城市行道树调查中无人机飞行方案的优化  

Optimization of a drone flight plan for an urban street tree survey

作　　者：任华章 孙圆[1,2] 李纪霖 郝雨 林子航 REN Huazhang;SUN Yuan;LI Jilin;HAO Yu;LIN Zihang(College of Forestry and Grassland,College of Soil and Water Conservation,Nanjing Forestry University,Nanjing 210037,China;Co-Innovation Center for Sustainable Forestry in Southern China,Nanjing Forestry University,Nanjing 210037,China)

机构地区：[1]南京林业大学林草学院、水土保持学院,江苏南京210037 [2]南京林业大学南方现代林业协同创新中心,江苏南京210037

出　　处：《南京林业大学学报(自然科学版)》2024年第6期175-182,共8页Journal of Nanjing Forestry University:Natural Sciences Edition

基　　金：江苏省科技厅自然科学面上基金项目(BK20191388);南京林业大学大学生创新创业项目(202210298074Y);江苏高校优势学科建设工程资助项目(PAPD)

摘　　要：【目的】借鉴林业调查中的无人机遥感摄影测量技术,探索一套适用于城市行道树调查的无人机飞行方案优选流程,解决多种飞行参数和飞行方案导致的行道树调查数据源参差不齐的问题。【方法】针对行道树调查,设计6种无人机飞行方案,使用层次分析法(AHP)优选出兼顾精度与效率的飞行方案。根据优选方案提取的树高建立胸径反演模型,通过决定系数(R^(2))、平均相对误差(MRE)和均方根误差(RMSE)等指标选取最佳胸径模型。对株数、树高、冠幅和胸径获取精度与提取效率进行比较,证明经过优选的飞行方案可以在保持一定精度的前提下提高行道树调查效率。【结果】经过AHP分析,“井”字形(飞行模式)和70 m飞行高度为兼顾精度与效率的无人机飞行最佳方案;采用该方案提取的株数精度达到查准率(Precision值)94.44%,查全率(Recall值)80.95%,以及二者衡量指标(F_(1)值)87.18%;提取树高的平均相对误差为8.66%,均方根误差为1.22 m,决定系数为0.88;提取冠幅的平均相对误差为27.77%,均方根误差为1.30 m,决定系数为0.69;所建胸径预测模型决定系数达到0.81。该方案节约时长70%,且获取株数、树高、冠幅和胸径数值的准确性在所有方案中较高。【结论】经过优选的无人机飞行方案在达到行道树调查所需精度的前提下节约了70%时长,并获得较好株数、树高、冠幅和胸径数据。本研究提出的无人机飞行方案优选流程有效,可应用于城市行道树调查。【Objective】Drawing on unmanned aerial vehicle(UAV)remote sensing and photogrammetry techniques from forestry surveys,this study aims to develop an optimal selection process for UAV solutions tailored to the investigation of urban street trees.This approach addresses the problem of uneven data sources caused by the various flight parameters and flight plans.【Method】Six different drone flight plans for urban street tree surveys were designed,and the analytic hierarchy process(AHP)was applied to select a flight plan that balanced accuracy and efficiency.Based on the selected plan,a corresponding model for estimating tree height using tree diameter at breast height(DBH)was established.The best predictive model for DBH was determined by evaluating the coefficient of determination,mean relative error(MRE),and root mean square error(RMSE).By comparing the accuracy of the tree number,tree height,crown width,and DBH,as well as the efficiency of data extraction,it was demonstrated that an optimized flight plan for urban tree surveys could improve efficiency,while maintaining the required level of accuracy.【Result】After conducting an AHP analysis,a Tic-Tac-Toe flight plan at a flight altitude of 70 m was identified as the optimal solution that balances accuracy and efficiency.The optimized flight plan achieved a precision of 94.44%,a recall of 80.95%,and an F_(1)-score of 87.18%for tree number extraction.The average MRE for tree height extraction was 8.66%,with an RMSE of 1.22 m,and a coefficient of determination of 0.88.The average MRE for crown width extraction was 27.77%,with an RMSE of 1.30 m and a coefficient of determination of 0.69.The established model for predicting DBH achieved a coefficient of determination of 0.81.Compared to other flight plans,the optimized flight plan was conducted in 70%less time,while maintaining a high level of accuracy in tree number,tree height,crown width,and DBH measurements.【Conclusion】The optimized UAV flight plan,after achieving the required accuracy for street tree s

关 键 词：行道树调查 倾斜摄影测量 无人机 层次分析法 回归模型 

分 类 号：S757.3[农业科学—森林经理学] P231[农业科学—林学] TP79[天文地球—摄影测量与遥感]