Point Cloud (LIDAR) Processing Demonstration

hello welcome today we will be talking about point cloud processing with eros imagine alright here we have our dots imagine we have the ribbon interface and if you look on the terrain tab we have the point cloud tools here you can find all the tools needed to process your point clouds so let's look at how to open a point cloud so one of the ways we can do this is just file open and just like any other file or raster a vector you can open a point cloud layer another really easy way is to simply right-click and select open point cloud layer and from here we can open any any file that has an L a s format which is pretty standard there are other formats that we do support so I want to go ahead and just select the dot L a s I'm going to select this point cloud right here open it up and you can see that it defaults to color coded by elevation now it is a point cloud so it doesn't display like a raster and if i zoom in you can see that we just have points and it may be a little difficult to tell what's going on but we can change that so when I open the point cloud data set you will notice that up here on the ribbon interface there is an orange highlighted tabs that make available point clouds tools so if I had opened a raster data set you would have raster tools available if I open the vector data set you would have vector tools available in this case we have point cloud tools you know one of the first things we might consider is how to display our point cloud data and over here on the far left we have color by elevation if I select this then we have other options and typically point clouds come with an intensity band and again the intensity is measures the strength of the return of each signal I mean zoom in here and you can see very clearly here's a residential area with some rare roads buildings and this must be tree canopy we can also display by return whether it's a first second or third return I'll go back to elevation and also in addition we can display in 3d there is a way to we can just by using my mouse wheel I can zoom in and I can rotate on its edge and one of the things here is to make the display a little bit easier to see we have this tool here that will increase or decrease the point size so if I increase the point size it will update in both the 2d and a 3d view one of the things you can do is you can also do a follow so anything I update over here will be that view will be updated in the 3d view and you can do a clip as well so if i zoom in and roam around you will see that it will just display a clipped area you zoom out over here a little bit and then anything we change over in the 2d view will also update in the 3d view see so let's get rid of the 3d view here I want to close now let's when we take a delivery from our vendor point cloud data it typically happens that we receive a number of LS files so when you download or copy those files to your local disk drive they may look something like this to deal with each one of these one at a time would take a long time to do that so inside of or das imagine on our terrain tab we have this tool called commands and what you can do is when you first take a shipment a delivery of your lidar data you might want to run it through something like this and I'll show you what I'm talking about so I have some raw data here I'm just going to grab my very first one for this data set I want to do a few things and I want to do several things to each one so I don't want to spend the time that it that it would take to do that for each la s file individually I want to create a batch process to process all of these at once so that I can start my project work so in this case we're going to compute pyramid layers and spatial index this will give the data set some varying levels of detail so that when you zoom in and out it'll make the rendering of the data a little smoother it's always a good idea to compute statistics so that if you need to do some type of analysis the statistics will already be there in all cases you definitely want to add horizontal projection you need to know where your data is on the Earth's surface so in this case we're going to identify the horizontal projection this is state plan so if I just hit my S key on my keyboard it'll take me down to where I need to go and I'm going to grab state plane and then if I hit my G key this is in Georgia and hit return I can identify my projection system now also I want to define the horizontal units in this case want to use u.s. survey feet now I have this set up for 1 if I want to apply the same settings for all of the files that have been delivered to me I can use the batch tool here we can just add the remaining files that we want to add I can run this now or I can wait till later and run it so I've already done this I don't have to really go through the process I'm going to open my file browser and you can see all the LEDs all the files that were extra files that were created so now I can just simply take my selection of ls files and drag them into my viewer and you'll see that they'll populate over here on the Left content minimize my file browser I will zoom out here so this is all the tiles that I have and I have about 17 here so I have all these into my viewer I just did a drag and drop so let's say for my project I need to merge these together and then maybe create a subset of that so I can start my work so to merge these together you can simply select all of the files in your contents menu over here underneath on your point cloud tab and tools you want to select merge and it will automatically populate all those into the merge tool so I have them all in here I can very easily just name a file and then hit OK I have already done that so I'm going to open up my merged tiles so we're open that again here we go so now I just have one file and this is all of my point cloud tiles merged together into one file one thing to note here when I cover the way the lidar is acquired you can see this in this data set we have a lot more dense point spacing here in the overlap areas from adjacent flight lines so the airplane flew this way north to south and I went to south to north so then the overlap Barry you're going to have a lot more dense point spacing so if I just zoom in to one of these real tight you can actually see the scan lines so you can see that these scan lines are this way and these scan lines are that way but here there you have double scan lines and then maybe if I increase the size you can see it better so this is a scan line so forth so I thought that would be good to show so next thing I want to do is I'm going to subset this I'm going to open up an acquire box and this just gives me a utility to draw a box from the point cloud suite I am going to go on they're tools and select subset the subset utility opens when I opened my Inquirer box a box appeared in my viewer and here it is here I'll move it so you can see it so if I want a subset an area let's say I have a project and I'm doing some you know watershed delineation or or I can just stretch this box out and I want to subset this area here in my subset tool when I hit from Inquirer box my my upper left X&Y and lower right X and Y's will update if you watch it there they go they're updated now I can specify my output file hit OK and it'll create a subset of this data set so I already have that just close all of that so let's open up my point cloud data set again see I put it in here so here we are we have our point cloud data set it's subsetted to our project area and just based on the elevation data I can see my drainage I want delineate or or whatever to help me out here I'm also going to right-click and hit open raster in this case I will be using an ECW that's a compressed format this is County ortho photography so I've compressed it into a 1 and then I've also subset it so I can do my project work cause I want to go ahead and select that it appears on top so it is over here is move it down let's go find an interesting area I like this area right here this will kind of help me get things set up so I have some buildings here let's see you want to select my LS file go to point cloud and I can make them thicker I can change it so that the intensity so I get a sense of you know everything looks good I'm going to do some work so let's say I am interested in vegetation well as we recall when I display it by returns anything after the first return is likely associated with vegetation so if I turn off my first returns you can see that all these points are very closely tied to vegetation you have a few in here you know bouncing off the sides of the houses but for the most part we have all of our points closely associated with vegetation if we zoom in here we can look at this house there's no points there so in this case if I wanted to just extract out these points because I know their vegetation I can do that I have my point cloud file here I can just go to the filter tool and from filter you can go over here to the return tab and say ok I want and I'm going to create a new point cloud and in that point cloud I want the points that are the last returns and the intermediate returns if you want to do it differently you can also select them like this - all the way up but this is easier someone just do it like this and there and you can just name your output file if you want to clip it further you can clip it further so I name my output file and just hit OK I could do that and I would have a point cloud file that just represents my vegetation so I know where my vegetation is now based on the lidar data another option is to turn this back over to elevation another option here we can use the classify tool and using the classify tool we can take your point cloud data and you can specify the parameters needed to classify both man-made features such as buildings and then vegetation and also from here there will be ground points so I can set these parameters if I don't know what something is I can consult the help page and this talks about the classified tool for point cloud and here you know when I first started working this I really didn't know what plane offset was so I could read about it or I didn't know what roughness was I can read about it and then the reality is that you would have to run a few processes to see how your point cloud data is going to behave I mean if you have really dense point cloud data it's gonna you're going to get some really good results if you have data that's not so dense then your the results just aren't going to be they're not going to be as good so when you talk to your vendor and their extract collecting point cloud data for you make sure you tell them that you want your point cloud to hit a dense you'll probably pay a little bit more for it but you're going to get a lot more for so I can run this and I've already done it for this data set and just to show you what the results would look like once you run your classification you can color by classification I can do my drop down here and I could show you the classes the ones that we're going to pay attention to here are the ground the different vegetations in the building so what I like to do after I do a classification is turn off my ground view it with the imagery and actually it looks pretty good I have red buildings green vegetation so this would be yet another way you could get your vegetation in your buildings now if I wanted to create a bare earth remember because lidar has the multiple returns it can map the surface of the ground through the tree canopy so I can create a bare earth from these points and you can notice that there are no points where there are houses and all these points in the tree canopy are actually on the ground and we'll talk about this in a future webcast on how to convert these ground points to an diem and then contour it and then also how to convert the tree points into a vector polygon representing tree canopy and also with buildings as well so stay tuned for that so this is classification a few other things to be aware of so you have some profiling tools if I wanted to select a polyline profile I could very quickly see come up here draw a line just to show you what it looks like and you'll have a couple displays over here that will have the the full range of your profile and then a side view and a front view so if I hit up here on the profile room you can move this box down your profile you can stop it at any time and if you want it to do measuring you can do that so I can measure the slope of the rooftop or how high the house is roofline is but you get a good sense of profiling so you have that capability another capability is you have a rectangle profile this is zoom back in here so another capability you have is a rectangle profile so I want to grab my rectangle profile I'm just going to draw a box around one of these houses here and click in here zoom in a little tighter just do this that makes some sense want to take a little cross-section of this roof so you can see that this roof actually has a couple other pitches on it the rooftop at the ground if I wanted to go in and measure this I could I can move my profile box around to the next house if I wanted to if I wanted to come in here and select some of these points that are classified as vegetation you can select those all right real quickly here I want to show you a few other things before we have to go one of the other neat tools on the terrain is this RGB in code you can take your existing point cloud data set your lis file specify an output and encode the RGB values from an overlapping image and I have one of those so let's go open point cloud RGB and here's our point cloud let's try to zoom into that same area so here's our point cloud if I wanted to bump this up so you can kind of see and you can tilt this 3d and so forth and roam around so that's RGB encoding the point cloud one of the other neat things is I will show quickly show this is we have our compression format and we're going to show this in our next webcast and how to create these so if I select this zoom in to the area the compression actually displays a lot better than the original lis and there's not a whole lot of rendering issues you can tilt this on your on your side and we get 3d zoom in with my mouse wheel move it Center click or center hold down and move it so you have that capability so this is RGB and coded point-cloud data now if you recall the semi Global matching the point cloud data from stereo we're going to cover this in a future webcast and I have just just so you have an example of what that looks like we're going to go to LS file this is some pretty neat stuff here so I want to zoom into these buildings and you can see how tightly how nice these look and this is again a point cloud RGB encoded and it comes from stereo and what you don't get from this that you get in lidar is the multiple returns so you're not going to get that bare earth through the tree canopy mapping but you do get a very dense point cloud that can be used for feature extraction and so forth that's all I had today so thank you for listening and watching I hope you have a nice day you

Channel: Hexagon Safety, Infrastructure & Geospatial

Views: 75,550

Keywords: point cloud, LIDAR, ERDAS, ERDAS IMAGINE, LPS, IMAGINE, Photogrammetry

Id: lOwWNV9rpcc

Length: 18min 47sec (1127 seconds)

Published: Thu Dec 17 2015