Abstract
The traditional practice to assess accuracy in lidar data involves calculating RMSEz (root mean square error of the vertical component). Accuracy assessment of lidar point clouds in full 3D (three dimension) is not routinely performed. The main challenge in assessing accuracy in full 3D is how to identify a conjugate point of a ground-surveyed checkpoint in the lidar point cloud with the smallest possible uncertainty value. Relatively coarse point-spacing in airborne lidar data makes it challenging to determine a conjugate point accurately. As a result, a substantial unwanted error is added to the inherent positional uncertainty of the lidar data. Unless we keep this additional error small enough, the 3D accuracy assessment result will not properly represent the inherent uncertainty. We call this added error “external uncertainty,” which is associated with conjugate point identification. This research developed a general external uncertainty model using three-plane intersections and accounts for several factors (sensor precision, feature dimension, and point density). This method can be used for lidar point cloud data from a wide range of sensor qualities, point densities, and sizes of the features of interest. The external uncertainty model was derived as a semi-analytical function that takes the number of points on a plane as an input. It is a normalized general function that can be scaled by smooth surface precision (SSP) of a lidar system. This general uncertainty model provides a quantitative guideline on the required conditions for the conjugate point based on the geometric features. Applications of the external uncertainty model were demonstrated using various lidar point cloud data from the U.S. Geological Survey (USGS) 3D Elevation Program (3DEP) library to determine the valid conditions for a conjugate point from three-plane modeling.
Highlights
As point densities of airborne lidar datasets increase, the need for full 3D absolute accuracy assessments of the associated lidar point clouds is becoming more important
To properly assess the absolute accuracy of airborne lidar data, one must have ground truth data that are of higher quality and accuracy than the data being tested
This paper documents extensive airborne lidar simulation modeling with a large array of pyramid targets in order to estimate the uncertainty associated with identifying a conjugate point
Summary
As point densities of airborne lidar datasets increase, the need for full 3D absolute accuracy assessments of the associated lidar point clouds is becoming more important. To properly assess the absolute accuracy of airborne lidar data, one must have ground truth data that are of higher quality and accuracy than the data being tested This involves using survey-grade global positioning system (GPS) checkpoints across the project. Dal Poz and Fernandes utilized high-resolution areal image along with lidar data in automatic extraction of building roof boundaries [26] With these other example methods in mind, this paper investigates geometric feature-based methods to identify a conjugate point from “ground truth data” represented in the airborne lidar point cloud for accuracy assessment. This paper documents extensive airborne lidar simulation modeling with a large array of pyramid targets in order to estimate the uncertainty associated with identifying a conjugate point. The development of the external uncertainty model is a crucial component in establishing a foundation of the 3D absolute accuracy assessment of the lidar point cloud
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call