Why Altavian Doesn't Use Lidar Systems--Yet
Altavian is often asked when—and—if we will be offering a Lidar payload. The new era of mapping has brought a flood of professionals who speak, eat, and breathe Lidar (or more technically: point clouds). But, is Lidar the appropriate technology? Or are we all just to use point clouds and photogrammetrically derived models to fit the bill? In this article, we’ll look at the pros and cons of deploying a Lidar system for an sUAS, and give some insight into our upcoming path.
Viable Systems Today
Let’s begin with some of the early visibility and success of the technology in our industry stretching all the way back to the DARPA Grand and Urban Challenges in the early-to-mid 2000s. In the development of aiding sensors for self-driving cars, Lidar was adapted from large and complex manned systems as a potential mapping tool. It wasn’t until just recently that the emerging market for small unmanned aircraft produced a requirement for lower-cost scanning time-of-flight (TOF) laser ranging systems.
So far, of the systems we have surveyed, two stand out as viable candidates for UAS: Velodyne HDL-32 and VLP-16 and Reigel’s VUX-1UAV. Riegl’s is the only system on offer from the Big Three of the survey-grade Lidar firms that is found in the sUAS market. However, at a minimum weight of 3.5 kg (7.7 lbs), without counting an INS or storage/control system, it pushes the boundary of what one could call an sUAS payload. In fact, Riegel quickly developed the RiCOPTER platform to specifically carry this sensor, as manufacturers were not quick to adopt this heavy offering.
Meanwhile, Phoenix Aerial Systems took the leap with Velodyne and produced complete payloads capable of 15 minute flights onboard a DJI S1000 class vehicle. So, what can we expect of the quality of data per the manufacturer? Phoenix Aerial’s website says at altitudes approaching 60m and at 15-20 m/s cruise speeds (typical of Altavian’s F7200 fixed-wing aircraft), point density of the AL3 systems hovers around 50-60 points/m2. Moreover, the MEMS IMUs in the Scout lead to georeferencing errors between 8 and 16cm, depending on options. Then, there is the wide beam divergence of 2.79 milliradians (mrad), compared to the VUX-1UAV’s 0.5 mrad along with the range repeatability nearing 2cm at one sigma for the Velodyne systems vs. 5mm for the Riegl.
Analyzing the Cost Difference
However, the cost difference is more than significant. Consider that the difference between the systems in the AL3 and the Scout series, disregarding the laser ranging sensor used, differ greatly in the quality of the IMU’s integrated, with price points ranging from around $1,700 to well beyond $10,000. Thus, it is important to keep in mind the ROI of such systems, given that one could easily spend up to $180,000 on state-of-the-art technology.
Perhaps unsurprisingly, the workflow remains a sticking point in Lidar surveys. The photogrammetric community has seen an explosion of growth in processing packages; from Pix4Dmapper, SimActive’s Correlator3D, and new online offerings such as DroneDeploy. Yet, the Lidar community has not experienced such a diversification due to the market size for sensors.
Systems must routinely be calibrated, both to determine boresight angles and to account for sensor systematic error. Lastly, and most critically from the standpoint of being both a fixed-wing and VTOL platform manufacturer, we would reach the absolute usable limits of a Velodyne system aboard the Nova F7200 at 400ft and under normal cruise conditions. The Galaxy R8700 is a different story as you have direct control over altitude and forward velocity.
Examining Lidar as a Tool
Now that we’ve explored the limitations of using Lidar on an sUAS taking into account the platforms we produce, let’s explore some cases why it would be beneficial (or the only option).
It is our conviction as a company that every tool has a use. Is a Lidar point cloud cleaner than a photo-derived one? Absolutely. Can Lidar penetrate vegetation, even with a single-return design? Yes, under some conditions. What about water surfaces? Lidar can measure directly the surfaces of water bodies in places where photogrammetric approaches are not even worth getting out of bed for. Lidar data is less dense on the ground, all things being equal, but you can dial density on a VTOL just by height and speed settings—if you are willing to trade total area covered per flight. Aside from that, density can be a dangerous stand-in for the perception of a dataset’s accuracy.
Perhaps the most telling fact of examining Lidar as a tool is that its technology alternative is catching up. Previously, it could be argued that Lidar was able to capture certain targets that photogrammetric approaches could not, such as power lines.
In this older comparison, you can see how only Lidar was able to capture power lines as they extended over a sandy, reflective road. This held true with early versions of photogrammetric data processing software, but as these programs improved, hard-to-capture objects such as power lines began to appear.
In this second example, power lines are easily captured by photogrammetric means without the use of expensive Lidar systems.
Altavian has been keeping a close eye on all these developments since 2012. Some may know that its CTO has actually been involved in Lidar mapping technologies for half his professional career. So why isn’t there a Lidar offering from Altavian yet? The answer to that is based in a combination of technological, and strategic considerations. Every year, good INS and active sensors get smaller and cheaper. Our philosophy is to aim for producing industrial-grade equipment that will not wear out easily and, thus, give our customers heartburn in the field and on the job. Industrial-grade hardware is what remains when the cutting edge has antiquated. When the market and technology is suitable and cost-effective, Altavian will be prepared and ready to put a Lidar system on our aircraft with expertise.