Aerial Photogrammetry or Aerial LiDAR for Mapping?
Posted on July 14 2017
Short answer? Both. The long answer is a bit more complex.
Aerial mapping using photogrammetry is fairly common these days. If you haven't seen an aerial 3D map, today's your lucky day! The map below can be rotated to change perspective and has a zoom function as well. Take a look.
That map was created using photogrammetry (the use of photography in surveying and mapping to measure distances between objects) and software to stitch multiple 2-dimensional photos into a 3D image that can be manipulated and measured. Traditional photogrammetry derived elevation by triangulating the ground from two different images of the same area on the ground. The images are displaced by a ‘baseline’, allowing one to use triangulation to discern elevation. The weakness of this was the fact that to measure elevation, a “ray” from each of the two images had to converge at the ground point one wished to measure. This is difficult in environments with overhead obstructions such as trees. Laser scanning (Laser Imaging, Detection and Ranging, LiDAR) has the distinct advantage of requiring only a single ray to measure elevation.
Drone technology has completely changed the considerations for companies trying to decide which option is best for their specific company or project. New software has made the collection and construction of the 3D models completely autonomous. With the click of a button, a drone can be sent up to collect images of a specific area with specific overlap, every aspect of the flight and image capture is controlled precisely by the program.
All that automation only counts for so much, though. If you’re doing photogrammetry, as long as the user follows the exact requirements that you need to create a 3D model, it can be quite accurate. For those areas where you have to measure a property and figure out how much dirt you’re going to clear, especially if it’s all bare earth, it’s an easy choice to use photogrammetry. It will be quite accurate even when compared to the info gathered from a LiDAR system with an expensive IMU and navigation system. But as soon as you add complex structure to it, you need an alternative.
photo: Geoffrey Morrison
Photogrammetry also depends heavily on lighting and contrast. When faced with an image such as the one above, the model generating software will likely have a difficult time discerning any differences in terrain. LiDAR, on the other hand, does not need good lighting or contrast to be able to map accurately.
When it comes to budget, photogrammetry wins hands down. You can get a Phantom 4 Pro for about $1,500 and get a decently accurate 3D map using photogrammetry with the only additional costs being for the software, which are largely subscription based.
LiDAR, which stands for Light Detection and Ranging, is a remote sensing method that uses light in the form of a pulsed laser to measure ranges (variable distances) to the Earth or another target as each individual application needs. The LiDAR instrument principally consists of a laser, a scanner, and a specialized GPS receiver. So good lighting isn't necessary since the laser makes its own light!
LiDAR is generally deployed in a manned aircraft, which is incredibly expensive for a single use scenario. This also limits the size for a project as well. Most companies cannot justify spending large sums of money for a manned LiDAR scanning crew for a small project. There are slightly less expensive ways to deploy LiDAR scanning, such as from a truck or crane. But these options still require crews to be in precarious positions.photo: Aly El-Kadi
LiDAR, because of its more precise nature, will yield better measurement data. That being said, LiDAR does not offer texture recreation which makes visualization much more difficult. For example, the image below is a model comparing LiDAR and photogrammetry.
The lidar image (on the right) shows much more accurate elevation data. You can see more precisely where obstructions like trees and buildings stick above the ground, whereas on the orthomosaic (photogrammetry map, on the left) the image is very flat. On the other hand, the LiDAR image doesn't offer details that can be easily discerned by texture (is that a playground or a scrapyard, and which is more important for this project?).
Getting back to the first sentence: having both is the best case scenario. But depending on the needs of the project, one or the other can be sufficient to complete the project.
Looking to dig a bit deeper? Check out these articles:
Even deeper? 3D Flash LiDAR may be the future of LiDAR mapping with 16,384 data points per single flash frame. Advanced Scientific Concepts - Technology Overview: 3D Flash LiDAR