地球信息科学学报

地球信息科学学报 ›› 2023, Vol. 25 ›› Issue (5): 935-952.doi: 10.12082/dqxxkx.2023.220766

• 地球信息科学理论与方法 • 上一篇    下一篇

林区数字高程模型修正方法：顾及高程自相关的后向传播神经网络模型

李琳叶(), 李艳艳(), 陈传法, 刘妍, 刘雅婷, 刘盼盼   

  1. 山东科技大学测绘与空间信息学院， 青岛 266590
  • 收稿日期:2022-10-09 修回日期:2023-01-03 出版日期:2023-05-25 发布日期:2023-04-27
  • 通讯作者: *李艳艳（1987— ），女，山东潍坊人，副教授，主要从事空间数据质量改善及高频GNSS数据处理研究。 E-mail: yylee@whu.edu.cn
  • 作者简介:李琳叶（1996— ），女，四川广安人，硕士生，主要从事空间数据质量改善。E-mail: lilinye2022@163.com
  • 基金资助:
    国家自然科学基金(42271438);山东省自然科学基金项目(ZR2020YQ26);山东省高等学校青创科技支持计划(2019KJH007)

Method for the Correction of Digital Elevation Models Over Forested Areas: Back Propagation Neural Network with the Consideration of Spatial Autocorrelation

LI Linye(), LI Yanyan(), CHEN Chuanfa, LIU Yan, LIU Yating, LIU Panpan   

  1. College of Geodesy and Geomatics, Shandong University of Science and Technology, Qingdao 266590, China
  • Received:2022-10-09 Revised:2023-01-03 Online:2023-05-25 Published:2023-04-27
  • Contact: LI Yanyan
  • Supported by:
    National Natural Science Foundation of China(42271438);Shandong Provincial Natural Science Foundation, China(ZR2020YQ26);Project of Shandong Province Higher Educational Youth Innovation Science and Technology Program(2019KJH007)

RichHTML

PDF (PC)

赞

摘要：

受植被遮挡影响，卫星遥感技术获取的全球数字高程模型（Digital Elevation Model，DEM）在林区难以准确描述真实地表形态，且在不同林区类型表现出不同系统偏差。为提高林区DEM精度，本文提出了一种顾及高程自相关的后向传播神经网络（Back Propagation Neural Network，BPNN）模型。该模型首先对训练区高程点拟合最优半变异函数以确定其变程，并将距离目标点变程以内的高程点作为高程自相关最优邻域，然后将地形因子（坡度、坡向、地形起伏度）、植被因子（植被高度、植被覆盖度）以及变程范围内高程点作为影响因子，DEM与对应LiDAR（Light Detection And Ranging）DEM高程差作为预测值，构建并训练BPNN模型，最后用训练好的模型修正测试区DEM。为了验证模型的实用性和高效性，本文以4种林区（常绿阔叶林、常绿针叶林、混交林、落叶阔叶林）DEM为研究对象，分别训练BPNN模型。同时，将修正结果与4种模型进行比较，包括综合利用4种林区类型数据训练的BPNN模型（BPNN-T）、没有使用地形因子的BPNN模型（BPNN-W）、没有顾及高程自相关的BPNN模型（BPNN-R）和多元线性回归模型（Multiple Linear Regression，MLR）。对上述4个林区的DEM（包括SRTM1、AW3D30、TanDEM-X（TDX）90）修正结果表明：① 与修正前相比，采用顾及高程自相关的BPNN模型显著提高了4种林区DEM精度，使其平均误差（Mean Error，ME）降至0~1 m，均方根误差（Root Mean Square Error，RMSE）降低了46%~70%； ② 修正前、后TDX90误差受坡向影响显著，AW3D30其次，SRTM1最小；修正前、后DEM的RMSE均随坡度和地形起伏度的增加而增大；③ 修正前DEM误差随植被高度和植被覆盖度的增加而增大，修正后该规律消失，表明BPNN有效消除了植被对DEM精度影响；④ 5种模型中，顾及高程自相关的BPNN预测效果最优，BPNN-T略次之，MLR和BPNN-W次之，BPNN-R效果最差。因此，充分考虑地形因子并分别对4种林区类型构建高程自相关训练模型可显著提高DEM精度。

关键词: 数字高程模型, 机器学习, 误差, 修正, 精度, 高程自相关, 林区

Abstract:

Due to the limitation of earth observation technology, the existing global Digital Elevation Model (DEM) datasets usually contain information of vegetation, buildings, and other non-ground objects. Especially in forested areas, the DEM data usually cannot describe the bare-earth surface precisely and show large systematic deviations. This study proposes a Back Propagation Neural Network (BPNN) model that takes into account the spatial autocorrelation of elevation to reduce the errors of bare-earth DEM in forested areas. This model first fits the optimal semivariogram to determine the spatial variation of elevation and takes the elevation points within the variation range from a target point as the optimal spatially autocorrelated neighborhood. Then, we train the BPNN model by using the terrain factors (i.e., slope, aspect, and terrain undulation), vegetation factors (i.e., vegetation height and vegetation coverage), and elevation points within the range of variation as the influencing factors, and using the elevation difference between DEM and Light Detection And Ranging (LiDAR) DEM as the predicted value. Finally, the trained model is used to correct the DEM in different forested areas. In order to verify the practicability and efficiency of the model, this paper takes the DEM products including SRTM1, AW3D30, and TanDEM-X (TDX) 90 of four types of forests (evergreen broad-leaved forest, evergreen coniferous forest, mixed forest, and deciduous broad-leaved forest) as the research objects, and trains the BPNN model respectively for each forest type. The correction result is compared with BPNN trained with all four types of forest data (BPNN-T), BPNN trained without terrain factors (BPNN-W), BPNN trained without spatial autocorrelation of elevation (BPNN-R), and multiple linear regression model (MLR). The results show that: (i) The BPNN model significantly improves the accuracy of DEM in the four forests, with the Mean Error (ME) close to 0-1 m and the Root Mean Square Errors (RMSE) reduced by 46%~70%; (ii) The aspect has the largest influence on the DEM correction for TDX90 while has little influence on SRTM1 DEM correction. Before and after correction, the RMSE of each DEM increases with the increase of slope and relief; (iii) The DEM error increases with the increase of vegetation height and vegetation coverage before correction, but this trend disappears after correction, indicating that BPNN effectively eliminates the impact of vegetation on bare ground DEM; (iv) BPNN has the highest prediction accuracy, followed by BPNN-T, MLR, and BPNN-W. And BPNN-R has the worst prediction accuracy. Therefore, the accuracy of DEM can be significantly improved by fully considering terrain factors and spatial autocorrelation of elevation for different forest types.

Key words: DEM, machine learning, error, correction, accuracy assessment, elevation autocorrelation, forest region