Author: Lewis Graham
The Challenge
We recently performed a drone LiDAR Mapping experiment with a GeoCue TrueView 410 client. The test was to see how well the TrueView 410 Imagery/LiDAR sensor (3D Imaging System, 3DIS) could perform when mapping relatively steep terrain.
Steep Terrain Test Site
We selected a site about 1 hour away from our GeoCue office located in the foothills of the Appalachian Mountains in northeastern Alabama. The site, referred to as the Rock Farm (you can guess the reason for this name!), is depicted in Figure 1. The grid spacing in the planimetric view is 10 m x 10 m and 5 m x 5 m in the profile view. The approximate LiDAR/imagery coverage area is 0.64 km2 (160 acres). The project distance along the example profile line shown in white in Figure 1 is about 1,100 m (1.1 km). The rise over this segment is 195 m. This gives an average slope of 18%.
Figure 1 – The Rock Farm
Conducting the Flight with TrueView 410
We normally recommend a flying height of 75 m. We performed the 75 m test with excellent results and then the client wanted to experiment with a 100 m flight. We used Litchi (from VC Technology Ltd) mission planning software and flight management system for the 3D flight. The Litchi software uses an elevation database to model the terrain for automatic 3D mission planning (note that by year-end, we should have 3D mission planning added to our True View Evo processing software)
We consider the useful maximum range of the True View 410 to be around 75 m. The LiDAR sensor we use is rated to 200 m, but we like to give you very useful, practical specifications. We were very pleased with the 100 m range results for this type project. We observed densities in the ground class in treed areas ranging from 125 pts/m2 down to 25 pts/m2. The conditions are leaf-off. Thus you can see that we had excellent results with this project. The 75 m recommendation comes in to play for very low reflectivity surfaces such as pavements.
Post Processing with TrueView Evo
The time to process a True View 410 project to the level of a fully colorized point cloud is about 1/3 the flight time. This project required about 30 minutes of flight time in two separate flights and thus post-processing was around 10 minutes.
We used our TrueView Evo automatic ground classification to separate the ground points from all other points (in this project, mostly trees). Running the ground classifier required 2 minutes and 4 seconds. The results, even with no manual cleanup, were sufficient to support a gridded elevation file or topographic contours. As an example, we used TrueView Evo to produce a set of 1m contours of the site using the ground surface computed from the automatic ground classification algorithm (see Figure 2).
Figure 2 – 1 m Topographic Contours of the Rock Farm
Summary Lessons
LiDAR with colorization algorithms from fused cameras is significantly faster to post-process (minutes versus hours) than photogrammetry
For vegetated areas, LIDAR is a requirement. Photogrammetry cannot get enough match points on the ground to create a surface model
Steep terrain is fairly easily managed using 3D mission planning software such as Litchi
The TrueView 410 has completely acceptable performance at 100 m range for this type project
Post-processing software needs to be specifically designed for the sensor in use. For GeoCue systems, this is True View Evo.
I think it is quite remarkable to fly a site such as this in the morning and deliver derivative products in the afternoon. The Return on Investment of the technology really is a step jump as opposed to an incremental improvement. With subscription plans available for the True View 410, testing the waters is easy and very low risk. If you feel that the True View 410 system would fit your needs, feel free to contact below or give us a call at 256-461-8289.