基于VTCI和分位数回归模型的冬小麦单产估测方法

条件植被温度指数(VTCI)是一种综合了归一化植被指数(NDVI)与地表温度(LST)的遥感干旱监测方法,在关中平原的近实时干旱监测中具有其适用性。分位数回归能全面反映因变量的条件分布在不同分位数处的特征,回归结果稳健可靠。为了进一步研究VTCI干旱监测结果与小麦单产之间的关系及提高冬小麦单产估测精度,构建了不同分位数τ(0.1,0.3,0.5,0.7,0.9)下关中平原各市2008—2014年的冬小麦主要生育期VTCI与单产之间的线性回归模型,并基于中位数(τ=0.5)回归模型对研究区域的冬小麦单产进行了估测。结果表明,分位数回归模型比较全面地反映了不同分位数下冬小麦单产分布与VTCI之间的相关程度,弥补了最小二乘估产模型回归结果单一、易受异常值影响等的不足。中位数回归模型的单产估测结果与实际单产之间的相对误差和均方根误差的最小值及平均值均低于最小二乘回归模型,估测精度较高。此外,中位数单产估测模型获取的冬小麦估产结果在年际变化规律与空间分布特征上与实际产量均较相符,说明分位数回归在研究VTCI与产量之间的关系及冬小麦单产估测中具有其适用性与可靠性。     

基于SVR逆向模型的拖拉机导航纯追踪控制方法

针对行驶过程中轮胎侧滑量、路面性质等不确定性因素导致传统拖拉机二轮运动学模型难以准确描述拖拉机运动轨迹的问题,提出了一种基于SVR(Support vector regression)逆向模型的拖拉机导航纯追踪控制方法。采用粒度支持向量回归(Granular support vector regression,GSVR)方法建立了拖拉机前进航向的逆向模型,实时获得实际转弯曲率与运动学理论转弯曲率的函数关系,逆向模型对纯追踪导航模型输出进行,校正提高了纯追踪导航控制方法的适应性和动态性能。拖拉机导航系统的路径追踪路面试验结果表明:当行驶距离大于125 m、行驶速度为1.2 m/s时,直线追踪最大横向偏差小于0.061 4 m,较常规纯追踪模型导航方法具有更好的直线追踪性能;田间试验结果表明:该导航控制方法适用于2BFQ-6型油菜精量联合直播机自动对行作业。     

单品种淡水活鱼数量的被动声学估计

针对单品种淡水活鱼数量估计问题,以鲫鱼和鳊鱼为研究对象,采用不同数量淡水鱼的被动水声信号作为样本,通过4~6层小波包分解算法,对比分析了样本的若干特征提取方案,明确了鲫鱼水声信号特征的提取方案为短时平均过零率+6层小波包分解频段能量,而鳊鱼只需提取6层小波包分解频段能量作为其信号特征。根据不同样本集划分方法比较结果,确定了鲫鱼样本集划分方法为Rank-SPXY(m=10),鳊鱼为Rank-SPXY(m=5)。利用竞争自适应重加权采样,筛选出样本的关键特征,并使用多元线性回归和偏最小二乘回归建立了若干淡水活鱼的数量估计模型,结果表明:多元线性回归模型的性能较好,鲫鱼和鳊鱼数量估计模型的复相关系数分别为0.835和0.893,相对分析误差分别为1.79和2.01。     

基于介电频谱的采后苹果可溶性固形物含量无损检测

根据10~4500MHz间采后21周贮藏期间无损富士苹果介电参数的频谱特性,建立了苹果可溶性固形物含量的支持向量回归（SVR）预测模型和BP网络预测模型;并综合比较了采用原始频谱（FF）、主成分分析（PCA）和连续投影算法（SPA）优选频率对模型预测效果的影响。研究结果表明,PCA—SVR建模效果最好,其预测相关系数为0.883,均方根误差为0.552,PCA—BP的建模效果较PCA—SVR稍差。并且发现经SPA处理后的数据建立的模型,均方根误差普遍较小;经PCA处理后的数据建立的模型,预测相关系数普遍较高。     

基于随机森林模型的山体滑坡空间预测研究

滑坡灾害空间分布的准确预测是实现防灾减灾的重要途径。以2010年福建省顺昌地区滑坡资料为基础数据,分别应用随机森林模型和逻辑回归模型对福建顺昌地区山体滑坡发生与滑坡因子之间的关系进行实证分析,通过模型变量筛选、模型精度分析,探讨了随机森林模型在我国南方山体滑坡空间预测中的适应性。结果表明:随机森林模型对滑坡发生数据的拟合效果比逻辑回归模型好,其对顺昌地区滑坡发生数据的预测精度为90.8%,而逻辑回归模型的预测精度为81.8%;随机森林模型对研究区滑坡发生的泛化能力比逻辑回归模型好,其预测出高危险区和较高危险区所包含的滑坡比总和为66.05%,而逻辑回归模型为63.34%。研究结果表明随机森林模型的性能优于逻辑回归模型,可用于顺昌地区基于滑坡因子的未来滑坡发生的预测预报。     

柴油机无载测功法的试验研究

以4125A4柴油机为试验对象，运用多项式回归的方法，寻求柴油机功率与耗油量（最高空转速及额客转速）之间的内在定量关系。所建立的模型具有较高的准确性，相对误差可以控制在3%以内。无载测功方法简便、有效，对柴油机的田间测试具有较强的实用性。     

基于自然光照反射光谱的温室黄瓜叶片含氮量预测

利用便携式光谱辐射仪测量了自然光照条件下温室黄瓜叶片的光谱反射率,并计算了反射率光谱的一次微分光谱。反射率光谱以及一次微分光谱与叶片含氮量的相关分析表明,温室内光谱特性与叶片含氮量相关性最大的敏感波段分别是505～664 nm和685～722 nm。当利用原始光谱进行分析时,通过变量筛选得到了4个敏感波长,分别是568、596、640和664 nm。偏最小二乘回归分析(PLSR)以及归一化颜色指数(NDCI)分析都表明,建模时的相关系数R_C>0.800,模型验证时的相关系数R_V>0.700。对微分光谱进行的相关分析结果表明,利用单一敏感波长520 nm就可获得理想模型,建模时的相关系数为0.880,模型验证时的相关系数为0.787。对比原始光谱的PLSR模型与一阶微分光谱的一元线性回归模型可以得知,原始光谱以及一阶微分光谱都可用于温室内叶片含氮量的预测,而且一阶微分光谱在一些特殊的波长处具有更高的预测能力,这些模型将成为开发便携式作物长势诊断仪器的技术基础。     

土壤持水曲线van Genuchten模型求参的Matlab实现

土壤持水曲线是研究土壤水力学性质必不可少的 ,在已经建立的众多数学模型中 ,vanGenuchten模型是目前运用最广泛的模型 ,而运用该模型的关键是其 4个参数的求解。为此 ,本文对同一组东北褐土的土壤水吸力和对应的土壤含水量数据较详细地介绍了Matlab软件的非线性拟合和非线性回归函数的运用 ,得出了该土壤vanGenuchten模型的 4个参数值 ,分别建立了该土壤的vanGenuchten模型 ,并利用Mat lab强大的绘图功能对它们进行了直观比较。最后运用方差分析和残差分析对该模型的计算值与实测数据进行了分析 ,结果表明 :非线性拟合和非线性回归函数求参结果的显著水平均达到p<0 0 0 0 1,残差平方和均小于 0 0 0 0 5 ,其中非线性回归函数的求参结果较非线性拟合好。因此 ,运用Matlab软件的非线性拟合和非线性回归函数对土壤持水曲线的vanGenuchten模型进行求参是切实可行的 ,从而为土壤学工作者寻求出了一条运用数值计算方法的新途径。     

Methods to interpolate soil categorical variables from profile observations: Lessons from Iran

The paper compares semi-automated interpolation methods to produce soil-class maps from profile observations and by using multiple auxiliary predictors such as terrain parameters, remote sensing indices and similar. The Soil Profile Database of Iran, consisting of 4250 profiles, was used to test different soil-class interpolators. The target variables were soil texture classes and World Reference Base soil groups. The predictors were 6 terrain parameters, 11 MODIS EVI images and 17 physiographic regions (polygon map) of Iran. Four techniques were considered: (a) supervised classification using maximum likelihoods; (b) multinominal logistic regression; (c) regression-kriging on memberships; and (d) classification of taxonomic distances. The predictive capabilities were assessed using a control subset of 30% profiles and the kappa statistics as criterion. Supervised classification and multinominal logistic regression can lead to poor results if soil-classes overlap in the feature space, or if the correlation between the soil-classes and predictors is low. The two other methods have better predictive capabilities, although both are computationally more demanding. For both mapping of texture classes and soil types, the best prediction was achieved using regression-kriging of indicators/memberships (κ = 45%, κ = 54%). In all cases kappa was smaller than 60%, which can be explained by the preferential sampling plan, the poor definition of soil-classes and the high variability of soils. Steps to improve interpolation of soil-class data, by taking into account the fuzziness of classes directly on the field are further discussed.     

Empirical analysis and prediction of nitrate loading and crop yield for corn-soybean rotations

Nitrate nitrogen losses through subsurface drainage and crop yield are determined by multiple climatic and management variables. The combined and interactive effects of these variables, however, are poorly understood. Our objective is to predict crop yield, nitrate concentration, drainage volume, and nitrate loss in subsurface drainage from a corn (Zea mays L.) and soybean (Glycine max (L.) Merr.) rotation as a function of rainfall amount, soybean yield for the year before the corn-soybean sequence being evaluated, N source, N rate, and timing of N application in northeastern Iowa, U.S.A. Ten years of data (1994-2003) from a long-term study near Nashua, Iowa were used to develop multivariate polynomial regression equations describing these variables. The regression equations described over 87, 85, 94, 76, and 95% of variation in soybean yield, corn yield, subsurface drainage, nitrate concentration, and nitrate loss in subsurface drainage, respectively. A two-year rotation under average soil, average climatic conditions, and 125 kg N/ha application was predicted to loose 29, 37, 36, and 30 kg N/ha in subsurface drainage for early-spring swine manure, fall-applied swine manure, early-spring UAN fertilizer, and late-spring split UAN fertilizer (urea ammonium nitrate), respectively. Predicted corn yields were 10.0 and 9.7 Mg/ha for the swine manure and UAN sources applied at 125 kg N/ha. Timing of application (i.e., fall or spring) did not significantly affect corn yield. These results confirm other research suggesting that manure application can result in less nitrate leaching than UAN (e.g., 29 vs. 36 kg N/ha), and that spring application reduces nitrate leaching compared to fall application (e.g., 29 vs. 37 kg N/ha). The regression equations improve our understanding of nitrate leaching; offer a simple method to quantify potential N losses from Midwestern corn-soybean rotations under the climate, soil, and management conditions of the Nashua field experiment; and are a step toward development of easy to use N management tools.