单目视觉定位与UWB技术融合的育苗温室移动平台定位方法研究  

Research on the localization method of seedling greenhouse mobile platform based on fusion of visual beacons and UWB

作　　者：张岳阳 王宇珂 吕宝轩 周林凡 甄军全 王东[1,2,3] ZHANG Yueyang;WANG Yuke;L Baoxuan;ZHOU Linfan;ZHEN Junquan;WANG Dong(College of Mechanical and Electronic Engineering,Northwest A&F University,Yangling,Shaanxi 712100,China;Key Laboratory of Agricultural Internet of Things,Ministry of Agriculture and Rural Affairs,Yangling,Shaanxi 712100,China;Shaanxi Key Laboratory of Agricultural Information Perception and Intelligent Service,Yangling,Shaanxi 712100,China)

机构地区：[1]西北农林科技大学机械与电子工程学院,陕西杨凌712100 [2]农业农村部农业物联网重点实验室,陕西杨凌712100 [3]陕西省农业信息感知与智能服务重点实验室,陕西杨凌712100

出　　处：《西北农林科技大学学报(自然科学版)》2025年第5期146-156,共11页Journal of Northwest A&F University(Natural Science Edition)

基　　金：国家自然科学基金项目(31701326);陕西省重点研发计划项目(2020zdzx03-04-02)。

摘　　要：【目的】研究一种单目视觉定位与超宽带(ultra wide band,UWB)技术融合的育苗温室移动平台定位方法,为育苗温室移动作业平台的高精度定位提供技术支持。【方法】首先定义一种可用于单目视觉定位的视觉信标编码规则,并制作一套视觉信标;然后依据测定视觉信标2个顶点的像素距离以及相机像平面与视觉信标之间的实际距离,使用非线性最小二乘法拟合单目视觉测距函数,在此基础上构建二维定位模型;之后采用联邦卡尔曼滤波算法将单目视觉定位与UWB技术融合,构建育苗温室移动平台定位系统,并对该定位系统的性能进行了测试。【结果】对视觉信标中2个顶点的像素距离以及相机像平面与视觉信标之间的实际距离进行拟合,得到的测距函数拟合度为0.999;随机选取8个位置对测距函数进行验证,结果表明,最大测距绝对误差为6.8 cm,平均测距绝对误差为3.3 cm,信标与相机之间的距离越远,测距误差越大。选取15个位置进行的静态定位试验表明,最大绝对误差为10.5 cm,最小绝对误差为4.6 cm,平均绝对误差为6.5 cm。单目视觉定位与UWB技术融合后,定位系统平均漂移绝对误差从融合前的6.2 cm降低至3.0 cm,降低了51.6%。【结论】基于单目视觉定位与UWB技术融合的定位方法明显提高了育苗温室移动平台的定位精度,可将该技术应用于温室设施农业机器人的开发。【Objective】This research aimed to find out a method for positioning a seedling greenhouse mobile platform by integrating monocular visual beacon positioning with ultra wide band(UWB)technology to provide technical support for the seedling greenhouse mobile operation platform positioning of high precision.【Method】Firstly,a visual beacon encoding rule that can be used for monocular visual beacon positioning was defined,and a set of visual beacons was created.Based on the pixel distance between the two vertices of the visual beacon and the actual distance between the camera image plane and the visual beacon,a nonlinear least squares method was used to fit the monocular visual ranging function,and a two-dimensional positioning model was constructed.Afterwards,the federated Kalman filter algorithm was used to integrate monocular visual beacon positioning with UWB technology to construct a mobile platform positioning system for seedling greenhouses,and the performance of the positioning system was tested.【Result】The pixel distance between two vertices in the visual beacon and the actual distance between the came-ra image plane were fitted,and the fitting degree of the ranging function turned out to be 0.999.8 positions were randomly selected to verify the distance measurement function.The results showed that the maximum absolute distance measurement error was 6.8 cm,and the average absolute distance measurement error was 3.3 cm.As the distance between the beacon and the camera increased,the distance mea-surement error became larger.Fifteen positions were selected for static positioning experiments,and the results showed that the maximum absolute error was 10.5 cm,the minimum absolute error was 4.6 cm,and the average absolute error was 6.5 cm.After the fusion of monocular visual beacon positioning and UWB technology,the average drift absolute error of the positioning system decreased from 6.2 cm before fusion to 3.0 cm,a reduction of 51.6%.【Conclusion】The positioning method based on the fusion of monocul

关 键 词：育苗温室 移动平台定位 单目视觉 超宽带 视觉信标 

分 类 号：S24[农业科学—农业电气化与自动化]