Dense Cultivation and Fertilization for Higher Yield of Thyme (Thymus vulgaris L.)

Thyme ( Thymus vulgaris L.) plants were spaced at 15, 30 or 45 cm distances in a clay-loamy soil. They received different levels of nitrogen and phosphorus, besides a constant level of potassium. The results showed, that the wider spacing promoted the growth and production of herb and oil per plant, however dense cultivation significantly increased the yields of herb and oil per unit area. In all cases, the applied fertilization treatments significantly increased the productivity per unit area. Dense cultivation accompanied with higher levels of fertilization proved to be very usefull. The essential oil content was not influenced by either the plant spacing or fertilization treatments applied in this study.     

Comparison of nuclear and capacitance‐based soil water measuring techniques in salt‐affected soils

A field calibration experiment was carried out on salt‐affected clayey soil in Syria, to compare the sensitivity to soil electrical conductivity (ECe), and bulk density (ρ_b) of two instruments for estimating soil moisture: the neutron probe (NP) and the Diviner 2000 capacitance probe (CP). The results showed that the values of the correlation coefficient of the calibration were decreased when the ECe and ρ_b values increased; this decrease was more pronounced for the Diviner 2000, indicating that it was more sensitive to ρ_b and ECe than the NP. When only scaled frequency was used in the fitted equation, the Diviner 2000 in wet soil underestimated soil water content significantly at all depths, but especially in the top layer, by up to 0.09 cm³/cm³ compared with gravimetric determinations. However, in dry soil, the Diviner 2000 overestimated the volumetric water content by up to 0.05 cm³/cm³ in the top 15 cm, and by 0.03 cm³/cm³ at 30‐45 cm depth. The performance of the neutron probe was better overall; using a factory calibration curve no significant differences were observed between NP estimates and the gravimetric values. Including both ρ_b and ECe in the calibration equations improved the fits, although the regression coefficient (R²) for the Diviner 2000 remained low.     

Comparison between Calibration Procedure and Econometric Estimation

The two methods of estimating parameter in computable general equilibrium(CGE) model are introduced and compared:the calibration procedure and econometric estimation. The conclusions are:the estimation of parameter in CGE model must use the calibration procedure coupled with the econometric estimation method;the elasticity of output with respect to labor input,the marginal expenditure share for households and price elasticity of export demand are estimated by econometric estimation method;and other parameters of the CGE model can be get by calibration procedure.     

A Research on the Way of EMBET in the Tracking Systems'' Accuracy-analysis During Satellite Launch

In the past, the criterion for calibration has almost invariably been the stellar oriented ballistic camera,the method is a point to point comparison on the identical coordinate system. Deficiencies is follows: The request of weather is severe(cloudless night); The period is often long; the measure equipment of high precision can't be calibrated. While EMBET methods is used in calibration, it does not need criterion and it demands only that the measure elements is larger than 3,and the trajectory and many system errors of the measure elements can be simultaneously calculated with the measure data sequences. The error models of related equipments' measure elements and the EMBET of typical equipment group are given. The EMBET method is successfully used in Accuracy-analysis of Satellite Launch.     

土壤水分特征曲线几种标定的对比分析

土壤水分特征曲线的标定是土壤水的能量和数量关系在不同尺度间转换的方法,也是植被生态需水机理研究的重要组成部分。通过对传统的迭代法、一次法和建立的2种混合遗传算法进行比较,标定的结果表明,多项式遗传算法能综合考虑2种方差的影响,并能给出一致性的标定结果,这种经验式混合遗传算法对传统的标定算法进行了改进。     

基于物联网的日光温室环境监测在线校准

[目的]温室环境条件特别是温度对于作物生长和发育具有十分显著的影响。日光温室调控的主要环境因子之一是温度。然而,自然环境下的光照对温度产生作用,影响空气温度的监测精度。大多温度传感要求将传感器置于避光处,然而实际应用中难以保证。[方法]采用机器学习中的支持向量机算法(SVM),对日光温室内的温度智能监测算法进行了研究,根据光照情况对实时监测的温度数据进行校准。[结果]通过与实验测量的数据进行对比分析,结果表明,所提出的监测方法可以较为准确地实时监测空气温度,从而无需使用隔热材料或者遮阳处理,就可以基于监测的数据更精确地对相应的环境因素进行调节。[结论]基于该方法,可采用常用的工业设备实现温室大棚内实时温度数据的监测,既可以节约设备和人力成本,又可以为温室控制提供准确的数据。     

颗粒肥料离散元仿真摩擦因数标定方法研究

对分体圆筒法、倾斜法、抽板法和斜面法4种颗粒特性测试方法进行Plackett-Burman多因素显著性筛选试验,试验方差分析结果表明,不同的测试方法影响测量结果的显著因素与因素显著程度。根据分体圆筒法、倾斜法和斜面法的方差分析结果,提出一种基于颗粒物料整体特性的摩擦因数标定方法,将仿真试验与真实试验相结合,依次标定出尿素颗粒与PVC材料间静摩擦因数为0. 41,颗粒间静摩擦因数为0. 36,颗粒间滚动摩擦因数为0. 15。将所标定的摩擦因数采用无底圆筒法进行验证试验,休止角仿真试验结果为30. 57°,真实试验结果为31. 74°,相对误差为3. 69%,不同含水率下的实际试验休止角与所标定摩擦因数下的仿真休止角相对误差均不大于4. 59%,仿真试验结果与真实试验结果无显著差异,验证了所标定摩擦因数的有效性。本方法可用于其他颗粒状物料间摩擦因数的标定试验。     

双目立体视觉在果树采摘机器人中的应用

果树采摘机器人是未来智能农业机械化的发展方向,具有广阔的应用前景。为此,针对果树采摘机器人作业环境的复杂性与定位目标的特殊性,从双目定位的几何模型出发,归纳了双目定位的基本过程,总结了立体视觉的摄像机标定问题,提出变焦距双目视觉定位果实的研究方向,并对双目图像的匹配问题进行了分析。     

基于残差BP神经网络的6自由度机器人视觉标定

基于视觉伺服控制的机器人手眼标定和逆运动学求解一直是视觉伺服领域的核心问题。随着应用场景的逐渐复杂,传统手眼标定方法已无法满足需求;随着网络深度的增加,单一BP神经网络逆运动学求解算法的误差趋于饱和甚至变大,无法进一步提升网络性能。针对以上问题,本文将手眼标定和逆运动学求解融合为拟合目标图像坐标到机器人各关节角之间的映射关系问题,提出了一种残差BP神经网络算法。使用多个残差网络模块的方式加深BP神经网络的深度,残差模块的输入信息可以在网络内跨层传输,较好地解决了因深度增加网络模型容易产生梯度消失而无法提升网络性能的问题;通过6自由度机器人雅可比方程对逆运动学解的空间进行划分,确定了8个独立的区域,基于独立区域方法对训练数据进行处理,从而避免了多自由度机器人逆运动学多解对网络学习的影响,网络训练精度提升了2个数量级,训练速度提高了2倍。在REBot-V-6R型6自由度机器人输送线分拣系统中进行二维平面抓取和三维实物抓取实验,实验结果验证了该方法的准确性。结果表明,该方法比1层BP神经网络、3层BP神经网络、5层BP神经网络的训练精度分别提高了4个数量级、2个数量级、5个数量级,测试精度提高2个数量级;与传统标定方法相比,本文方法节约了逆运动学求解过程的计算成本,抓取位姿精度提高了1个数量级。