【作者】 崔怀雷；

【导师】 邵陆寿；

【作者基本信息】 安徽农业大学 ， 农业机械化工程， 2004， 硕士

【摘要】 农作物的病虫草害的防治是农业生产中最普遍的一项农事活动。有关资料表明，如果农业生产中不使用农药，农作物收成将减少30％，农副产品价格将至少上涨50％～70％。总之发展农业就离不开农药的使用。但是在农药使用上存在着很多问题，如施药过程中忽视了田间病虫草害发生的差异性和作物信息的差异性，从而使大量的农药流失到非靶标环境中，造成严重的环境污染。要彻底地克服这些传统方法的弊端，唯有实行精准施药技术，然而发达国家发展起来的精准施药技术是建立在“3S”等一系列高科技基础上的。而我国现阶段国情使得精准施药技术难以推广和应用。所以我国精准施药技术不应该是“3S”技术的简单照抄，应该是“变量施药技术思维”方式指导下的农药施药设备与过程的革新，改造。基于这个思想，本文提出了一种适合我国现状的新型模糊施药控制系统。本系统主要包括自动物体识别子系统和施药量模糊控制子系统。 自动物体识别子系统中提出用归一颜色分量法处理了光照强度和阴影对图像的影响，以及采用直方图均衡化来解决由此而产生的灰度值过度集中的问题：列举了三种图像增强方法，通过实验我们选用了自适应邻域平均法；通过对比分析，选用最小误差法来自适应地实现图像的二值化；运用形态滤波消除了二值图像中的一些小空洞和孤立点；最后通过比较决定从二值化图像提取相对面积比这个特征信息。实验结果表面这个特征可以准确地反映作物密度的变化。 本文根据专家知识和实践经验设计了模糊控制器。这其中包括，输入输出变量的论域、模糊子集以及各子集隶属度函数的确定；模糊控制规则的设计；模糊推理系统的设计等几个方面的内容。然后利用Simulink对所设计的模糊模型进行了仿真，并分析了仿真结果，结果说明所设计的模糊控制器基本符合实践经验 此外本文在编程方面采用的是VC++与MATLAB混合编程的方法。主程序采用VC++编写，而一些具体的算法采用MATLAB编写。结合两种软件的优点，从而保证准确性，提高了编程效率。 综上所述，本文详细地阐述了这种新型精准施药控制系统的设计，并对设计出的系统讲行了大量试验，并通过仿真的方法评价了系统的准确性。更多还原

【Abstract】 The controlling of agricultural Crops Pests is the most common agricultural activity .The relative data shows that if we don’t apply pesticide ,the harvest of agricultural products will decrease by 30%,the price will increase by 50%~70% at least. In a word, to develop agriculture couldn’t lack the applying of pesticides. But there are many problems in pesticide applications. For examples, The difference of crop pest and crop information is ignored during the process of pesticide application, It makes large quantities of pesticide is lost into non-target environment, and caused a series environment problem. In order to thoroughly overcome the corrupt practice of those traditional ways, what we only can do is to apply precision applying pesticide. But the technique of precision applying pesticide developed in the developed country is based on "3S". At present our country situation make it difficult to popularize and apply the technique. So we shouldn’t follow "3S"blindly,On the contrary we should follow "Variable applying pesticide" to reform the equipment and the process. Based on the thought ,the paper advance a new precision applying pesticide controller system. The system is composed of automatic recognition subsystem and pesticide rate fuzzy control subsystem.In the automatic recognition subsystem, it applies the normalization color component method to rectify the impact on image caused by light variety and shade; apply the technology of histogram equalization to resolve gray level excessive concentration; enumerate three image enhancement methods and select the collative adjacent region average method; through experiment we select the minimum error method to approach image binaryzation. By contrast analysis apply the composite morphological filtering to eliminate the little hole-filling and isolated point in the binary image, Extract the relative areas ratio as signature information after analysis . The experiment results show that the signature can reflect the variety of crops density precisely.The paper design a fuzzy controller based on experts knowledge and practice experiment. Which includes the choice of input-output variable domain fuzzy sets and membership function ,the design of fuzzy controller rules and fuzzy logical .after it was designed we used the SIMULINK to simulate the designed controller. The simulation result indicate that it fits the practice experiment.Furthermore, the paper apply the combination program method of VC++ and MATLAB. The main program is written with VC++. The application Algorithms isaccuracy is ensured and the programming efficiency is greatly improved.Sum up, The paper expatiate the design of the new system in detail. Many examinations have been done on it. The simulation result shows it is useful.更多还原