A Dithyramb on GNSS and Mobile LiDAR Mapping
by Jeff Fagerman, Lidar USA
Location, location, location. For most businesses to succeed this is absolutely essential. While the business location is not nearly as important for mobile mapping as a restaurant or clothing store, location is very important for a mobile LiDAR system to be useful. It is not possible to generate a point cloud from a moving platform without knowing the position of the system. This position does not have to be an absolute position and can in fact be a relative position. However, the most readily available means of obtaining positioning today is via GPS which is an absolute positioning system.
Often times when discussing mobile LiDAR people begin by saying they don’t need absolute coordinates so GPS is unnecessary. While this statement may be true, what they are really saying is they want to remove the GPS hardware and consequently reduce the cost of the system. They don’t really care about the coordinates as long as relative accuracy is preserved. They do care about costs. It seems to be a recondite principle, however, that GPS is the most readily available and affordable means of providing positioning.
Recall that the product generated by a mobile LiDAR system is a point cloud containing XYZ coordinates and usually intensity at the very least. Again, to build a point cloud the system must know precisely where the sensor is (position) and where it is looking (attitude or orientation) at virtually every instant. It is true, it does not require knowing where it is in the global sense. It could simply be in a localized coordinate system.
One could also argue that certain algorithms eliminate the necessity of knowing the position, which is only partly true. The position may not be a direct result of instrumentation but rather an indirect result of other observables and a suitable algorithm. This is very typical for surveying and photogrammetry where a resection is computed giving the final position. The solution in this case uses a least squares minimization of the error in the projection equations. It then determines the best estimate of the position of the instrument at the time the observations were taken. This is the very reason that so many photogrammetric and surveying solutions now have GPS built-in. At the very least, the GPS data provides a very good initial estimate enabling a rapidly converging solution often more accurate than the best GPS solution. We see this prevalent in most tripod scanners and (robotic) surveying total stations.
The same can be done using raw LiDAR data but it is definitely a complicated and compute intensive solution. It generally requires a calibrated multi-laser system to provide some structure in order to work. Such solutions are not much different than photogrammetric solutions from the 1970’s and early 1980’s using strip-wise adjustment solutions. These solutions will experience similar challenges from drift or cantilever extension over longer distances and will require overlap or external control to minimize errors.
Do non-GPS derived positioning systems commence an epicedium for the GPS solutions? It’s doubtful. By far the majority of the exterior scanning jobs require fairly accurate geo-referenced data. It is more likely that just as LiDAR and photogrammetry are synergistic, the same will hold for GPS positioning versus indirect observation derived positioning. We already have this with photogrammetry and it makes perfect sense that this will happen with LiDAR as well.
In summary, GPS is readily available everywhere and all the time. It ranges in accuracy from five (5) or more meters to centimeter level. The consumer cost continues to spiral even after more than two decades. We find GPS in many consumer grade products (watches, phones, cameras). It isn’t perfect. The positioning isn’t exact. It changes instantaneously. It only works in clear sky (for the most part). We can’t even think of a world without it now. For mobile LiDAR it remains an essential component and the prevalent means of positioning and will continue to do so indefinitely. Expect improvements and supplemental solutions.
For more information please contact Jeff Fagerman at Lidar USA, or please fill out the form below and we will contact you.