【作者】 周利明；

【导师】 王书茂； 张小超；

【作者基本信息】 中国农业大学 ， 机械电子工程， 2014， 博士

【摘要】 随着我国棉花种植规模发展与机械化采收程度的提高,对棉花机械播施与收获过程的智能检测技术也提出了更高的要求。准确获取区域内的棉花产量分布信息一方面能够检验当年精准农业措施的实施效果,同时也是来年播种施肥精准变量作业处方决策的重要参考；棉花播种量检测则是实施高精度变量播种作业的关键,也是判断播种作业质量水平的基础。因此研究一种能在线检测棉田作业过程中棉花产量与播种量的方法和技术,对于减少棉花生产资料投入、增加棉花产量以及减少环境污染等都有极其重要的意义。论文提出了一种基于电容法的采棉机棉花产量和棉花精密播种机播种量的检测方法,并在深入研究基础上实现了工程应用。论文首先研究了籽棉的介电特性,并结合籽棉的气流输送特点,提出了基于双电容相关法来检测籽棉的流速,基于单电容法来检测籽棉的质量流量。双电容相关法是利用两个传感器信号的时间相关性,结合传感器间距,实现籽棉运动速度的检测；单电容法是基于采棉机作业过程中籽棉运动速度基本稳定的条件下,利用电容传感器信号来检测籽棉的质量流量。论文在利用ANSYS有限元分析并结合试验验证基础上,设计了差分型电容传感器,研制了检测电路和单片机信息采集和处理系统,实现了籽棉流速和质量流量的检测。论文以CASE CPX620采棉机为试验平台,构建了基于CAN总线的棉花产量监测系统。监测系统主要由人机交互终端、籽棉质量流量传感器、微波谐振式含水率传感器、GPS等构成,通过关键技术研究和系统集成,实现了基于经纬度坐标的棉花产量在线测量,以及采棉机作业工况参数等相关数据的实时获取。基于Labwindows/CVI测控软件开发平台,完成了上位机的数据采集与处理,为多信息融合处理奠定了基础。为检验棉花产量监测系统的稳定性和环境适应性,搭建了籽棉气流输送试验台,通过温度变化试验,验证了研制的差分型电容传感器对温度有比较好的适应性；通过改变籽棉品种试验,验证了研制的检测系统对多品种籽棉的适应性。文中还研究了棉花质量流量、含水率与电容信号响应的关系,建立了基于相对电容变化率与籽棉质量流量、含水率的回归模型。实验室静态试验表明,测产模型的预测平均误差不大于4.7%,动态试验表明,测产模型预测误差不大于9.07%。田间收获试验表明,测量装置的累积电容变化量与籽棉重量的拟合决定系数R2为0.94。另外,论文还研制了螺旋型电容籽粒传感器,准确检测棉花播种机播种管内棉籽的通过信息,实现了棉花精密播种机播种量的检测。研制的螺旋型电容籽粒传感器,可将棉籽通过播种管电容检测区域的信息转换成脉冲信号,以此来获取棉花精密播种机的播种量、漏播及重播信息。通过最小二乘移动平滑方法实现籽粒脉冲的平滑处理,得到最佳的信噪比,以一阶导数寻峰结合阈值判别条件快速识别棉籽脉冲,并获得峰位与脉冲边界信息,结合脉冲积分面积判断双籽粒同时下落,提高了播种量检测精度。同时构建了基于CAN总线的精密播种机播种监测系统,实现了播种质量水平的高可靠性监测。实验室试验表明,研制的系统能够实现漏播、阻塞报警,其对播种量监测精度为94.6%,漏播量监测精度为93.5%,重播量监测精度为88.1%。更多还原

【Abstract】 With the development of mechanical planting and harvesting for cotton, the demand for precise and intelligent detection system for seed cotton yield and seeding rate are rising. The accurate cotton yield information can examine the implementation effect of precision agriculture in the current year and provide the important reference information for variable rate prescription map in the coming year. The monitoring of seeding rate is the key to successfully implement the variable seeding, but also the basis to determine seeding performance. Therefore, studying on the online detection method for seeding rate and cotton yield has very important significance to reduce cotton production inputs, increase cotton production and reduce environmental pollution.In this research, a new method based on capacitive sensing for detecting seed cotton yield and seeding rate was proposed. Some thorough research work on the method were done to assess it.Then, the project application was realized.The dielectric property of seed cotton was studied firstly. Meanwhile, according to the pneumatic conveying characteristics of seed cotton, the detection method for seed cotton conveying velocity based on dual capacitor cross correlation technique and method for seed cotton mass flow based on single capacitive sensor were proposed.The correlation method was to use the time correlation between the two capacitive sensor signals,combined with the sensor spacing,for detecting seed cotton conveying velocity.The other method was to use capacitive sensor signal to detect the mass flow based on the conveying velocity under conditions of stable in the harvest process.The differential shaped capacitive senson was densined. The finite element method was used to analyze the sensor by using ANSYS and the experiments were conduct to verify the simulation results.Based on the design of capacitive detection circuit and microcontroller information acquisition and processing system, the conveying velocity and mass flow of seed cotton was obtained.The cotton yield monitor system base on CAN bus was developed in CASE CPX620cotton harvester, which includes interactive computer terminal, seed cotton mass flow sensors, microwave moisture sensor, GPS receive.Through research of key technologies and system integration,the cotton yield based on latitude and longitude information,as well as cotton harvester working condition parameters were acquired online. With the help of the Labwindows/CVI software development platform, upper computer data acquisition and processing was completed and the foundation for the multi-information fusion was laid.To assess the cotton yield monitor system, the free-falling experimental device was build.According to the results of the temperature and variety change test,the sensor have a better adaptability to temperature and variety. One kind of seed cotton named "xinluzao43"is used as sample to study the influence of cotton moisture content on the output capacitance value of mass flow sensors, and the relationship of cotton mass flow and sensor output capacitance. The relationship among cotton moisture content, mass flow rate and capacitance value can be described by a bivariate regression model. The laboratory experiment results show that, the predictive error of static model is4.7%and error of dynamic model is9.07%. The field experiment results indict that the determination coefficients of the model for cotton mass and capacitance variation is0.94.In addition, the spiral shaped seed sensor was designed in the paper.The sensor could capture the dynamic information when seed went through it, then the seeding rate for drill was detected. Meanwhile, the dynamic information was converted into digital pulse signal in order to get the information of seeding rate and missing rate and multiples rate.The seed pulse was smoothed by Savitzky-Golay smoothing method to get the optimum signal noise ratio. Combined with the first derivative and threshold condition, the cotton seed pulse was identified rapidly.The pulse position and boundry information were also obtained.According to the pulse integral area, the double seed could be detected, which could help to improve the seeding rate monitor accuracy.At the same time, the monitoring system based on CAN bus was constructed to realize high reliability monitoring of sowing quality levels. The laboratory experiment results showed that the system can realiz the missing and blocking alarm, and the system accuracy for single seed detection is up to97.3%, and the system has an accuracy of94.6%in seeding quantity,93.5%in missing detection and88.1%in multiples rate.更多还原