基于Hyperion影像的玉米冠层叶绿素含量估算

采用高光谱卫星数据进行玉米叶片和冠层尺度的叶绿素含量估算,对现代农业技术的发展有重要意义。首先,采用以α为倾斜度参数的双曲正切S型函数为基础的误差反向传播(back propagation,BP)算法前馈神经网络(hyperbolic tangent sigmoid function-back propagation,Htsf-BP)构建叶片尺度的叶绿素含量高光谱遥感估算模型;以几何光学辐射传输模型(4-scale模型)为理论依据,对叶片和冠层尺度的光谱转化函数进行推导,实现Hyperion影像冠层尺度光谱到叶片光谱的转化,同时获取叶片尺度叶绿素含量估算结果;最后,结合叶面积指数(leaf area index,LAI)进行冠层尺度叶绿素含量估算。结果表明:当隐含层结点数为6时,Htsf-BP神经网络法对叶绿素的估算精度最高,验证精度达78.68%;在波长750与980 nm处,采用光谱尺度转化方程进行模拟的冠层光谱与实测冠层光谱间的相关系数R2值分别达到了0.784和0.706;实测叶片尺度叶绿素含量与模拟结果间的相关系数R2值达0.726。该方法可为高精度快速估算叶片和冠层尺度玉米叶绿素含量提供参考。     

基于人工神经网络的土壤有机质含量高光谱反演

研究了土壤有机质含量与土壤高光谱之间的关系,在对原始光谱进行了预处理分析后,运用多元线性逐步回归法(MLSR)和人工神经网络法(ANN)建立了土壤有机质含量的反演模型,并对模型进行了验证。结果表明:人工神经网络所建立的反演模型普遍优于回归模型,网络集成模型优于单个BP网络模型,网络集成是提高反演模型准确性与稳定性的有效途径。网络集成模型为最优模型,总均方根误差为1.31,可以用于土壤有机质含量的快速测算。     

基于遗传神经网络的玉米叶色的自动测定研究

利用计算机图像处理技术和遗传神经网络技术,建立了一个多层前馈神经网络,实现了大田玉米和背景图像的正确识别,并且通过获取玉米叶的色度直方图提取了玉米叶表面颜色特征,进而求得了玉米叶色的测定值。实验结果表明,玉米叶色值自动测定系统,识别玉米的准确率可达91.6%,可以有效地测定玉米的叶色。该研究为实现大田玉米的化肥精确施用提供了理论依据。     

水稻叶片氮素及籽粒蛋白质含量的高光谱估测模型

研究水稻叶片氮素和籽粒蛋白质含量的高光谱快速、无损监测方法,对于水稻营养诊断、籽粒品质监测及氮肥高效利用具有重要意义。本文通过水稻盆栽试验,测定水稻叶片氮素、籽粒蛋白质含量和冠层光谱,采用不同的光谱建模方法来提高氮素、籽粒蛋白质含量的估测精度。先用主成分分析(PCA)方法进行特征波段的提取,再用多元线性回归(MLR)、人工神经网络(ANN)和偏最小二乘回归(PLSR)进行建模。结果表明,水稻叶片氮素和籽粒蛋白质含量与特征光谱存在很好的模型关系,3种模型预测的决定系数(R2_p)均在0.847以上,并以PLSR模型的预测效果为最好,可以实现水稻氮素营养和籽粒品质的高光谱估测。     

基于高光谱图像处理的大豆品种识别(英文)

大豆组分(油,蛋白质,脂肪等)在不同的大豆品种间差异很大。对于提高大豆品质来说,大豆品种识别是一个关键因素。该文利用高光谱图像技术对不同的大豆品种进行识别。利用高光谱成像系统获取大豆样本1 000～2 500 nm范围的光谱反射数据;应用主成分分析法(PCA,principal component analysis)对获取到的光谱数据进行数据降维并去除冗余数据;在分类算法中将得分高的主成分值作为输入特征,通过PCA方法从每个特征图像中提取4个特征变量(能量、熵、惯性矩和相关性);对于具体特征提取,从16个特征变量中提取8个重要特征参数;根据选择的特征,应用神经网络方法构建分类器;训练精度精度达到97.50%,平均测试精度达到93.88%以上。结果表明,应用高光谱图像技术结合神将网络建模方法可以对大豆品种进行分类。     

Non-destructive estimation of chlorophyll and nitrogen content in leaf of Rosa damascena (Mill)

Chlorophyll (Chl) and nitrogen (N) status of leaves provide valuable information about the physiological condition of plants. The conventional methods for measuring Chl and N contents in leaves are destructive, costly, time-consuming, and do not allow repetitive measurement of the same sample. The Damask Rose (Rosa damascena Mill) is an important aromatic crop in the western Himalaya region in India. Generally, flower yield and oil yield of the Damask rose are correlated with nitrogen, phosphorus, and potassium (NPK) levels in the leaf at the bud development stage. The dynamics of N within the rose plant have not been reported clearly. Thus, there is a pressing need for non-destructive techniques to estimate Chl and N content in the leaf of the Damask rose. Our objective was to establish an appropriate mathematical relationship between the Chl content index (CCI) value and the total Chl/N contents for non-destructive estimation of total Chl and N in the leaf of the Damask rose. The regression models were developed with destructively measured parameters (total Chl and N) as the dependent variable and a parameter derived from CCM-200 as the independent variable (CCI). We found that polynomial regression models are suitable for non-destructive estimation of total Chl, and the model predicted values were very close to traditionally measured values with a root mean square prediction error (RMSEp) less than 0.20?mg?g^?1 of Chl. In the case of N estimation, a power regression model was appropriate with lowest Akaike's information criteria (AIC) and root mean square validated error (RMSEv) value. Significant correlations (P?≤?0.001) were observed between traditionally measured values and our model predicted values in both cases.     

基于BP神经网络模型的荔枝树叶面积测定方法

为了准确、快速地测定荔枝树叶面积，设计了一个BP神经网络模型，输入参数为叶片长度和叶片最大宽度，输出参数为叶面积。用LI-3000A型叶面积仪测量所得到的样本数据对网络进行训练，测试样本的网络输出与网络目标的相关系数达0.99609，网络模型是有效的。用训练后的网络模型对10组未参加建模的样本数据进行叶面积测定，误差平方和为1.2929，优于回归方程法的2.511。训练好的BP神经网络模型可以在不破坏叶片的情况下，简单、快速、经济地测定大量的荔枝树叶片面积。     

BP神经网络算法在河西绿洲玉米生产碳排放评估中的应用及算法有效性研究

针对作物生产碳排放预测较为困难的实际问题,提出基于BP神经网络算法的玉米生产碳排放预测模型。选择地处河西走廊石羊河下游的民勤绿洲246家农户,面对面调查玉米种植户农场内生产投入数据,将玉米生产投入数据作为神经网络输入层;查阅和梳理国内外相似区域玉米生产环节碳排放系数,运用碳足迹生命周期法计算得到的碳排放值作为神经网络输出层;基于BP人工神经网络算法,运用试凑法确定网络隐含层节点个数,建立河西绿洲玉米生产碳排放预测模型,选择多元线性回归模型、多元非线性回归模型,对该模型有效性进行评估。研究结果表明,3层且各层节点数9、10、1的神经网络结构能够准确预测河西绿洲玉米生产碳排放,其碳排放预测值为0.763 kg(CO_2-eq)·kg~(-1)(DM);9-10-1结构的神经网络预测模型的相关系数(R~2=0.984 7)高于多元线性和非线性回归模型,该神经网络结构模型的均方根误差(RMSE=0.069 1)、平均绝对误差(MAE=0.051 3)均低于其他模型,BP神经网络算法预测性能明显优于其他预测模型。该研究为准确预测农业生产碳排放提供了新思路和可操作方法。     

基于遗传神经网络的黑龙江浅表地层水分预测

针对BP神经网络预测土壤墒情容易出现较大空间内存在局部极值点的问题,采用GA算法对BP网络进行优化,根据大豆作物在不同生长阶段的根系分布及吸水情况,划分3个不同发育阶段,5个地层深度,建立3种对应的土壤含水量遗传神经网络预测模型,并应用于黑龙江垦区红星农场大豆田间土壤水分预测,分别对3种模型的整体预测误差进行了分析,2009年大豆播种前期及其全生育期土壤体积含水量预测的平均绝对误差为1.83%,能较好地反映大豆田间土壤水分具体情况,为大豆节水灌溉与管理提供可靠的科学依据,该预测方法亦可为寒地大豆或其他农作物田间土壤水分预测提供借鉴。     

Estimation of Sunflower Seed Yield Using Partial Least Squares Regression and Artificial Neural Network Models

Statistical models can efficiently establish the relationships between crop growth and environmental conditions while explicitly quantifying uncertainties. This study aimed to test the efficiency of statistical models established using partial least squares regression(PLSR) and artificial neural network(ANN) in predicting seed yields of sunflower(Helianthus annuus). Two-year field trial data on sunflower growth under different salinity levels and nitrogen(N) application rates in the Yichang Experimental Station in Hetao Irrigation District, Inner Mongolia, China, were used to calibrate and validate the statistical models. The variable importance in projection score was calculated in order to select the sensitive crop indices for seed yield prediction. We found that when the most sensitive indices were used as inputs for seed yield estimation, the PLSR could attain a comparable accuracy(root mean square error(RMSE) = 0.93 t ha-1, coefficient of determination(R~2) = 0.69) to that when using all measured indices(RMSE = 0.81 t ha-1,R~2= 0.77). The ANN model outperformed the PLSR for yield prediction with different combinations of inputs of both microplots and field data. The results indicated that sunflower seed yield could be reasonably estimated by using a small number of crop characteristic indices under complex environmental conditions and management options(e.g., saline soils and N application). Since leaf area index and plant height were found to be the most sensitive crop indices for sunflower seed yield prediction, remotely sensed data and the ANN model may be joined for regional crop yield simulation.     

引入地形因子的黑土区大豆干生物量遥感反演模型及验证

为了对田块尺度农作物地上干生物量进行估测,提高大豆地上干生物量反演模型的精度和稳定性,该文获取了研究区地块2016年7、8月份的SPOT-6多光谱数据,并测定不同地形坡位的大豆地上干生物量,以归一化植被指数(normalized difference vegetation index,NDVI)和增强型植被指数(enhanced vegetation index,EVI)为输入量,建立田块尺度大豆地上干生物量一元线性回归模型;加入与地上干生物量相关的地形因子,建立逐步多元回归和神经网络多层感知反演模型.结果表明:1)使用传统的单一植被指数模型预测大豆地上干生物量有可行性,但模型精度和稳定性不高.2)加入地形因子(海拔、坡度、坡向)的神经网络多层感知器模型,有较高的精度和可靠性,模型准确度达到90.4％,验证结果显示预估精度为96.2％.反演结果与地块的地形、地貌、气温和降水特征基本吻合,反映了作物长势的空间分布特征,可以为田块尺度大豆地上干生物量动态监测和精准管理,提供借科学依据.     

Soil Particulate Organic Matter (POM) Prediction in a Mountainous Watershed using Artificial Neural Networks

Soil organic matter is a very important component of soil that supports the sustainability and quality in all ecosystems, especially in arid and semi-arid regions. A comparison study was carried out to verify when the artificial neural network (ANN) and multiple linear regression (MLR) models are appropriate for the prediction of soil organic matter (SOM) and particulate organic matter (POM). Discussions of advantages and disadvantages are given for both methods. Three different sets of easily available properties (soil properties alone, topographic and vegetation index, a combination of soil and topographic data) were used as inputs and the one affecting the model the most was determined. The smallest prediction errors were obtained by the ANN method; however, the prediction accuracies of the constructed MLR models using different data sets were closed to the ANN models in many cases.     

艾比湖流域不同土地覆盖类型土壤养分 高光谱反演模型研究

土壤养分影响着土壤的质量,也影响着植被、农作物等的生长。为快速准确地估测艾比湖流域土壤养分状况,选择艾比湖流域精河县作为研究区,以精河县内不同地表覆盖类型土壤为研究对象,基于实地采集的75个土壤样品的室内ASD Field Spec3实测光谱数据和3种光谱变换形式,利用10 nm间隔重采样进行去噪处理,再结合多元逐步回归法(SMLR)、偏最小二乘法回归法(PLSR)、人工神经网络法(ANN)分别建立土壤养分预测模型,以探索最优模型。结果表明:土壤实测光谱的一阶微分、二阶微分变换形式能显著提高光谱与土壤养分之间的相关性,尤其是一阶微分变换与土壤有机质和全氮的相关性最高分别达0.87和0.91,光谱变换技术能显著增强土壤养分与高光谱之间的敏感度,达到更好的建模效果;SMLR、PLSR和ANN这3种模型都具有良好的预测能力,其中,ANN建立的模型预测效果最好,二阶微分变换的ANN模型对有机质、全氮的预测决定系数(R2)分别为0.886和0.984,均方根误差(RMSE)分别为2.614和0.147,PLSR次之;全氮的预测效果明显优于有机质的预测效果,说明高光谱和全氮之间的敏感性更高。总体来说,光谱二阶微分变换形式的人工神经网络模型可以最精确稳定地完成土壤养分含量的快速预测,能够实现艾比湖流域的土壤养分空间分布状况和动态变化特征的动态监测。     

基于高光谱和BP神经网络的玉米叶片SPAD值遥感估算

为了进一步提高玉米叶绿素含量的高光谱估算精度,该文测定了西北地区玉米乳熟期叶片的光谱反射率及其对应的叶绿素相对含量（soil and plant analyzer development,SPAD）值,分析了一阶微分光谱、高光谱特征参数与 SPAD的相关关系,构建了基于一阶微分光谱、高光谱特征参数和 BP 神经网络的 SPAD 估算模型,并对模型进行验证;再结合主成分回归（principal component regression,PCR）、偏最小二乘回归（partial least squares regression,PLSR）以及传统回归模型与 BP 神经网络模型进行比较。结果表明：SPAD 值与一阶微分光谱在763nm 处具有最大相关系数（R=0.901）;以763 nm 处的一阶微分值、蓝边内最大一阶微分为自变量建立的传统回归模型可用于玉米叶片 SPAD 估算;将构建传统回归模型时筛选到的光谱参数作为输入,实测 SPAD 值作为输出,构建 BP 神经网络模型,其建模与验模 R2分别为0.887和0.896,RMSE 为2.782,RE 为4.59%,与其他回归模型相比,BP 神经网络模型预测精度最高,研究表明 BP 神经网络对叶绿素具有较好的预测能力,是估算玉米叶片 SPAD 值的一种实时高效的方法。     

地基激光雷达提取大田玉米植株表型信息

玉米个体表型信息对于玉米的高产高效发育规律研究、玉米遗传育种中基因型的确定具有重要意义,该文针对传统的玉米表型信息提取方法费时、费力、效率低下、主观性强等问题,提出一种基于TLS(terrestriallaserscanning,地面激光扫描)技术的大田玉米个体表型信息提取方法。利用地基激光雷达获取毫米级精度的玉米个体植株三维点云数据并进行海量点云数据预处理,构建玉米叶片三角网模型和叶片骨架点云;基于叶片三角网提取绿叶叶面积,基于叶片骨架点云提取叶长和叶倾角,基于未去穗的玉米植株点云提取株高。试验结果与实地手动测量值相比,真实叶面积、叶长、株高、叶倾角的均方根误差(RMSE)分别为12.69 cm~2、1.31 cm、1.30 cm和5.12°,平均绝对百分比误差(MAPE)分别为2.38%、1.32%、0.61%和8.96%。试验结果表明本文提出的基于TLS提取玉米个体表型参数的方法精度较高,具有可行性,为辅助玉米育种、生长监测等提供了一种有效手段。     

Performance of European Artificial Neural Network (ANN) Calibrations for Moisture and Protein in Cereals Using the Danish Near‐Infrared Transmission (NIT) Network

Three problems need to be addressed in networks of Infratec Grain Analysers: 1) the networks are not interconnected, 2) the partial least squares (PLS) calibrations used so far have to be individually adjusted for bias when transferred to the slave instruments, and 3) the calibrations are not entirely stable over time. Nonlinear artificial neural network (ANN) calibrations based on a large common European data set (≈4,000 samples in the training sets and ≈1,000 samples in the stop sets) were introduced to overcome these constraints. The performance of these ANN calibrations was compared with Danish PLS models for protein and moisture in cereals during the 1998 harvest in Denmark, and subsequently with PLS models based on the same European data set. ANN models were more accurate than PLS and, unlike PLS, were linear and transferable up to 25% moisture. It is suggested that the improved performance of the ANN models is attributable to the modeling technique rather than the size and nature of the European data set. In most cases, ANN models could be applied directly and without bias adjustment to slave instruments. The ANN models were also more stable, they required fewer bias adjustments or remodeling over time compared with Danish PLS models. ANN calibrations using shared data have been adopted for commercial use in several European countries and work is in progress to develop global ANN models for determination of protein in wheat and barley.     

基于CGA-BP神经网络的好氧堆肥曝气供氧量预测模型

为提高好氧堆肥曝气供氧量的曝气效率以及预测精度,该研究利用遗传算法（genetic algorithm, GA）对标准反向传播（back propagation, BP）神经网络的初始权值和阈值进行优化,再利用克隆选择算法（clonal genetic algorithm, CGA）优化遗传算法中的变异算子并复制算子,加快获取最优参数的速度,构建基于CGA-BP神经网络的曝气供氧量预测模型。为验证CGA-BP模型的有效性,与BP模型、GA-BP模型预测结果进行对比。试验结果表明：克隆遗传算法优化BP神经网络能加快获得最优解,效率相比BP模型和GA-BP模型分别提高了75.36%、51.30%;在曝气供氧量预测模型中,CGA-BP模型具有更准确的预测效果,预测精度为99.65%,而BP模型与GA-BP模型预测精度分别为96.99%、99.26%;CGA-BP模型评价指标的均方误差、平均绝对误差、平均绝对百分误差分别为0.003 4、0.038 9和0.350 6,均小于BP神经网络和GA-BP神经网络模型评价指标的误差;利用CGA-BP好氧堆肥曝气供氧量预测模型对好氧堆肥发酵过程进行精准...     

地形对漫川漫岗黑土区大豆产量的影响

为研究黑土区田块尺度地形对大豆产量造成的影响,在海伦东兴合作社具有明显地形起伏的地块,采集大豆田间试验数据,考虑温度、太阳辐射、坡度、土壤养分等因素,运用作物生长模型DSSAT(Decision Support Systemfor Agrotechnology Transfer)模型对各样点进行参数率定及验证,得出以下结论:1)DSSAT模型的模拟产量与实际产量的相对均方根误差为7.9%,模拟结果表现为优,表明运用作物模型模拟不同地形上的产量变异具有可行性;2)地形通过影响作物生长环境因子的时空差异决定产量差异,田块尺度温度、水分和坡度是影响产量差异的主要因素;3)坡顶和坡底的产量相对较高,且产量变异性较小,阳坡虽然接收到更多的光照,却由于水分胁迫造成减产,坡底和平缓坡顶水肥保持较好,易获得高产。研究成果为田间精细管理与田块尺度耕地高效利用提供科学依据。     

Development of a Near‐Infrared Imaging System for Determination of Rice Moisture

The objective of this study was to develop a near‐infrared (NIR) imaging system to determine rice moisture content. The NIR imaging system fitted with 15 band‐pass filters (wavelengths of 870–1,014 nm) was used to capture the spectral image. In this work, calibration methods including multiple linear regression (MLR), partial least squares regression (PLSR), and artificial neural network (ANN) were used in both near‐infrared spectrometry (NIRS) and the NIR imaging system to determine the moisture content of rice. Comprehensive performance comparison among MLR, PLSR, and ANN approaches has been conducted. To reduce repetition and redundancy in the input data and obtain a more accurate network, six significant wavelengths selected by the MLR model, which had high correlation with the moisture content of rice, were used as the input data of the ANN. The performance of the developed system was evaluated through experimental tests for rice moisture content. This study adopted the coefficient of determination (r_val²), the standard error of prediction (SEP), and the relative performance determinant (RPD) as the performance indices of the NIR imaging system with respect to the tests of rice moisture content. Utilizing these three models, the analysis results of r_val², SEP, and RPD for the validation set were within 0.942–0.952, 0.435–0.479%, and 4.2–4.6, respectively. From experimental results, the performance of NIR imaging system was almost the same as that of NIRS. Using the developed NIR imaging system, all of the three different calibration methods (MLR, PLSR, and ANN) provided a high prediction capacity for the determination of moisture in rice samples. These results indicated that the NIR imaging system developed in this study can be used as a device for the measurement of rice moisture content.     

Surface Application of Lime and Cover Black Oat and Corn Residues for No‐Till Soybean Production

Abstract: Crop residues that are left on the soil surface to serve as mulch can diminish the soybean response to surface application of lime under no‐till management by ameliorating soil chemical and physical attributes and the plant nutrition. A field experiment was performed in the period from 2000 through 2003 in Paraná State, Brazil, on a clayey‐sandy Rhodic Hapludox. Soil chemical attributes and soybean [Glycine max (L.) Merrill] nutrition, grain yield, and quality were evaluated after surface application of lime and covering with crop residues of black oat (Avena strigosa Schreb) and corn (Zea mays L.) under a no‐till system. Dolomitic lime was surface applied at the rates of 0, 2.5, 5.0, and 7.5 t ha^?1 on the main plots, and three treatments with vegetable covering were applied on the subplots: (i) without covering, (ii) with covering of corn straw, and (iii) with covering of corn straw and black oat residue (oat–corn–oat). After 30 months, surface‐applied lime increased soil pH and the exchangeable calcium (Ca²⁺) and magnesium (Mg²⁺) levels down to a 10‐cm depth, independent of the vegetable covering treatments. The black oat and corn residues on the soil surface increased the soil exchangeable K⁺ level at the 5‐ to 10‐cm depth. Liming increased leaf potassium (K) content and phosphorus (P) content in the soybean grain and reduced leaf zinc (Zn) content and manganese (Mn) content in the soybean leaf and grain. There was no effect of liming on soybean grain, oil, or protein yields, independent of the vegetable residues kept on the soil surface. The treatment with black oat covering and corn straw increased leaf N content, P content in the leaf and grain, and the contents of K, Mg, copper (Cu), and Zn in the soybean grain. It also increased soybean grain and protein yields. The corn straw left at the surface after harvesting was very important to the performance of the no‐till soybean.