养指数的相关性,然后结合相关系数和方差膨胀因子,筛选出对氮营养指数敏感且图像指数间不存在共线性的图像指数,通过偏最小二乘法建立各生育期氮营养诊断模型,并利用挑旗期和开花期的诊断模型对无人机影像进行填图和可视化.[结果]结合相关系数和方差膨胀因子筛选出挑旗期的图像指数分别是b、g/b、(r-g-b)/(r+g)、NDI、WI,筛选出开花期的图像指数分别是b、r/b、(r-g-b)/(r+g)、VARI.就生育期而言,开花期建模的决定系数比挑旗期的决定系数高0.008 8,均方根误差低0.021 7,开花期可以较好地反应冬小麦氮素营养状况.[结论]挑旗期和开花期的数码影像,经填图和可视化处理后得到的氮营养指数分布图能较好地监测不同生育期氮素营养状况,为田间小麦氮素营养状况监测提供高效的技术手段.