基于云服务架构的田间信息采集与分析系统设计  

Design of Crop Information Storage Analysis System Based on Cloud Service Architecture

作　　者：马旭颖 张智勇[2] 高德华 李民赞[1,2] 孙红[1] 李松[2] MA Xuying;ZHANG Zhiyong;GAO Dehua;LI Minzan;SUN Hong;LI Song(Key Laboratory of Modern Precision Agriculture System Integration Research,Ministry of Education,China Agricultural University,Beijing 100083,China;Key Laboratory of Agricultural Information Acquisition Technology,Ministry of Agriculture and Rural Affairs,China Agricultural University,Beijing 100083,China)

机构地区：[1]中国农业大学现代精细农业系统集成研究教育部重点实验室,北京100083 [2]中国农业大学农业农村部农业信息获取技术重点实验室,北京100083

出　　处：《农业机械学报》2019年第B07期122-127,共6页Transactions of the Chinese Society for Agricultural Machinery

基　　金：广西科技重大专项经费项目(桂科AA18118037);国家自然科学基金项目(31501219);中国农业大学研究生实践教学基地建设项目(ZYXW037)和中国农业大学研究生课程建设项目(HJ2019029、YW2019018)

摘　　要：为了满足农田作物信息采集和分析服务的需求,将智能手机终端硬件、微信小程序软件与云服务平台相结合,设计了一款基于云服务架构的田间信息采集与分析系统。系统主要包括腾讯云服务器模块和手机微信小程序模块,其中,云服务器端使用MySQL搭建数据库,用于存储、处理和下载数据;使用CSS和JavaScript语言及小程序封装的组件开发微信小程序,用于交互实现数据的采集、上传与信息反馈。以田间小麦作物生物量指征参数调查为例,针对冠层覆盖度和植株行间距计算进行了系统应用测试。采集100幅出苗期的小麦冠层图像,由小程序端上传样本图像到后台处理。使用霍夫变换、图像掩膜和图像腐蚀获取定位图像后,利用HSV彩色空间突出样本像素点,计算冠层覆盖度;采用投影法和滤波法提取峰值,获取株行中心线,从而计算植株行间距。建立了图像识别像素株行间距与实测株间行距间的一元线性回归模型,建模精度R2达到0.911,可为田间作物信息检测和调查提供技术支持。In order to meet the demands of collecting and analyzing services of farmland crop information,a crop information collection and analysis tool was designed based on cloud service architecture with the combination of the smart phone terminal hardware,WeChat applet software and cloud service platform.The system mainly included Tencent-cloud server module and mobile phone WeChat applet module.The MySQL was used to build a database for data storing,processing and download on cloud server.The WeChat applet was developed by CSS,JavaScript and applet packaged components.It was used to realize the interaction between collecting and uploading data and information feedback.In order to apply and test the system,the survey of biomass indication parameters of wheat in the field was taken as an example.The researches were carried out for the calculation of canopy coverage and plant row-spacing.More than 100 sampling images of wheat were captured at seedling period,and they were uploaded from applet to background and processing.After the image preprocessing following the target area detection by using Hough transform,image mask segmentation and image enhancement with erosion processing,the canopy of wheat was segmented and the canopy coverage was calculated by using HSV color space to highlight the pixel of sampling plants.The algorithms were proposed to extract the peak line of plant by projection and filtering method.Then,it was used to calculate the plant line spacing in the row.The linear regression model was established to indicate the fitting accuracy between the line spacing of the image recognition pixels and the measured values.The result showed that the modeling accuracy R 2 reached 0.911.It could provide a technical support for crop information detection and investigation in the field.

关 键 词：作物信息 云服务 服务管理系统 微信小程序 图像处理 

分 类 号：S126[农业科学—农业基础科学]