本文提出了一种人工神经网络的田间信息插值新方法，并利用ArcView3.2 软件绘制碱解氮的BP 神经网络插值空间分布图和球状插值分布图，并对BP 神经网络插值方法和克立格球状插值方法的结果进行了误差分析。结果表明，BP 神经网络的插值方法优于克立格球状插值，该方法有利于田间信息空间分布特性准确、直观的表达，有利于农田精确施肥、灌溉、播种等精细农业生产管理。

该文针对传统土壤电导率测量仪器准确性低、实时性差等问题，根据土壤电导率的测量原理，采用NI 公司的虚拟仪器开发平台LabVIEW 和数据采集卡，采集土壤温度和电导率等参数，设计了一种土壤电导率测量仪器，通过简单的操作便可测量不同温度下的土壤电导率，并实现数据的显示、保存、打印和网络传输等功能。该仪器利用温度校正方法消除温度变化给电导率带来的测量误差，对KCl 标准溶液电导率的测试结果表明，该仪器具有较高的测量精度，可以广泛应用于土壤电导率的测量中。该研究为多参数的正确测试提供了一种简便的方法。

为了解决精细农业中空间信息的快速获取与分析处理问题，提出并实现了基于掌上电脑的农田信息采集处理系统，并对具体实现该系统的若干关键技术如GIS、数据模型、空间索引、Windows CE 的数据库技术、采样控制等进行了分析与研究。作为一种实时系统，该系统可结合GPS 和各种传感器，快速地测量和分析农田信息。本文用该系统对两种GPS（AG132 与MAP330）进行了精度比较试验，快速而方便地证实了AG132 有更好的精确性与稳定性，更适于农业生产要求。试验的过程与结果也间接表明了该系统具有较好的实用性，为精细农业信息的快速采集和处理的提供了关键技术和设备。

根据拖拉机的特点和性能检测的要求，开发了一套基于多功能数据采集卡（DAQ）和图形化编程语言LabVIEW 的虚拟仪器拖拉机性能综合检测仪器，该仪器可以快速测定拖拉机制动性能、车速、轴重、牵引力、灯光、烟度、噪声，实现了一机多用。通过实际测试表明，仪器具有较高的可靠性和精度。

This paper reports a practical approach for detecting and diagnose engine faults in real-time based on both the historical and the real-time engine operation data using a specially design neural networks-based fault diagnosis expert system. This system consisted of multiple sensors for real-time monitoring, an engine database for historic data comparison, and a neural network-bases classifier for detecting faults based on both the realtime and the historic data. This neural network-based engine fault diagnosis system was evaluated in a series of validation tests. The results indicated that the system was capable to detect the predefined faults reliably.

在神经网络的应用过程中，经常遇到样本太多问题，采用模糊聚类分析，科学选取学习样本，使少量的学习样本本身包含全部样本的特性，很好地解决了由于样本多学习速度慢的问题，并提出了训练结果检验方法。此方法可以作为神经网络应用的一个拓展。

针对现有农机化区划评价方法与社会经济发展的不协调性，提出并建立了一种基于Map Objects开发技术的GIS 农机化区划系统。该系统通过创建地理信息数据库、农机信息数据库，并用数据统计处理以及计算区划值来进行农机化区划和分类。应用表明，该农机化区划系统可以方便、科学地进行区划和分类，并能进行各种因素对区划结果的灵敏度分析，为农机化区划提供了一种新的方法。

NIR spectroscopy appears as a rapid, convenient and simple nondestructive technique that can be useful to quantify several soil properties. This method was used to estimate soil N and OM content in a Dark soil of Zhejiang Province, Hangzhou county. A total of 125 samples were taken from the field. A total of ninetyeight samples spectra were used during the calibration and the validation stage to validate the calibration equation produced. Twenty-seven sample spectra were used to determine N and OM concentration. NIR spectra of these samples were correlated using partial least square regression. Correlation coefficients higher than 0.83 (r2>0.83) where obtained during the calibration stage and validation stage validation analysis. Statistic for predicted N and OM is shown in Table 2, Fig 3. The r2 between measured and predicted values in both constituents N and OM, 0.82 and 0.84, as well as RMSEP, 3.67 and 0.06, SEP, 0.07 and 0.06, respectively, demonstrated that NIRS method have potential to predict accurately this constituents in soil.

In order to acquire synchronous multi-information in precision agriculture, a mobile soil moisture instrument with GPS position function was developed. This device, which mainly consists of a single-chip microcomputer (SCM), a keyboard, a Liquid Crystal Display (LCD), several sensors and a GPS receiver, was designed to synchronously integrate the signals from both sensors and GPS receiver. In this paper, results showed it worked very well with a Standing Wave Ratio (SWR-2) sensor and GPS25 receiver in respectively obtaining high-precise multiple soil moisture information of vegetable terra, cropland and gravel terra. In addition, the following plane and section tests using the instrument and its corresponding interpolation analysis with software Arcview also afforded a feasible and effective method for further soil moisture researches.

ency of mixed oil composed of cottonseed oil and conventional diesel oil and for improving the performance of engine fuelled by the mixture. The experimental results obtained showed that a mixing ratio of 30% cottonseed oil and 70% diesel oil was practically optimal in ensuring relatively high thermal efficiency of engine as well as homogeneity and stability of the oil mixture. A quadratic regressive orthogonal design test method was adopted in the experiment designed to examine the relationship between specific fuel consumption and four adjustable working parameters (intake-valve-closing angle (α), exhaust-valve-opening angle (β), fuel-delivery angle (θ) and injection pressure (P, in 104 Pa)) when the above-mentioned oil mixture was used. The mathematical equations characterizing the relationship were formulated. The equation of specific fuel consumption derived from the regressive test under each operating condition was set as the objective function and the ranges for the four adjustable working parameters were the given constraint condition. Models of non-linear programming were then constructed. Computer aided optimization of the working parameters for 30:70 cottonseed oil/diesel oil mixed fuel was achieved. It was concluded that the predominant factor affecting the specific fuel consumption was fuel-delivery angleθ, the approximate optimal value of which, in this specific case, was 3-5degrees in advance of that for engine fuelled by pure diesel oil. The experimental results also provided useful reference material for selection of the most preferable combination of working parameters.