A 1.833Mb PDF of this article as it appeared in the magazine complete with images is available by clicking HERE
For commercial surveyors speedy data collection without degrading accuracy is the key. The less time you spend on-site, the lower your costs and the sooner the next project can start, so surveyors are always on the lookout for any tool or technique that can accelerate the survey process. That’s why I found myself out at a small mine in Germany one day in 2013 waiting for the small buzzing dot that was our Geo-Copter X-8000 UAV to land.
Some weeks earlier one of our clients had mentioned that they were getting excellent results with the Optech ILRIS terrestrial laser scanner that we had providedit was faster, safer and more detailed than surveying with a GNSS receiver. However, they wanted to shave off some of the time required for surveying open pit mines. Their usual survey technique involved scanning with their ILRIS from many different points around the mine, and they had noticed that they could capture the majority of their target, especially the vertical faces, with just their first few scanning positions.
However, the complex topography of open pit mines means that some areas can only be captured from particular positions. These areas are usually the horizontal surfaces of the mine, which are less important than the vertical surfaces and don’t really need the precision and detail of the ILRIS, but still have to be surveyed for a complete map. In the end, surveyors may spend 75% or more of their on-site time filling in these gaps and our client wanted to know if there was a faster way to do it.
Two Eyes for One Problem
This is where our new Geo-Copter X-8000 UAV entered the discussion. Designed for quick and easy surveying and mapping, it contains a Sony NEX-7 camera in a 2-axis gimbal and can survey an entire open pit mine very quickly. Afterwards, a surveyor can compare the overlapping images and use aerial triangulation to derive a rough digital elevation model of the survey area. The question was, if the ILRIS could quickly scan most of the mine from a small handful of positions, could the data from the X-8000 be used to fill in the gaps (Figure 1)?
To find out, I took our own ILRIS sensor and the X-8000 to a mid-sized 500500-m quarry in Germany for a proof-of-concept survey. Our goal was to cover the largest and most important parts of the mine with the ILRIS for maximum detail and accuracy, and then to use the UAV to survey the entire mine so that any gaps in the ILRIS data could be covered in the triangulated DEM and the points could be colorized.
Using traditional methods, such a mine would require around 12 hours of work to capture every nook and cranny, forcing the survey team to stay overnight near the mine or pull a very long day. By combining the ILRIS with the UAV, we intended to cut this down to 4 hours or less. If this approach worked a commercial surveyor would be able to drive a couple of hours to a mine, survey it quickly, and drop the data off at the office, all within a reasonable 8-hour working day.
Surveying at the Speed of Light
First things first. I knew we would need some ground control points (GCPs) to stitch the ILRIS LiDAR data and the UAV camera imagery together, so I sent part of my team with a geo-kombi GNSS receiver to collect 12 GCPs. These GCPs took about an hour to collect, since they were spread out across the mine for consistent coverage (Figure 2).
To save time, we completed the ILRIS survey while the GCP team was still collecting data. We limited ourselves to just 3 surveys with the ILRIS to minimize the time spent moving between survey positions and settingup/tearing down the sensor. Because such a small number of surveys would limit our fields of view, we were careful to position the surveys to maximize the mine area captured by the ILRIS.
This was made easier by installing the ILRIS on its Pan-Tilt base for some of the surveys, which let the sensor rotate 360 horizontally and 90 vertically. With this minimal set of survey positions, it took only 2.5 hours to cover 90-95% of the vertical details of the mine and 50% of the horizontal details. Now it was time to take to the skies.
As it turned out, the flight characteristics of the UAV work well for surveying open pit mines. The X-8000 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. Furthermore, 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 hunt for 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.
Twenty minutes later, the X-8000 had surveyed the entire mine and was back on the ground, ready to be packed up. The entire survey required only 3.5 hours on-sitewe had met our goal of covering the entire site in half a day. The only question now was whether the X-8000 data could be successfully merged with the LiDAR data from the ILRIS.
Into FocusMerging Two Sets of Data
Back at our processing center, we began by turning the collected data into 3D for comparison and merging. As the ILRIS LiDAR data is natively in point cloud format, we simply downloaded and registered the scans, and then georeferenced according to the ground control points and the GNSS receiver mounted on the sensor itself.
The imagery from the X-8000 camera required a little more post-processing work to turn it into 3D data. We georeferenced the imagery with Agisoft PhotoScan, which also automatically compared matching elements in overlapping images to generate a DEM. Finally, we transferred the triangulated DEM and the corresponding GeoTIFF into JRC 3D Reconstructor and merged it with the LiDAR point cloud based on the GCPs defined at the beginning of the survey (Figure 3).
Examining the data, it was clear that we should keep the accurate, high-resolution LiDAR data from the ILRIS wherever possible, while using the DEM from the X-8000 only to fill gaps in the LiDAR data. The LiDAR data had a point density of >115 points/ m and a precision of 2.5 cm, while the triangulated DEM had a point density of 10 points/m 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" (a series of rocks placed by the edge of benches), though it covered the entire mine.
The two surveying techniques complemented each other well. As the ILRIS was surveying from a horizontal perspective, it got its best results against the vertical walls of the mine but tended to get gaps where the line of sight was obstructed, such as areas blocked by vertical intrusions. By contrast, the X-8000 excelled at capturing the horizontal floors of the mine from its overhead perspective, but did not capture much detail on vertical faces.
Since the ILRIS had more precision and point density than aerial triangulation, the vertical faces were captured in higher detail than the horizontal faces. We considered this acceptable because the walls are the primary focus in many applications for open pit mines, such as rock-fall analysis and blast design. Furthermore, in many applications that require information about both the walls and the floors, such as safety and mine planning, far lower point density and accuracy is needed for the horizontal faces.
A Faster Way to Survey
The point extraction, georeferencing and merging were all completed in 4 hours, delivering a finished DEM that was ready for refined processing such as breakline extraction, volumetric analysis, and point cloud colorization based on the camera’s imagery (Figure 4). Combined with the 3.5 hours used for the survey itself, we had created a merged dataset of the whole mine from start to finish in less than a dayall while maintaining accuracy and resolution.
With this proof of concept in their hands, our customer decided to purchase their own X-8000 to fill out their ILRIS data. In the months since then, they have surveyed several mines for commercial projects and have reported good success and time savings, so we expect this method of combined ILRIS LiDAR and aerial photography to have a bright future among mine surveyors.
Jnos 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.
A 1.833Mb PDF of this article as it appeared in the magazine complete with images is available by clicking HERE