Terrestrial LiDAR and UAV Aerial Triangulation for Open Pit Mines

János Faust

Say what you will about the past few years, it has produced some useful tools for the surveyor. First, LiDAR technology made it possible to collect high-density 3D data with spectacular speed, so that now almost any surveyor can afford a terrestrial LiDAR scanner. This has created a large new market as customers find more and more inventive uses for 3D data. Now, ultra-light unmanned aerial vehicles (UAV) of every shape and configuration are invading every possible industry, drastically reducing the cost of survey techniques such as aerial triangulation.

geo-conzept GmbH has a great deal of experience with both technologies: We have used and sold Optech ILRIS terrestrial laser scanners (TLS) for 7 years and recently devised our own Geo-Copter X-8000 UAV and camera system for airborne imagery. Both technologies are capable of constructing 3D data (TLS natively and the UAV through aerial triangulation), so it is worth exploring the strengths and weaknesses of each. Since one of our client providing services to the open pit mining industry was interested in seeing how they could be combined, we recently took both a TLS and a UAV out to a nearby mid-sized quarry for comparison.

At the Site
We started by surveying the mine quickly with the Optech ILRIS TLS (Figure 1). One of the first TLS systems to come on the market, the ILRIS is well suited for open pit mines as its long range lets the surveyor scan from a safe distance, even from the opposite side of a mine. For some of the surveys we installed the ILRIS on its Pan/ Tilt base, which lets it rotate 360° horizontally and point either straight up or straight down. This was handy in such a wide, open environment. To assist with georeferencing afterwards, we also installed a GNSS receiver on the handle of the ILRIS.

Overall, we spent about 2.5 hours surveying with the ILRIS from 3 positions around the mine. Thanks to the Pan/Tilt base and the unit’s long range capability, we captured most of the mine from just these three positions.

Next, we brought out our new GeoCopter X-8000 UAV (Figure 2) for the aerial portion. Designed for quick and easy surveying and mapping, it contains a Sony NEX-7 camera in a 2-axis gimbal, which for this purpose would be used to generate a DEM using aerial triangulation.

As it turned out, the flight characteristics of the X-8000 work well for surveying open pit mines. It is extremely light, packing down to just 40 × 40 × 32 cm for transport, so we simply packed it into the back of our SUV alongside the ILRIS and drove it right to the survey site. As a rotary-winged aircraft, it can take off and land vertically in any open space near the survey site, so we did not have to rely on a runway near the mine.

Being designed for hands-off operation, the X-8000 follows a pre-defined survey plan entirely on auto-pilot, which helps prevent crashes and minimizes operator training. It also simplifies survey planning. Before the flight, we defined the geographical area to be surveyed from a map-based interface, set the operating parameters, and loaded the plan into the X-8000. Once we gave the command to start, it took to the air and covered the specified area automatically while we watched the data collection over the video link.

The UAV worked rapidly, with the flight completed in about 20 minutes, just under the UAV’s endurance of up to 25 minutes. A larger mine would have required multiple flights, but this would not have increased the survey time by much because the X-8000 operates on replaceable battery packs that can be quickly swapped in the field.

Processing Data
Back at our processing center, we were excited to compare the data collected by the two systems. LiDAR data lends itself easily to point cloud data, so it was easy to convert it all to 3D data. Georeferencing was also quite easy, as we had the data from the GNSS antenna on the ILRIS–we only needed to map the data to the ground control points collected before the survey.

We then turned the data from the X-8000 into a digital elevation model (DEM), which involved several steps more than with the LiDAR data. First, we used Agisoft PhotoScan to automatically merge the photos into a set of images connected through tie-points in overlapping photos. Then, we manually identified the 12 GCPs within the images and used PhotoScan to calculate the elevation of the tie-points in the imagery based on the GCPs, which produced a georeferenced DEM and a huge orthorectified aerial image. Finally, we loaded the DEM and the corresponding geoTIFF into JRC 3D Reconstructor for editing and analysis (Figure 3).

Data Results
Examining the data, there were some significant differences. First, the ILRIS data had considerably higher point density and precision than the UAV DEM. This was due to the inherent limits of aerial triangulation versus LiDAR: The LiDAR data had a point density of >115 points/m² and a precision of 2.5 cm while the triangulated DEM had a point density of 10 points/m² and a precision of 15 cm. As a result, the DEM missed small details that were picked up in the LiDAR data, such as the “pop stones” (the series of rocks placed by the edge of benches).

There was also a difference in how much of the mine the two systems covered. The mobility and airborne perspective of the X-8000 enabled it to take images of the entire mine, and therefore we could derive a DEM of the complete mine. This was important because we surveyed with the ILRIS from only a few locations around the mine, and the LiDAR data therefore did not include some areas that were occluded or hidden. While we could have caught these areas by surveying from more positions with the ILRIS, this would have increased the length of the survey.

The two techniques also differed in how well they mapped certain areas, although this had more to do with their perspectives than the technologies used. Since the ILRIS was scanning horizontally, the point density was higher on vertical faces like walls than on horizontal faces like floors. Likewise, because the X-8000 was surveying from above, it provided more detailed measurements on horizontal faces than on vertical ones.

Conclusions
The ILRIS TLS collects more precise and detailed data, especially from vertical faces, and its data is easy to process and georeference, but as a ground-based laser scanner it takes longer to cover sites with difficult topography. On the other hand, UAV-based aerial triangulation is extremely rapid and can easily cover the entire site, but it provides lower detail and precision and takes longer to post-process.

These strengths and weaknesses are complementary, and surveyors would benefit from using them together. While a full-scale survey with a TLS might take 1-2 days to complete, it can capture most of a mine within a few hours. It is especially adept at capturing the vertical faces of the mine, exactly the sections where detail is most critical for applications such as blast design, safety, and rock-fall analysis. For example, our 2.5-hour ILRIS survey captured 95% of the vertical faces of the mine.

If surveyors capture the most critical parts of a mine with the ILRIS TLS, they can then fill in the coverage gaps by using the aerial-triangulation DEM from the UAV. And since they then have both a point cloud and a georeferenced photo mosaic of the site, they can also colorize the point cloud to create a value-added deliverable for their client (Figure 4).

By combining these two technologies, a surveyor can acquire a complete survey of the target mine in only a few hours of onsite work, while still capturing the crucial aspects in high-density, high-precision data. For example, our survey of the mid-sized mine took only 3.5 hours on-site in total, which allows a surveyor to complete the whole survey in one day with ample time left for transit to and from the site.

Considering this, our customer decided to purchase their own X-8000 to fill out their ILRIS data for themselves. In the months since then, they have surveyed several mines for commercial projects in this way and have reported good success and time savings. We expect this method of combined ILRIS LiDAR and UAV aerial photography to have a bright future among mine surveyors.

János Faust works for geo-konzept in Germany. He completed his Master of Science degree in Geo-Ecology at Bayreuth University, focusing on different methods to survey soil erosion, such as 7Be-γ-ray-spectrometry and terrestrial laser scanners.

http://www.amerisurv.com/content/view/12214/