基于无人机航测数据的甘南高原森林冠层高度和冠幅的提取
Extraction of forest canopy height and width using UAV aerial survey data in the Gannan Plateau
目的 以甘南高原的35个针叶林样方为研究对象,探究基于无人机航测数据提取森林结构参数的可行性,为日趋自动化的林业调查工作提供参考。 方法 研究基于无人机激光雷达数据分离植被点和地面点生成数字表面模型和数字高程模型,运用冠层高度模型提取树高,基于无人机正射影像利用ESP工具筛选最优分割尺度进行多尺度分割,优化特征空间进行面向对象分类,在ArcGIS中提取冠幅,将提取树高、冠幅与野外实测数据对比,进行精度评价。 结果 树高提取值与实测值拟合R2为0.90,RMSE为1.336 m,具有明显的线性关系;冠幅总体分类精度为90.18%,Kappa系数为0.80,东西冠幅与实测值拟合的R2为0.842,RMSE为0.509 m,,南北冠幅R2为0.876,RMSE为0.479 m,东西和南北方向冠幅的均值R2为0.878,RMSE为0.430 m。 结论 基于无人机遥感影像的森林结构参数自动提取方法高效可靠,在森林资源调查中有重要的现实意义。
Objective Based on the data from UAV aerial survey of 35 coniferous forest quadrangles on the Gannan Plateau,the feasibility of extracting forest structure parameters was explored to provide insights for the advancing automatic forestry investigations. Method In this study,vegetation points and ground points were segregated based on UAV LiDAR data to generate digital surface models and digital elevation models.A canopy height model was employed to derive tree heights,while the ESP tool was utilized to determine the optimal segmentation scale for multi-scale segmentation based on UAV imagery.The feature space was optimized for object-oriented classification,and canopy amplitudes were extracted using ArcGIS.Extracted tree heights and crown widths were compared with field-measured data to assess accuracy. Result The fitting values for tree height exhibited a strong linear relationship,with R2 and RMSE values of 0.90 and 1.336m,respectively.The overall classification accuracy of crown width was 90.18%,with a Kappa coefficient of 0.80.The R2 fitting for east-west crown width against measured values was 0.842,with an RMSE of 0.509 m,while for north-south crown width,it was 0.876 with an RMSE of 0.479 m.The average R2 for east-west and north-south crown widths was 0.878 with an RMSE of 0.430 m. Conclusion The automatic extraction method of forest structure parameters using UAV remote sensing imagery proved to be efficient and reliable,which was of practical significance in forest resource investigations.
无人机 / 森林结构参数 / 点云数据 / 冠层高度模型 / 面向对象
unmanned aerial vehicle / forest structure parameters / point cloud data / canopy height model / object oriented
| [1] |
邓云,王彬,李强,张志明, |
| [2] |
毕凯,李英成,丁晓波, |
| [3] |
侯峰.LIDAR详细介绍及其应用举例综述[J].科技广场,2014(4):95-100. |
| [4] |
艾萨迪拉·玉苏甫,玉米提·哈力克,阿不都拉·阿不力孜, |
| [5] |
Dalponte,Orka,HO, |
| [6] |
|
| [7] |
|
| [8] |
|
| [9] |
边瑞,年雁云,勾晓华, |
| [10] |
张玉薇,陈棋,田湘云, |
| [11] |
赵雪雁,李巍.中国地理学中的甘南研究[J].地理研究,2019,38(4):743-759. |
| [12] |
张志强,孙成权,吴新年, |
| [13] |
林鑫,庞勇,李春干.无人机密集匹配点云与机载激光雷达点云的差异分析[J].林业资源管理,2020(3):58-62. |
| [14] |
邵为真,赵富燕,梁周雁.基于不规则三角网的渐进加密滤波算法研究[J].北京测绘,2016(6):17-21. |
| [15] |
张俊,朱国龙,李妍.面向对象高分辨率影像信息提取中的尺度效应及最优尺度研究[J].测绘科学,2011,36(2):107-109. |
| [16] |
黄慧萍.面向对象影像分析中的尺度问题研究[D].北京:中国科学院研究生院(遥感应用研究所),2003. |
| [17] |
|
| [18] |
周辰琴,余拥军,方陆明, |
| [19] |
曾霞辉,王颖,曾掌权, |
| [20] |
高飞,史海静,税军峰, |
| [21] |
刘思康,史泽林,宋宏阳, |
| [22] |
王娟,陈永富,陈巧, |
| [23] |
张吴明,李丹,陈一铭, |
| [24] |
尤号田,邢艳秋,王铮, |
| [25] |
|
兰州交通大学优秀平台项目(201806)
干旱气象科学研究基金“半干旱区生态脆弱型城市森林生态效应评价研究:以兰州市南北两山为例”(1AM201903)
/
| 〈 |
|
〉 |