针对当前场地土壤环境污染监测任务中存在的效率低下、自动化程度低等问题,本文提出了一种基于无人车(unmanned ground vehicle,UGV)的场地土壤环境监测方法.该方法利用激光雷达和同步定位与建图算法(simulta-neous localization and mapping,SLAM)预建场地地图及本地坐标系,通过自动导航算法控制UGV在场地中沿预设的监测路径行驶,同时利用UGV搭载的监测传感器,获取监测路径上各检测位置的污染数据,动态调整检测点间隔,检测结果在边缘计算单元中实时处理并存储.基于空间插值算法和检测结果生成场地土壤污染区域分布图,高效自动地完成场地土壤环境污染监测工作.本文以土壤挥发性有机物(volatile organic compounds,VOCs)污染监测场景为例,对方法进行验证,基于多分类逻辑斯蒂回归算法评估空间插值的预测精度,结果为79％.
Mobile Monitoring Platform and Method of Soil Environmental Pollution Based on Unmanned Vehicle
Aiming at the problems of low efficiency and low degree of automation in current site soil environmental pollu-tion monitoring tasks,this paper proposes a site soil environ-ment monitoring method based on unmanned ground vehicle (UGV). This method uses Lidar and simultaneous localiza-tion and mapping (SLAM) to pre-build the site map and local coordinate system,controls the UGV to drive along the pre-set monitoring path in the site through the automatic naviga-tion algorithm,and uses the UGV to acquire the pollution da-ta of each detection position on the monitoring path,dynami-cally adjusts the detection point interval,and processes the de-tection results and stores the data in real time in the edge com-puting unit. Based on the spatial interpolation algorithm and the detection results,the distribution map of the soil pollution area of the site is generated,and the monitoring of the soil en-vironmental pollution of the site is efficiently and automatical-ly completed. This paper takes the soil volatile organic com-pounds (VOCs) pollution monitoring scenario as an example to verify the method. Based on the multi-class logistic regres-sion algorithm,the prediction accuracy of spatial interpolation is evaluated,and the result is 79％.
万方数据
维普
