基于Copula 函数的边坡可靠度高效分析方法

将Copula函数与高效蒙特卡洛方法结合,提出了含相关随机变量的边坡可靠度高效分析方法。以两个岩质边坡稳定性问题为例验证了所提方法的有效性。结果表明,该方法相比于直接蒙特卡洛方法在保证边坡失效概率的准确性的同时计算效率更高。Copula函数可以构造出具有不同相关结构的岩土体参数的联合概率分布,与高效蒙特卡洛方法(即蒙特卡洛重要抽样方法和子集模拟)结合能高效地处理含多种相关随机变量的边坡可靠度计算问题,相比于现行的Nataf变换方法结果更能体现岩土边坡真实稳定性。此外,该方法也能高效地计算含有复杂的隐式功能函数的边坡可靠度,研究成果拓展了高效蒙特卡洛方法在边坡可靠度分析中的应用。     

Relationships of Spectral Reflectance with Plant Tissue Mineral Elements of Common Bean (Phaseolus Vulgaris L.) Under Drought and Salinity Stresses

ABSTRACT

Salinity and drought stresses are critical for Phaseolus vulgaris L. growth and development. They affect plants in various ways, including tissue mineral element content. Micro- and macro-elements in leaves of Phaseolus vulgaris L. (cv. ‘Blue lake’ and cv. ‘Terli’) subjected to deficit irrigation and salinity treatments were investigated, both analytically and with regards to their effect on the leaves’ spectral reflectance. B (boron), K (potassium), Mn (manganese), Na (sodium), Si (silicon) and Zn (zinc) appeared to be influenced by stress factors, mainly responding to salinity increase. The leaf spectral reflectance of the plants appeared to be significantly correlated with most of the elements under investigation. Multivariate regression identified a relationship of the reflectance at particular regions of the spectrum with phosphorus and NDVI (normalized difference vegetation index) and indicated a significant correlation with B, Fe (iron), K, Mn, P (phosphorus) and Zn. Moreover, customized spectral indices, exhibiting significantly high correlation with B, Fe, K, Mg (magnesium), Mn, Na, P, Zn and N (nitrogen), were developed.     

猕猴桃叶片叶绿素含量高光谱估算模型研究

【目的】研究猕猴桃叶片叶绿素含量的高光谱估算方法,为猕猴桃长势的遥感监测提供理论依据。【方法】以陕西杨凌蒋家寨村2018年不同生育期(初花期、幼果期、膨果期、壮果期、果实成熟期)的猕猴桃叶片为研究对象,分别测定其高光谱反射率和叶绿素含量(SPAD值),分析原始光谱和5个常见的植被指数(归一化植被指数、归一化叶绿素指数、改进的叶绿素吸收反射率指数、MERIS地面叶绿素指数、土壤调整指数)与叶绿素含量之间的相关关系,提取各生育期的特征波段,分别建立基于特征波段和植被指数的单波段叶绿素含量一元线性估算模型。利用主成分分析对原始光谱数据进行降维,将得到的主成分得分作为随机森林模型的输入变量,建立基于多波段信息的叶绿素含量多元估算模型,并对模型进行精度验证和分析。【结果】不同生育期猕猴桃叶片光谱反射率变化趋势基本一致,整体趋势为可见光波段反射率低,近红外波段反射率高;在可见光波段,光谱反射率随着叶绿素含量的升高而降低;在近红外波段,光谱反射率则随着叶绿素含量的增加而升高。通过相关性分析可知,初花期、幼果期、膨果期、壮果期、果实成熟期原始光谱的特征波段分别为729,548,707,707和712 nm,估算模型决定系数(R~2)分别为0.18,0.85,0.54,0.85和0.82,其中初花期估算模型未通过显著性检验,其余生育期均通过极显著性检验。在5个常用植被指数中,初花期与叶绿素含量相关性最高的是归一化叶绿素指数(NPCI),但是估算模型决定系数R~2只有0.1,未通过显著性检验;其他生育期与叶绿素含量相关性最高的是MERIS地面叶绿素指数(MTCI),所建立的估算模型拟合效果好,预测精度高。基于主成分分析和随机森林回归建立的不同生育期猕猴桃叶片叶绿素含量估算模型的R~2在0.91～0.98,均通过极显著性检验,其拟合效果和预测精度远高于单波段一元线性回归和基于植被指数的一元线性回归模型,是估算猕猴桃叶片叶绿素含量的最优模型。【结论】基于主成分分析的随机森林模型包含了更完整的波段信息,对不同生育期猕猴桃叶片叶绿素含量具有较好的预测能力。     

基于机器学习算法的土壤有机质 质量比估算

为快速高效地估测干旱、半干旱地区土壤有机质(soil organic matter, SOM)质量比,提出了一种结合竞争适应重加权法(CARS)和随机森林(RF)的估测模型.以内陆干旱区艾比湖流域为研究区,测定土壤高光谱反射率和SOM质量比,经预处理后,利用CARS对原始光谱(R)、一阶导数(R′)、吸光度(log(1/R))及吸光度一阶导数［log(1/R)］′4种光谱变量的可见-近红外光谱进行筛选,并结合RF算法,建立全谱段RF模型与CARS-RF模型.结果表明,基于CARS方法对光谱进行变量筛选后,得出4种光谱变量的优选变量集个数分别为35,26,34和121;在4种光谱变量中,R′和［log(1/R)］′的SOM估测模型精度较高,以［log(1/R)］′为基础数据获得的模型精度最高;CARS-RF模型精度优于全谱段RF模型,模型验证集决定系数(R²)、均方根误差(RMSE)、相对分析误差(RPD)分别为0.881,6.438 g/kg和2.177.该研究在预处理的基础上通过变量优选,应用较少的变量个数获得较高的估测精度,为干旱、半干旱区SOM高光谱估测提供了适宜高效的方法.     

基于无人机高光谱遥感的柑橘患病植株分类与特征波段提取

目的 结合传统与现代农业病虫害监测的优缺点，探索通过无人机高光谱遥感技术检测出患病的柑橘植株、通过人工田间调查方式判断其患病种类及患病程度的病虫害监测方法。方法 使用无人机获取原始高光谱图像，经过光谱预处理和特征工程后，采用连续投影算法提取对柑橘患病植株分类贡献值最大的特征波长组合，基于全波段使用BP神经网络和XgBoost算法、基于特征波段使用逻辑回归和支持向量机算法，建立分类模型。结果 基于全波段的BP神经网络和XgBoost算法的ROC曲线下面积(Area under curve，AUC)分别为0.883 0和0.912 0，分类准确率均超过95%；提取出698和762 nm的特征波长组合，基于特征波长使用逻辑回归和支持向量机算法建立的分类模型召回率分别达到了93.00%和96.00%。结论 基于特征波长建模在患病样本分类中表现出很高的准确率，证明了特征波长组合的有效性。本研究结果可为柑橘种植园的病虫害监测提供一定的数据和理论支撑。     

Detection and classification of Lepeophterius salmonis (Krøyer, 1837) using underwater hyperspectral imaging

Salmon louse, or sea lice, (Lepoptherius salmonis) represents practical, economical and fish welfare challenges for salmon farming (Hamre et al., 2013) and for the free-living stocks of salmon. There is an urgent need in the industry for a system that provides reliable numbers of louse on farmed salmon. Underwater hyperspectral imaging represents a potential new technique for louse counting in sea cages. In laboratory studies, the UHI technology could detect and classify pre-adults (pre-adult I and II), adult males and adult females (ovigerous) of sea lice based on the difference in their spectral characteristics. A model was built for detection of lice on the salmon and the UHI had a detection success ranging from 67 to 100 % with an average of 82%. A classification of the detected lice was performed for pre-adults, adult males and ovigerous lice and had a prediction accuracy of 85% when lice were divided into three groups (pre-adults, adult male and ovigerous lice) and 93% when lice were divided in two groups, ovigerous lice and all the other mobile lice (pre-adults and adult male). An automatic procedure for in situ measurements of louse infected salmon could deliver a data basis several times higher than the traditional counting system. The next generation of UHI louse detector should be developed with a higher spatial resolution to be able to detect also the sessile stages of lice. For succeeding with in situ classification of L. salmonis, correction algorithms to compensate for the impact of water between the UHI and lice need also to be developed.     

水稻工厂化育秧生长信息智能模拟采集系统的设计

水稻工厂化育秧可有效提高生产效率,确保国家粮食安全。为了给水稻工厂化育秧生长管理提供更为迅速、可靠的信息判决依据和决策支持,设计了一种水稻工厂化育秧生长信息智能模拟采集系统,包括针对水稻秧苗生长情况的高光谱图像信息采集系统和监测水稻生长环境各项参数指标的环境信息采集系统(CO2采集单元、温度采集单元、土壤水分采集单元、光照强度采集单元)。同时,采用Keil C语言开发系统的工作软件,以AT8 9 S5 2单片机为控制器、LCD1 6 0 2为液晶显示器等设计了系统正常工作的软硬件系统。     

Nutritive value of sorghum silage for sheep

This study compared ensiled sorghum (ES) from two varieties (first and second cuts) with corn silage (CS) for chemical analysis, total-tract nutrient digestibility, ruminal variables and blood biochemistry metabolites in sheep. Five experimental silages representing first cut ES var. Pegah (PS1), second cut ES var. Pegah (PS2), first cut ES var. Speedfeed (SS1), second cut ES var. Speedfeed (SS2) and CS were assessed for chemical analysis, silage fermentation and anti-quality compounds. Diets consisted of lucerne hay, silage and a premix (at the ratio of 53.5:46:0.5 on a dry matter [DM] basis). Feed consumption, total-tract nutrient digestibility, microbial nitrogen supply (MNS), ruminal variables and blood biochemistry metabolites were determined in a Latin square design of five periods with five male sheep. Compared to sorghum, CS had lower (p < 0.01) concentrations of ash, water-soluble carbohydrates and crude protein, but higher starch, lignin (sa), non-structural carbohydrates and apparent nutrient digestibility. The measurements of pH, lactic acid, NH₃-N and volatile fatty acids (VFA) confirmed that all silages were well-preserved. The anti-quality compound levels in the silages were below the safe limits for ruminants. No differences were observed in feed consumption (p > 0.05) and total-tract digestibility of DM and nutrients (p > 0.05). Rumen pH, NH₃-N, VFA concentrations and acetate to propionate ratio showed no difference among treatments (p > 0.05). Sheep fed on sorghum silage showed a decrease (p < 0.01) in total protozoa, and particularly in the Entodiniae population. There were no differences in cellulolytic bacteria counts (p > 0.05) among treatments. Urinary N, MNS and blood biochemistry metabolites showed no difference among treatments (p > 0.05). Silages from both cuts of sorghum had comparable total-tract nutrient digestibility to CS, therefore, these varieties could be used as a more drought resilient silage source in the feeding of sheep.     

基于SD模型的长沙市水资源承载力研究

基于长沙市水资源现状,为平衡居民生活用水、经济生产用水和环境生态用水之间的关系,运用系统动力学方法,建立长沙市水资源-人口-经济-生态环境的耦合系统,模型模拟了现状延续型、经济发展型、节水治污型和综合协调型四种发展模式,选取GDP、人口数量、居民生活需水量、经济需水总量、供需压力作为主要评价指标。结果表明:第一、二、三产业分别保持10%、12%、12%年增长率且节流治污情况与现状延续型保持一致的方案,供需压力将达到节流治污方案的两倍多;经济生产用水对水资源供需压力的影响远大于居民生活用水和环境生态用水;通过适当降低人均用水量来平衡因人口数量增长而增加的居民生活用水是可行的。     

基于混合遗传算法的谐波齿轮传动优化设计

以谐波齿轮传动体积最小为优化设计目标,建立了混合离散变量优化设计数学模型。引入混沌移民算子,对基本遗传算法进行了改进,以维持群体的多样性,提高全局收敛能力。提出了一种设计变量的离散化处理方法,结合染色体编码的修正技术,在遗传算法中染色体按离散化后的设计变量进行离散搜索。开发了混合离散变量优化的改进遗传算法程序,给出了谐波齿轮传动的混合离散变量优化设计实例,所得优化设计参数符合设计规范,不需进一步处理就可直接用于制造。