分布式移动农业病虫害图像采集与诊断系统设计与试验

发布时间：2018-10-29 10:03

【摘要】：为了便捷地采集和实时诊断农业病虫害图像,设计了一个分布式移动农业病虫害图像采集与诊断系统。该系统由多个便携式图像采集终端和一个图像处理服务器组成;其中,图像采集终端包括嵌入式相机、可伸缩的手持杆和装载控制App的手机;图像处理服务器包括农业病虫害诊断、信息记录和反馈模块等。手持杆可将安装在其前端的嵌入式相机送到人手或视觉难以企及的病虫害区域,手机可实时预览前端相机的拍摄画面和实现控制相机完成农业病虫害图像采集等功能;系统通过HTTP协议实现多个采集终端与图像处理服务器的数据交互,协同进行分布式计算,可以减少网络移动资费和服务器的负载。利用该系统对水稻纹枯病图像采集与诊断测试结果表明,该系统的图像采集终端可以便捷地采集到水稻纹枯病图像,手机端视频预览画面延时低,对相机控制命令无误,图像采集终端与服务器通信稳定,服务器端对水稻纹枯病图像处理和诊断实时,基于图像的水稻纹枯病为害等级诊断准确率为83.5%。如果服务器端加载不同的农业病虫害图像处理和诊断算法,该系统可广泛应用于各种农业病虫害图像的采集与诊断。
[Abstract]:In order to collect and diagnose agricultural pests and diseases images conveniently, a distributed mobile agricultural disease and pest image acquisition and diagnosis system was designed. The system is composed of a plurality of portable image acquisition terminals and an image processing server, wherein, the image acquisition terminal comprises an embedded camera, a scalable handheld lever and a mobile phone loaded with App. Image processing server includes agricultural pest diagnosis, information recording and feedback module. The handheld rod can send the embedded camera installed in the front end to the area of diseases and insect pests that are difficult to reach by hand or vision. The mobile phone can preview the picture of the front-end camera in real time and realize the function of controlling the camera to complete the image collection of agricultural diseases and insect pests. The system realizes the data interaction between multiple collection terminals and the image processing server through HTTP protocol, and carries out distributed computation in cooperation, which can reduce the network mobile charges and the load of the server. The results of image acquisition and diagnosis test of rice sheath blight show that the system can collect rice sheath blight image conveniently, the video preview screen of mobile phone has low delay, and the command of camera control is correct. The communication between the image acquisition terminal and the server is stable. The server can process and diagnose the rice sheath blight in real time. The diagnostic accuracy of rice sheath blight based on the image is 83.5%. If different image processing and diagnosis algorithms of agricultural diseases and pests are loaded on the server side, the system can be widely used in the collection and diagnosis of various agricultural pests and diseases images.
【作者单位】：浙江理工大学信息学院;中国水稻研究所;
【基金】：国家高技术研究发展计划(863计划)项目(2013AA102402) 浙江理工大学521人才培养计划
【分类号】：S43;S126

【相似文献】