棉蚜发生量信息快速获取方法与监测模型的建立研究

发布时间：2018-01-01 15:24

本文关键词：棉蚜发生量信息快速获取方法与监测模型的建立研究　出处：《石河子大学》2017年博士论文　论文类型：学位论文

【摘要】：棉蚜发生量信息的快速、精准获取是其科学精确防治的必要前提,而目前的棉蚜发生量信息采集依人工调查为主,受棉蚜分布不均匀和大量迁飞的影响,即使耗费大量的人力和时间,其数据调查精度和时效性仍然较差,致使棉蚜防治受到很大影响。因此,本研究针对目前棉蚜信息监测依靠田间人工调查的方法,而导致棉蚜信息获取滞后、采集图像存在背景复杂和粘连等一系列问题。以棉蚜极强的趋黄特性为理论依据,利用黄色粘虫板在棉田进行大田试验,通过对比分析不同高度和方向的黄色粘虫板诱虫效果,同时在人工调查对应小区不同类型棉蚜信息的基础上,确定棉蚜信息最佳监测条件。进而结合网络高清拍照、无线远程传输、机械自动化控制等现代信息技术,自主研发棉蚜发生量信息快速采集装置,实现棉蚜图像信息的自动采集,解决了采集图像的背景复杂与蚜虫粘连问题;进一步结合图像识别计数技术构建棉蚜发生量信息快速获取方法,在此基础上建立棉蚜信息快速监测预警模型,并开发了棉蚜信息快速监测预警与决策一体化系统,为农业一线生产者提供所在地区棉蚜发生量信息查询服务,同时为棉农大田实际生产提供棉蚜防治决策信息。主要研究结果如下:1.基于趋黄性棉蚜最佳快速监测条件的确定本研究基于棉蚜趋黄性进行大田试验,通过设置不同方向和高度的诱蚜黄板进行诱蚜效果的统计分析,以此确定棉蚜信息最佳监测条件。结果表明:棉田有翅蚜发生量和总蚜量之间的相关性极显著,通过监测有翅蚜的发生量可以很好的估测棉田总蚜量的发展趋势;同时,不同高度和方向的黄板诱蚜量之间均存在极显著的差异,并且通过两因素方差分析,表明黄板高度和方向作为单一因子和交互作用均对诱蚜量的影响达到了极显著水平,因此,黄板下边沿距离棉花冠层的高度和黄板的方向对棉蚜信息的监测具有较大的影响。经过2013年和2014年两年试验,结果证明,诱蚜黄板距离棉花冠层90cm高度且水平方向与60cm高度且向东方向的监测黄板(即距离棉花冠层60cm-90cm),这两个条件可以作为棉蚜发生量快速监测的最佳条件。2.棉蚜发生量快速获取方法的研究在棉蚜最佳监测条件确定的基础上,利用网络高清拍照、无线远程传输、机械自动化控制等现代信息技术,自主研发了棉蚜发生量信息快速采集装置,实现了采集图像的背景单一化,解决了蚜虫堆积导致棉蚜识别计数精度不高的问题。通过对装置进行测试试验,结果表明,该装置获取图片信息的有效性为95.83%、传输成功率达到99.4%、每次传输平均耗时为191s,其图片信息的平均传输时间不影响棉蚜快速监测预警的时效性。在此基础上,基于图像识别计数技术,以实用性、可扩充性、统一性和简单性为设计原则,设计开发了棉蚜识别计数软件,该软件系统实现了棉蚜预警参数信息的自动采集和统计分析等功能,经过人工计数对比测试分析,该系统识别计数的相对误差为2.67%,完全满足农业生产中棉蚜预警决策的需求。3.棉蚜发生量实时估测模型的建立与验证基于棉蚜发生量快速获取装置,于2015年进行了大田测试试验,同时人工调查了棉田棉蚜实际发生量,通过分析装置获取信息和大田实际发生量进行相关关系,并在此基础上建立了不同类型百株蚜发生量快速监测模型。经过2016年的模型验证试验表明:百株有翅蚜预测精度偏低,百株无翅蚜和百株全蚜量的预测精度较高。大田棉蚜爆发量的判断一般依据全蚜量的发生量,并且百株全蚜量的预测值和真实值的均方根误差(RMSE)仅为2180,因此,模型Y=0.0362x2-12.642x+16470对大田实际百株蚜发生量进行预测具有较好的可行性。与此同时,经过大田实际调查对棉蚜危害程度进行了分级,并与基于棉蚜快速监测装置的百株蚜估测分级进行了对比分析,结果表明其预测等级完全一致,正确率达到100%。因此,本研究建立的棉蚜发生量预警模型完全可以满足棉蚜发生程度的预警预测,再参考结合当地气象资料,能为大田棉蚜防治提供可靠的数据支持。4.棉蚜信息预警系统的研发本研究集棉蚜信息动态采集与图像识别计数技术,结合前文建立的棉蚜快速监测预警模型,集成计算机编程和数据库技术,开发出棉蚜信息快速监测预警与决策一体化系统,系统主要包括信息展示、系统管理、信息管理、棉蚜预警决策,信息查询与效益评价五个功能模块。实现了通过对棉田棉蚜发生量的估测计数,参考棉蚜发生程度和气象资料提出预警信息,以此提出对应的防治措施,为棉农提供准确的棉蚜防治时期和农药施用量,有效抑制棉花虫害的大面积发生,提高了对棉蚜测报的准确率和时效性,同时节省了农药施用量,达到节本增效的目的。
[Abstract]:The amount of cotton aphid occurrence information fast, accurate access is a necessary precondition for the scientific and accurate control, and the amount of information collected by the artificial cotton aphid occurrence survey, by the impact of uneven distribution of cotton aphid and a large number of migration, even spend a lot of time and manpower, the investigation is still poor data accuracy and timeliness, resulting in the cotton aphid prevention by great influence. Therefore, this study focuses on the information monitoring on field investigation of artificial cotton, cotton aphid caused information lag, a series of problems with complicated background image. Adhesion to the cotton aphid strong trend yellow characteristics as the theoretical basis, a field experiment was conducted in the cotton field by yellow sticky board, yellow sticky board through the comparative analysis of different height and direction of the trapping effect, at the same time in the corresponding cell based artificial investigation of different types of cotton aphid information, determine the best monitoring information of cotton aphid Measuring conditions. And then combined with the network high-definition camera, wireless remote transmission, automatic control and other modern information technology, independent research and development of rapid information acquisition device of cotton aphid, Aphis gossypii to realize the automatic acquisition of image information, to solve the complex background image acquisition and aphid adhesion problems; combining with the construction of cotton aphid occurrence fast acquisition method of image information recognition count technology, on the basis of the establishment of rapid detection and early warning model information and the development of the cotton aphid Aphis gossypii, information rapid monitoring and early warning and decision system for agricultural producers, a region is provided for aphid information query service, at the same time as farmers field production cotton aphid prevention decision information. The main results are as follows: 1. to determine the trend of yellow the best conditions for the rapid monitoring of Aphis gossypii based on field experiment was carried out based on the trend of yellow, by setting the Statistical analysis of aphid trap effect in the same direction and height of the Yellow aphid, Aphis gossypii in order to determine the optimum conditions of information monitoring. The results show that the correlation between the occurrence of cotton aphids and aphid significantly, by monitoring the amount of aphids can estimate the total cotton aphid the good development trend; at the same time, different height and the direction of the Yellow induced significant differences were found between the amount of aphids, and through two factor variance analysis showed that the yellow, height and direction as the single factor and the interaction of impact induced aphid reached a very significant level, therefore, has a great influence on the direction of the edge information monitoring cotton cotton canopy height distance the yellow and yellow. After 2013 and 2014 two test results show that monitoring of aphid with yellow cotton canopy height and horizontal distance of 90cm direction and 60cm direction and height to the East Yellow (distance of cotton canopy, the best conditions for.2. 60cm-90cm) these two conditions can be used as a cotton aphid Aphis gossypii quantity rapid monitoring the occurrence of the amount of fast acquisition method based on determining the optimal monitoring conditions of the cotton aphid, the use of network high-definition camera, wireless remote transmission, automatic control and other modern information technology, independent research and development of cotton aphid fast acquisition device information, the realization of the background of a single image, resulting in the accumulation of cotton aphid to solve the aphid recognition and counting accuracy problem. Through the test, the results show that the effectiveness of the device, the device obtains image information is 95.83%, the transmission success rate of 99.4%, average time for the transmission of each 191s, the average the transmission time the picture information does not affect the timeliness of Aphis rapid monitoring and warning. On this basis, image recognition and counting technology based on practicality, can Extensibility, unity and simplicity of design principles, design and development of the cotton aphid counting software, the software system realizes the automatic data collection and statistical analysis functions of cotton aphid warning parameter information, through the manual counting contrast test and analysis, the system recognition and counting of relative error is 2.67%, fully meet the rapid establishment and verification of real time estimation the amount of acquisition device model based on the amount of cotton aphid Aphis gossypii warning decision in agricultural production needs of.3. cotton aphid were conducted in 2015, and investigated the test, artificial cotton aphid and the actual amount, through the analysis of means of access to information and the actual amount of field correlation is established based on the amount of rapid monitoring model different types of 100 plant aphid. Through the model test in 2016 showed that aphids per plant low prediction accuracy, 100 and 100 strains of apterae aphid strains The prediction accuracy is high. The amount of cotton aphid outbreak field based on the general judgment amount of full amount of aphids, and 100 strains of aphid prediction value and the real value of the root mean square error (RMSE) is only 2180, so the Y=0.0362x2-12.642x+16470 model in the field of actual amount per plant aphid occurrence forecast is feasible. At the same time after field survey of cotton aphid, harm degree is classified, and compared with the rapid monitoring device of 100 plant aphid Aphis gossypii estimation based on the classification, the results show that the prediction level is completely consistent, the correct rate of 100%. because of this, this study established a cotton aphides early-warning model can fully meet the prediction of the occurrence degree of cotton aphid then, the reference combined with local meteorological data, research can provide reliable data to support.4. information early warning system for the field of cotton aphid Aphis gossypii control in the study of cotton aphid information dynamic recovery Set and the image recognition and counting technology, combined with the rapid detection and early warning model previously established cotton aphid, integrated computer programming and database technology, the development of cotton aphid information fast monitoring and early warning and decision system, including system information display, system management, information management, Aphis warning decision five function module evaluation information query and achieve benefits. Based on the cotton aphid occurrence quantity estimation of reference counting, the occurrence degree of Aphis gossypii Glover and meteorological data provides early warning information, put forward the corresponding prevention and control measures, provide the period and amount of application of pesticide control of Aphis accurately for farmers, a large area of effective suppression of cotton pests and improve the forecasting accuracy of cotton aphid and timeliness, at the same time save the amount of pesticide application, to reduce cost and increase benefit.

【学位授予单位】：石河子大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：S435.622.1;S126

【相似文献】