基于机器视觉的森林昆虫行为自动跟踪记录系统设计研究

Research on the Design of Automatic Tracking and Recording System for Forest Insect Behavior Based on Machine Vision

我国的森林面积广阔,多种病虫害和复杂环境给防治工作带来了诸多困扰。为了达到对森林病虫害防治的目的,从昆虫的行为出发,对森林活体昆虫-花绒寄甲的行为特征进行观察与研究。基于机器视觉技术与昆虫学、林学相结合,自主研发了一套森林昆虫行为自动跟踪记录系统,主要包括:总体设计、硬软件设计和功能测试。利用基于机器视觉技术的OpenCV和基于camshift、meanshift等的跟踪算法,通过可视化界面操作实现了对不具备飞行能力与跳跃能力的活体昆虫——花绒寄甲的行为跟踪观察和实时录制,且能实时显示昆虫的位置坐标、偏移距离以及速度等参数。结果表明,可实现对活体昆虫的行为跟踪及观察;在此基本功能上添加了视频录制按钮,可对观察的森林昆虫进行实时录制等。以期为采用机器视觉自动检测法有效开展森林有害昆虫防治或对有益昆虫进行保护提供有益参考,具有较好的应用价值。

China has a vast forest area, and various pests and diseases and complex environment have brought many difficulties to the prevention and control work.In order to achieve the purpose of prevention and control of forest diseases and insect pests, this study starts from the behavior of insects, and observes and studies the behavioral characteristics of the living insects in the forest-Dastarcus helophoroides.Based on the combination of machine vision, entomology and forestry,this paper independently developed a set of automatic tracking and recording system for the behavior of insects in forests. The main contents include overall design,hardware and software design and functional testing. In this paper,using Open CV based on machine vision technology and tracking algorithms based on Camshift and Meanshift,the observation and recording of Dastarcus helophoroides can be realized through visual interface operation. It realizes the behavior tracking observation and real-time recording of the living insects( e. g.,Dastarcus helophoroides) that do not have the ability to fly and jump and can display the parameters such as the position coordinates,offset distance and speed of the insects in real time. The conclusion shows that this design result can realize the behavior tracking and observation functions of living insects. If we add a video recording button to the basic functions,the system can record the observed harmful insects in real time. The above can provide a useful reference for the effective control of harmful insects in forestry and forests in my country or the protection of beneficial insects by using the automatic detection method of machine vision and has good application value.