利用误差熵确定激光点云变形可监测指标

Using Error Entropy to Determine Deformation Monitoring Indicators

  • 摘要: 利用三维激光扫描确定变形区域的主要方法是对相同区域的点云进行对比分析,根据对比值确定变形区域及变形量,这种方法虽然能够简单地实现变形监测,但对于监测结果的可靠性并没有进行评价。因此,为了提高变形监测结果的可靠性,对点云误差及点云配准误差进行分析,并由此确定点云变形监测的可监测指标。为了避免相邻点位误差之间的相互影响及误差空间大小的不确定性影响,利用误差熵来确定点云误差空间,并根据其实际大小和误差极值的关系来确定变形可监测指标。通过不同距离和入射角下模拟的平面板变形来验证其可行性,并将该方法应用于某个滑坡场景,以确定该滑坡的变形区域和变形大小。

     

    Abstract: Deformation extraction and magnitude is typically estimated by comparing three-dimensional laser scanner scans over the same area but at different time epochs. Although, these methods can easily be realized, the reliability of deformation extraction has not been evaluated. In order to improve the reliability of deformation monitoring, the point cloud error and registration error need to be analyzed in this paper, and to determine the deformation monitoring indicator. In order to avoid the influence of adjacent point error and the uncertainty of error space, the error entropy is introduced to the determination of cloud error space and thus the deformation monitoring indicator can be determined according to the relationship between error space and error extreme value. The feasibility of deformation monitoring indicator is verified according to the simulated experiment of plane board deformation. The method proposed by this paper is applied to assure the region and magnitude of the landslide.

     

/

返回文章
返回