Webinar: Popular UAV LiDAR Use Cases

all right well we're gonna go ahead and get started if you if there's any audio problems please let us know in the feedback in the webinar chat and throughout the presentation there's a Q&A button that you can click to to let us know what you would like us to ask the panelists live and we'll answer as many as we can if we can't we'll we'll respond to all your questions at least afterwards through email so as an introduction my name is vu Winn I'm the director of marketing here at Phoenix lidar systems a lot of you may have heard of us and I hope so but just in case Phoenix lidar systems we specialize in compact customizable survey grade mapping systems so both mapping for lidar and photogrammetry and fully automated cloud lidar post-processing will touch on these a little bit later would you advance the next slide Conrad so first of all just a little bit of background why why are we doing this um it's because we started we launched the first commercial UAV lidar system way back in the day so in 2012 the guy who is our CEO Grayson tomans had this great idea which is can we combine lidar sensors with drones so that we can do really precise aerial surveying and bridge a gap that that exists with between human and lidar and an aerial lidar all these can work together and how do we get lidar to augment the strengths of photogrammetry which are which are great so we worked with he got in touch with Valentine and so our very first the industry's first commercial UAV lidar system featured a valid ein sensor anyways I just got a phone call but I won't rattle through all the rest of those points here but the the history of Phoenix just has gone through since then so five years real time point cloud VTOL fixed-wing UAV lidar system and we're really excited about our post processing platform lidar mil and the terrain following flight planner tool which which we just launched about a month ago so without further ado I'd like to introduce you to a few of our great team members first we have Conrad who's going to be leading the conversation along with IRA Conrad is our senior post-processing manager and IRAs one of our post processing engineers on the call also we have David our Director of Sales I'm sorry Eric our Director of Sales I'm sure I'll hear about that afterward and in David one of our regional sales managers so they'll be helping with a lot of the chat questions and the Q&A questions and so now I'd like to pass it over to Conrad everybody hope everyone's having a good day so far today we are going to go over three different data sets that illustrates popular uses of UAV lidar will touch on how these data sets provide a unique advantage compared to other remote sensing methods the first data set we're going to look at is a disaster response data set then we're gonna move on to a high precision infrastructure monitoring data set and finally take a look at a data set that was collected for a challenging topographic survey so first we'll take a look at a response to disaster hello all I right here wanted to walk you through this first data set response to disaster as you are all very familiar with Houston was recently devastated with a 125 billion dollar damage hurricane event major flooding made the city planners and emergency planners relook at the Bayou system as well as the impenetrable surfaces of the city itself and realized that their their high logic mapping and subsequent emergency planning needed a revamp so we were tasked to evaluate a small section of Bayou that was recently constructed at a golf course constructed near it so the riparian area and subsequent hydrologic modeling and flow was altered and thus their floodplain modeling and emergency planning needing to be updated as well since this was a relatively small section of Bayou traditional lidar was not necessarily going to be the best application for them the slow response time getting a plane or helicopter in the air requires a lot of FAA planning and other acquisition planning that can be streamlined through UAV use also the added risk in an emergency situation putting manned operations up and up in the air is counterproductive to this sort of response thus making UAV acquisition that much more valuable photogrammetry could possibly be used for this application but because we're looking at ground models and hydrologic modeling of this area any areas with vegetation would pose quite a problem as well as any water surfaces require quite a bit of a difficult processing with photogrammetry and both of which made photogrammetry and not the ideal method in which to update these floodplain models so you have a light art was a perfect application for this small area over four flights each about ten minutes long phoenix lidar systems surveyed this section of bayou and subsequent golf course and ended up with about a hundred and one acres all perfectly processed in which to feed straight into their existing hydrologic models without any added acquisition or added processing orthos were also or imagery was also taken during this acquisition via the same platform and same flights to produce north of mosaic and also allow for RGB extraction into the point cloud we utilized the Phoenix lidar Ranger system because it has very very good vegetation penetration a high resolution point cloud and very very accurate data sets I believe that we ended up with a two centimeter RMS ez absolute vertical accuracy on this project the outputs that came out from this project included a dense calibrated classified RGB extracted lidar point cloud along with that we produce seamless ortho mosaic water delineation for their hydrologic modeling one foot contours and a highly detailed canopy model to allow for their right period analysis as well through all of these data sets actionable data was given to the clients that they could directly feed into their hydrologic models and hit the ground running in a short period of period of time to update their bayou hydrologic models we are now gonna go into the actual data of this we have our own data viewers and pardon us if there's a little bit of lag we will try to move this slowly as depending on your connectivity you may see a little bit of a lag yeah we're loading this data through the web so you might see a little bit of tiling kind of as the data is loading but hopefully this should look smooth every one so you can see the colorization of this point cloud is half the RGB values mostly the intensity and saturation of the RGB along with an elevation ramp and here you can see very clear capture of the vegetation as well as the banks of the Bayou itself as this was a newly constructed golf course the the interplay between the Bayou Bank and the golf course itself was exceptionally important to capture and from this vantage you can see that there's a lot of detail along those banks yeah zoom out a little bit we'll be able to toggle off some of these layers this is a classified point cloud so we can easily turn off the vegetation layers and buildings and powerlines because those were also present in the data and I'll zoom out a little bit more to help illustrate this I'm gonna go to the materials and we can stretch this color ramp to help kind of show us the range of elevation values along the ground surface here so even though we had a highly vegetated data set the ground capture was very very good I think I believe that we had over 30 points per square meter just on the ground alone under vegetation upwards of 200 shot 200 points per square meter in exposed ground areas this type of density allows for extremely precise and accurate contours and other hydrologic data for input into their models and provide the client with the kind of confidence that they're looking for when it comes to saving lives and responding to emergency situations so we'll go back to the slideshow and take a look at some of the products that were created from this data set the client was very interested in the various applications of UAV lidar and photogrammetry and therefore a variety of products were created for this this way one data set could enter this the municipality of Houston and be disseminated into the various institutions that may want to investigate riparian analysis analysis perhaps even permitting of the golf course itself along with the absolute purpose of hydro hydrologic modeling and emergency response for future hurricane events on the top left you'll see the highest hit model this is the highest return in any given 1 foot cell for the entire data set in the bottom left of that image you can just barely make out a power line going across the Bayou as well as the majority of the data which is vegetation this allows extraction of treetop data canopy height canopy density and allows for any further studies to allow for a change analysis or comparison between data sets on the bottom left you'll see the digital elevation model this is the ground surface itself along with building footprints from the neighborhood nearby this Bayou this the combination of building footprints as well as the DM allows for very very accurate hydrologic modeling when it comes to impervious services and the interplay between hydrologic modeling and urban environments on the bottom right we have our one foot contour dataset of the entire area you can see the detail that was captured in this from very small drainages within the golf course to the natural drainages on the north side in the forested area this once again is exactly what hydrological model errs need to provide an actionable confident data set in which to make plans for emergency situations and on the top right is our seamless ortho mosaic that was utilized to extract RGB values into the point cloud and give a macro overview of the entire area makes the data a little bit more digestible when you're digging into the more nitty-gritty deliverables once again we captured over 750 points per square meter over these four flights over a course of 40 minutes 32 points per square meter under the ground came under the canopy itself with a sub 2 centimeter RMS ez absolute accuracy these sorts of stats allow our client to take this data without question and move forward with it and react in the speed that is necessary for these sorts of disasters with that said we will move on to our next case study you lidar has been used for infrastructure modeling and monitoring quite extensively traditionally from fixed-wing and helicopter applications this has been an industry that has been successful for quite a while but has led to certain holes in the industry certain data sets that were unable to capture in this case study exemplifies one of those Phoenix lidar system had a power company that needed to investigate the distribution lines within a neighborhood level anybody who is familiar with the utility industry knows that or even driving down the street knows that these small distribution lines going from power poles to individual households or even commercial units quickly become very small tiny wires that turn into quite a maze of utility management comparing traditional utilities surveying which often targets high kilovolt lines say 230 kilovolt lines floating 200 feet above the ground those are quite easily captured via traditional lidar but but these small distribution lines are very often missed in such datasets this power company happened to know this and wisely sought us out to utilize you guys lidar to find these distribution lines so once again traditional lidar was not nearly dense enough to capture these distribution lines also for a footprint such as a neighborhood level survey getting a plane in the air a whole crew an operator of pilot etc in the sky adds risk and adds costs is such a appropriately sized project UAV based photogrametry while has its certain applications utility monitoring is not one of those is not precise enough and photogrammetry 3d point clouds via photogrammetry often use or don't even capture power lines let alone distribution lines right the photogrammetry based point cloud could be extremely dense on hard surfaces but the little lines that are connecting these homes to the to the infrastructure most likely wouldn't get picked up you would need ridiculous image resolution to to collect those and that would be prohibitive in a urban neighborhood environment as well these distribution lines are extremely hard to survey for property rights and people want letting utilities onto their property so this allows both the large the overview of the neighborhood as well as the small details of each individual unit and property to be surveyed appropriately a Phoenix lidar Ranger system was utilized for this forty a turn three acre neighborhood you can see on the image on the right the flight lines via image of utilize two oblique RGB cameras in conjunction with the lidar in which to get oblique imagery that better better shows the vertical surfaces of such an urban neighborhood one RGB camera pointing straight down will often miss the side walls of buildings the profiles of trees and any other vertical surfaces that are very present in these urban environments this 40-acre flight took 10 minutes and ended with a dense calibrated RGB fused point cloud as well as a seamless ortho mosaic let's take a look at that data we've got that ready as well and our viewer so this data set is extremely dense and will have to be a little patient as we zoom in to get the full resolution of the lidar because there are so many points as we come in here we can see that the lines in the buildings are all showing up with pretty amazing detail we've got it colored based on the RGB values right now and the wires and towers are highlighted in white just to kind of make them pop and illustrate the data a little bit better as we kind of circle around if we focus on one of these towers we can see just how much detail there is in this point cloud compared to traditional lidar which often collects say twenty to a hundred points per line per span we see a nearly continuous point cloud for every single main line as well as distribution line in this data set even such details as individual resistors on insulators are visible transistor boxes and subsequent utilities on the ground are also easily captured both the RGB and lidar the detail that's captured in these small distribution lines weaving in and out bushes and through people's backyards is truly what is important to these these utility companies as the management of these small distribution lines quickly becomes a nightmare when you get beyond the neighborhood level this level of detail and the subsequent models that can be created from this dense of a point cloud allow for very actionable easily digested datasets for utility companies to then properly send out crews for proper trimming and utility management pls CAD models can easily be created from this data as attachment points pole stalks pole bottoms exact measurement of stag is all very calculated from this data set precisely what utility engineers need for proper management um like the other data set we just finished looking at this one is classified and so we can toggle off certain levels so I'll turn off all of the buildings and vegetation class I'm also going to switch the view from showing RGB to showing elevation values and then we can get a nice separation of our utility infrastructure from the rest of the above-ground objects this kind of gives us an indication of just how complex these these systems are so if we go back to our presentation we'll be able to take a look at some of the products that this lidar data set was able to well allowed us to create you can see from this detailed CAD model screenshot on the left that individual features such as utility boxes transistors attachment points even the poles holding the power the streetlights themselves were usually found in the point cloud and created into a model on the bottom right you can see an overview of the entire entire neighborhood and the connectivity of all of the distribution lines from their 65 kV mains all the way to the tiny little phone lines connecting individual buildings on the top right this is a screenshot from the actual point cloud itself showing the ground in comparison to the rest of the above-ground features this 610 point per square meter data set captured all of the distribution lines the client was expecting plus some it's amazing how many can be hidden in between trees and routed via their residence 125 points per square meter per flight line made CAD modeling easy and an actionable and confident this is these are the words that utility companies are looking for and the exact type of data these systems can deliver all right let's uh let's move forward to our third case study the third data set we're going to go over is from a challenging topographic survey an engineering company was assessing the development of a specific plot of land this plot of land was densely vegetated and in order to determine the cost of development they have to factor in how much ground is going to need to be moved we learned during this project that leftover soil after the I guess design of the neighborhood is very costly to remove and you know you can quantify that by how many dump trucks it takes to to remove all of this excess Earth and making an accurate prediction of that can really make or break a project and the the typical methodology of for this kind of a project is to send out a field crew and have them shoot topo points and use their survey to build a digital elevation model just based on gridded gridded survey points from the from a field crew in this particular case the area of interest was so densely vegetated that field crew couldn't really get through the brush below the canopy was just simply too dense and UAV lidar turned out to be the the appropriate and best solution for getting an accurate elevation model in this area for this project a Phoenix Ranger system was used lidar and high-resolution imagery were collected it was a fairly small area 60 acres and so it only took about nine minutes to do this survey it was flown at 60 meters above ground level you can see a kind of a fun image from the landing zone on the top right there we had a bunch of cow friends that were out in the field with us and just below that you can see a screen capture of our base station that we were controlling the lidar from and we're looking at a real-time point cloud that we were able to use to ensure that we were getting the proper coverage with with our lidar system we've got this data that we're gonna show you as well so let's toggle to that one so starting off you can see that our our point cloud has really nice RGB values that were extracted from the the photos that were Co acquired during the lidar acquisition the photo processing to create this RGB extraction is really automated the photos are directly geo-reference based on the trajectory which is the same trajectory that the lidar has so it's a pretty quick to get to this a really awesome data set I'm going to use one of our tools in the viewers toolbox to create a profile view so we can see kind of just what this what this vegetation layer looks like I'll just cut a cross-section kind of right right through here and I'm gonna maximize this high profile view so that we can all get a good look at what this data data set really entails through this 1 meter wide cross-section we can see that the the canopy up top is pretty continuous there's really not a view of the the ground surface at all through the upper canopy and then when you look a little closer at the the lower canopy you can see that the the bush is extremely dense going down below that we can see that despite those two thick layers of vegetation we still have quite a good coverage of points along the ground so this is really the the ultimate goal of this project was to get these ground points without having to you know take a machete and chop your way through this this vegetation we can also in this data set see I will toggle on and off the veg layers there aren't any buildings so we'll just turn off the veg and change the symbology to show us elevation and then do a little bit of stretching so we can kind of see the relief of the of the ground in this area so with with everything off except the ground well I've got the wires on there as well but with just the ground layer visible we can see that despite that that thick vegetation we've got a continuous dense detailed shot of the topography of this area so that that was really the ultimate goal of this this acquisition we were able to take this data and turn it into a couple of products that were of particular interest if you can see that grayscale digital elevation model image that that's really the main purpose of this and we can see some fine details in that that definitely wouldn't be visible without UAV lidar we can kind of see some some trails and some drainages going through there this high resolution allows for really accurate understanding of exactly how much ground is there and where it is overall below the canopy we were getting greater than 30 points per square meter on the ground surface so that that's pretty great or compared a gridded field crew out there surveying this that's pretty amazing detail and below that canopy luckily during this project we had ground checkpoints that were taken using a robotic total station and compared to those we were hitting three centimeters RMS easy which is insanely good that that added a level of confidence that makes this data pretty much the best solution for for this particular situation we've also got a little interesting thing to it to show here we noticed on Google Earth over the past well in about the span of a year the Google Earth imagery had been updated a couple of times and we could see the progress that was happening post post the lidar survey so that was really fun and exciting for us those were the three projects that we wanted to go over cool well thank you very much to to Conrad and to Ira for both presenting they're both are some of our post-processing engineers and thanks also to Eric for for answering a lot of questions in the background now as we as we move forward with these webinars we hope we're providing a lot of valuable data there are so many topics that we could cover I'm especially with five years of working in this industry so tell us what you would like to see so it could be UAV lidar use cases or it could be best practices on post-processing or how to how to set up your system so that you can avoid errors with lidar acquisition so Conrad where you go to the last slide or the next slide so with that go ahead and we'll sign off and please keep us in mind let us know your suggestions you can email me directly at marketing at Phoenix live.com that just goes to me and if you'd like connect with us on social media the handles all of those instagram twitter facebook youtube linkedin they're all phoenix light art and with that unless unless the team has anything else to say i want to say thank you again one last time to Conrad IRA Eric and David and thank you all for joining in it was our pleasure thanks everyone thanks for joining in everybody

Channel: Phoenix LiDAR

Views: 12,118

Keywords: uav, lidar, drone, surveying, mapping, photogrammetry, webinar, phoenix lidar, velodyne, post-processing

Id: -y6HkfUauzs

Length: 35min 23sec (2123 seconds)

Published: Thu Jan 25 2018