双重定位系统的视觉神经网络监测农作物的室内无人机

为了解决传统室内农作物监测工作中人力物力消耗高、效率低、无法实施动态计标机监测等问题,设计并实现了用室内监测无人机代替检测人员完成各项检测。详述了室内检测无人机的机械结构组成;无人机由视觉神经网络系统、双重定位系统、通信系统等系统组成。双重定位系统包括激光雷达定位和双目视觉定位,同时通过视觉神经网络、通信协议模块、四旋翼飞控装置,能够实现室内双重精准定位、监测农作物生长情况、通信传输等功能。实验结果表明,室内无人机飞行误差和激光落脚点误差精度在1 cm之内,监测农作物生长情况的判断正确率高达97.5%。

In order to solve the problems of high consumption of manpower and material resources,low efficiency,and inability to implement dynamic marking machine monitoring in the traditional indoor crop monitoring work,the indoor monitoring drone is designed and realized to replace the inspectors to complete all the tests.The mechanical structure of indoor testing UAV is described in detail.UAV is composed of visual neural network system,dual positioning system and communication system.The dual positioning system includes laser radar positioning and binocular vision positioning.At the same time,through visual neural network,communication protocol module and quadrotor flight control device,it can realize indoor dual accurate positioning,monitoring crop growth,communication and transmission,etc.The experimental results show that the accuracy of indoor UAV flight error and laser foothold error is within 1 cm,and the judgment accuracy rate of monitoring crop growth is as high as 97.5%.