基于物联网的桔小实蝇诱捕监测装备设计及试验

    Design and test of remote monitoring equipment for bactrocera dorsalis trapping based on internet of things

    • 摘要: 为了实现对桔小实蝇诱捕的实时监测和快速诊断,设计了一个基于物联网的桔小实蝇诱捕监测装备。该装备包括诱捕监测装置、太阳能供电装置和监测控制装置3个主要部分,其中诱捕装置包括顶盖、透明的连通件和诱捕瓶;太阳能装置包括太阳能板、蓄电池以及太阳能板支架;控制装置包括Fit-pc控制器、3G通讯模块和自主研发的桔小实蝇监测计数系统软件。该装备结合了机器视觉技术、远程通讯技术以及太阳能供电等技术,实现了集病虫害信息采集、处理、传输与自供电为一体的桔小实蝇诱捕监测装备,可长期的、实时的、远程的监控桔小实蝇诱捕过程和精确的计算桔小实蝇数量,且可自动传输到远程服务器并保存在本地存储卡中。在实验室环境下采用该装备测试,在830 s内有138头桔小实蝇进入该装备,系统检测出的结果是131头,检测成功率为94.9%。采用该装备在杨桃公园从2013年11月到2014年12月进行了一年多测试,系统软硬件可以稳定地协同工作,仅在光照严重不足太阳能供电不力的情况下出现过系统停止运行。基于物联网的桔小实蝇诱捕监测装备能自动跟踪计算桔小实蝇数量,从而向区域监控人员提供简洁有效的监控信息,在农业上有着广泛的应用前景。

       

      Abstract: Abstract: For bactrocera dorsalis field monitoring, the current method had some disadvantages: a heavy workload, low efficiency, poor reliability, low accuracy, and it could not large-scale and fast monitor the orchard pest situation in real time. Agriculture experts eager to have a solution to automatically count the number of bactrocera dorsalis and remotely observe the trapping result in real time to reduce their labor so that they could focus more on the study of the characteristics of insects. Therefore, combining the image target detection technology and the target tracking technology to develop an automated counting system by using a video image sensor would be necessary. In order to realize the real-time monitoring, the bactrocera dorsalis trapping, and a rapid diagnosis, an IOT-Based remote monitoring equipment for bactrocera dorsalis trapping was provided in this paper. The equipment included a trap monitoring device, a solar power supply device, and the monitoring control device. The trap monitoring device was comprised of a top cover, a transparent funnel, a trap bottle, a LED, and a camera; the solar energy device was comprised of a solar panel, a storage battery, and a solar panel bracket; the bactrocera dorsalis monitoring control system device was comprised of a Fit-pc controller, a 3G communication module, and the independent software for counting bactrocera dorsalis' numbers. This equipment combined machine vision technology and telecommunication technology with solar power technology. The purpose of the equipment was to achieve a whole function for bactrocera dorsalis trap monitoring with plant diseases and insect pests information collection, together with processing, transmission, and self-supply. It could monitor the trapping process and precise calculation of the number of bactrocera dorsalis anytime and anywhere, and also automatically transmit the results to the remote server or store it in a local storage card. For object extraction,this paper used an HSV color space for image filtering, then used median filtering and morphological filtering for the image to reduce white noise, eliminating holes in the target area to improve the image quality, and then divided the image into blocks based on the adjacent pixels of the image and used these blocks for Geometric feature matching, so that the bactrocera dorsalis area could be extracted. Finally, this paper used the watershed algorithm based on weight for an image segment to get the tracking object and tag the object. For bactrocera dorsalis tracking, this paper used a Kalman filter to predict the target movement position, narrowing the range of target searching and target matching, reducing the amount of calculation of the target matching; and established a cost model by using centroid Euclidean distance, survival time, and the color difference of the target between two consecutive frames; updated the cost model of each tracked target, and handled the missing target to ensure the stability and accuracy of the tracking algorithm. For bactrocera dorsalis metering, this paper studied a counting strategy for moving targets. This equipment was tested in the laboratory environment. There were 138 bactrocera dorsalis entered in the equipment in 830 s, the detected result of the system was 131, and the detection success rate was 94.9%. Also this equipment was tested in the Yangtao Park for more than a year (from November 2013 to December 2014), and the system hardware and software could work cooperatively and stably. The system stopped running only when there was a shortage of solar power, rendering it ineffective under grim weather for lack of the light. The IOT-Based remote monitoring equipment for bactrocera dorsalis trapping could automatically track and count the number of bactrocera dorsalis. It could also provide simple and effective monitoring information to the regional monitoring personnel, improve work efficiency, greatly improve the performance of the monitoring system, and had a wide range of applications in agriculture. This equipment had important practical applications.

       

    /

    返回文章
    返回