Research On The Key Technologies Of Pest Monitoring Based On Internet Of Things

Agricultural Internet of Things is the organic combination of modern agriculture and Internet of things technology,and promotes the development of modern agriculture from computer agriculture,digital agriculture,and precision agriculture to smart agriculture.The extensive interconnected technologies,flexible sensing technologies,and sophisticated intelligent technologies in smart agriculture have made modern agricultural systems more standardized,smarter,and more intensive,thus promoting the sustainable development of agriculture.The Internet of Things is the supporting technology of smart agriculture.The sensing devices of agricultural internet of things are developing in the direction of micro-power consumption,low cost,high reliability and self-adaptation.The sensor network is also gradually distributed,self-organizing,multi-protocol compatible and high-pass,quantitative and other functional characteristics,real-time,accurate and efficient data processing.However,the Internet of Things is still not widely used in pest monitoring and other areas.The main reasons are that the performance of the sensor node does not meet user requirements,the data type collected by the node is one dimensional,and the management and application level of multi-sensor data fusion are not high.Therefore,the research on the Internet of Things technology integrated with multiple types of sensor nodes for pest monitoring is of practical significance in promoting the development and application of Internet of Things and sensor networks in the field of pest monitoring,improving the level of intelligence in pest monitoring,and promoting the development of agricultural internet of things.The paper combines the latest technology of the agricultural internet of things,uses pest monitoring as the application object,and monitors the system architecture of the machine-based pest monitoring wireless acquisition node software and hardware design,multi-sensor integration pest monitoring remote wireless network design,pest monitoring wireless network system,the research on key technologies such as the application of UAV monitoring data in visualized pest monitoring,and the main contents of the research are as follows:(1)A pest acquisition monitoring wireless node adapted to scene lighting changes was designed.The node consists of the thorny fruit fly trapping monitoring device,control device,network transmission module and solar power supply.The pest trap monitoring device includes a top cover,a PVC shell and a trapping bottle.The monitoring control device is composed of a processing module,a storage module and a video acquisition module.The solar power supply includes a solar panel,a battery,an intelligent controller and a solar panel bracket.Insect pest target acquisition algorithm,cost model-based pest target tracking algorithm and pest automatic measurement algorithm are designed,and the effectiveness of the algorithm is verified through experiments.After the real experiment of farmland environment,in order to improve the robustness of the algorithm,the pest target detection algorithm was optimized and improved.Using the method of background difference,a detection algorithm for the orange fruit fly that can adapt to the change of scene lighting was proposed.Detailed design of solar powered devices extends the life cycle of the nodes.A visualized remote monitoring network system for Bactrocera dorsalis was designed and a monitoring and tracking program,a remote server program and a client program were developed.Farmland experiments show that the packet loss rate of node network transmission is 0.7%.The improved detection algorithm has 7.21%and 12.4%error rate under the influence of moderate light and severe light,and the time-consuming reduction is 42.2%.The pest detection accuracy rate is 98.7%.(2)A multi-type sensor node fusion pest remote monitoring wireless network was designed.Combining the features of the designed pest monitoring wireless acquisition node,incorporating multiple types of meteorological sensors and soil sensor nodes,as well as UAV image monitoring nodes,according to the needs of practical applications,a set of software and hardware combinations,ground surface underground and more Collaboration pest monitoring remote wireless network architecture,including multi-sensor integration of the sensor layer,network pest monitoring remote system network model,application layer pest monitoring IoT cloud platform software.(3)Pests monitor transmission models and transmission control mechanisms for remote wireless networks.In combination with practical needs,the topological structure,network construction and transmission model of pest monitoring remote network were designed in the network model,and communication protocols,data packet formats,data backup mechanisms for multi-level storage,error control mechanisms,clock synchronization strategy and transmission control mechanisms were designed.A web-based virtual internet of things on cloud platform software for pest monitoring is designed to integrate and effectively manage the data collected by multiple types of nodes,and at the same time provide users with reliable network services.(4)A comprehensive test of the veracity monitoring remote wireless network system was conducted.In accordance with the designed test plan,multiple types of meteorological sensors and soil sensor wireless network nodes were deployed in farms and scientific research bases in four different administrative regions of Guangzhou,and high resolution orthographic images were acquired by using UAV image monitoring technology.Image maps and digital surface models,established a remote wireless network system based on the Internet of Things pest monitoring,carried out a long-distance network comprehensive test across regions.After statistical analysis of the data,the average data packet loss rate of the network is 2.965%.The statec~2 test method was used to detect outliers.The probability of occurrence of outliers for sensors,such as air temperature sensors,air humidity sensors and soil temperature sensors was 2.23%,0.83%and 1.69%respectively,and the maximum spatial correlation of temperature data for different monitoring areas was0.9934.Based on the quantitative relationship and correlation analysis between sensor data and pest occurrence after data fusion,the six most important environmental factors affecting pest occurrence are soil temperature,leaf moisture,air temperature,rainfall,soil moisture,and wind speed,which established a regression analysis equation as well as a pest warning model.The drone data analysis shows that the optimal flying height in the monitoring area is 15 meters.The G component of the orthographic image can effectively monitor the impact of pests on lettuce growth.The main innovative work of this paper is embodied in the design and implementation of a pest monitoring wireless acquisition node,using the BDDA-LV algorithm which can adapt to the change of scene illumination,and the remote wireless network architecture of pests monitoring integrated with multiple types of sensor nodes.