LiDAR Processing in Global Mapper

hello everyone welcome to the global mapper webinar for the month of December my name is tae-young's Application Specialist here at Global Blue Marble Geographics with me today is David McKittrick December indeed December means it's almost the end of 2014 almost 2015 who would have believed it thank you all for signing up today thank you for attending are the latest in the global mapper webinar series as you're seeing on the slide in front of you today's webinar is gonna be looking at light our light our processing in global mapper specifically as those of you who follow the development cycle of global mapper will be aware with put a lot of work into enhancing some of the lighter processing capabilities and it's been a while since we've actually talked about lidar in a formal way like this so we figured this month we'd set aside the webinar to talk about the entire process of working with lidar in global mapper from importing doing some cropping filtering perhaps some QA of your data and ultimately generating some files maybe around terrain files or even going as far as generating contour files I'm not sure full of chance to do that today given our time but we'll we'll talk about the entire process of working with lidar from importing lidar to generating data and hopefully eventually at the end of our session to exporting lidar prada lidar as well the session is being recorded and will be available on our website within a couple of days before we begin there are a few housekeeping issues to cover the webinar is in listen-only mode so in other words you can hear us but we can't hear you if you want to submit a question about today's topic you can use the questions tool on the GoToWebinar panel to the right of your screen Blue Marble product specialists will try to answer your question in written form as they're submitted and we'll also answer some questions verbally as we can fit squeeze in and our time limits and we anticipate there might be quite a few questions today we'll get a very large audience in today's webinar I do make one request and we try to that is to try to keep those questions relevant to what we're talking about I'm as Tim mentioned we have some product specialist sitting the sidelines eagerly waiting for your questions so antair will be monitoring those questions as they come in as well based on our time and based on availability will actually fire some of those questions out and will attempt to answer those questions verbally as they come in so the question answer session will be active during the presentation as you'll see on the slide in front of you we also have an email inbox those of you have attended previous webinars will be familiar with this email address this is our inbox for help as far as global mapper is concerned any questions you have the perhaps you're watching the recording of this if you have questions about any of the content or about global mapper in general feel free to use that inbox we monitor that inbox on a continual basis and try to respond as quickly as we can and we also have the forum I don't have information about the forum on this slide but at the end of today's presentation I'll show you the URL for the forum you can even get to the forum from the help menu within the application form is an excellent resource as well for finding out about the application and specifically today for finding out about some of the lidar processing functions so with that let's take a look at specifically what we're going to be covering today as I mentioned today's topic is lidar so we're gonna begin today by answering the question or by asking the question what is lidar we're gonna introduce the technology of lighter I'm not gonna go into too much depth as far as that's concerned but there may be some of you attending today who are not familiar with this technology it's very much an emerging technology so those of you here firmly knew we'll go over some of the basics we will introduce the liner module in global mapper those of you who have not seen the module or maybe you're not aware of the module we'll talk about the functionality that that brings to play in regard specifically to working with these point cloud data sets and if you do currently use the lighter module I'll also show you how you can activate that in your setup and try it out with a trial option you can try the lidar module or if you're interested in activating up permanently how you can register it and make it part of your global mapper setup when we get into actually looking at the application we're going to talk about some of the filtering and cropping options you have with if you're working with a raw point cloud you may need to pare it down to perhaps a geographical area or limit the geographic the elevation extent of some of the points in your point cloud or just simply create crude input just a sample perhaps just maybe a preview just a small subset of your data we'll talk about some of those options when we have our point cloud imported into the application we'll talk about some of the visualization options we can render your point cloud based on various characteristics that are inherently associated with each point things like intensity or classification and we'll go obviously go into more detail when we actually look at the those components of the software we also talked a little bit about rendering your point cloud in in 3d as a 3d model and also generating a profile a cross-sectional profile another way of visualizing your data so a lot of things to cover covering the basic visualization options some of the new tools we've introduced with the release of version 16 of global mapper include tools for automatically detecting certain types of points within a point cloud we're gonna look at some of those specifically we're gonna look at where's in which you can automatically identify points that are likely to be ground points very useful if you're obviously doing terrain modeling and also we'll look at examples of point clouds where we can identify buildings and identify vegetations and apply the appropriate classification to those so very very powerful functionality that has just recently been introduced and hopefully will give you some information on that it'll be towards the end of today's session I'm related to that previous function where we identify vegetation and buildings we'll also show you a tool that will physically extract those in vector form and allow you to delineate the outline of a building and generate a three-dimensional vector feature and allow you to model that building in a three-dimensional environment as well I should mention that this process is only this development process is ongoing and the extracting of vector features is something that we're looking at expanding significantly throughout the life history of version 16 and in be going into version 17 of global mapper you'll see some major enhancements as far as that capability is concerned obviously many users of lidar data are using it as a means to an end rather than a end of itself so one of those perhaps the most common use of using lidar is for terrain modeling generating a surface model of the ground basically and we'll talk about that it's something that's obviously not new to global mapper we've had these this function for many many years but in the context of looking at lidar we'll talk about generating a surface model from our lidar point cloud and no process in global mapper will be complete without addressing the price of exporting data light our data sets can be imported obviously we're going to talk about that at the start but it can also be exported so whatever processing you do whatever reclassification filtering or whatever analysis you have derived from your point cloud or whatever information have associated with your point cloud ultimately you can export that again perhaps making it available in another platform perhaps making it a file that you deliver to a client so importing and exporting r2 begins in the middle of us you will be a lot of the processing work as well so that is a plan of attack for today and obviously we have a lot to cover so we're gonna go ahead and start by at asking or answering I should say that question we'll start by asking and then we'll answer what is lidar and ter do you know what lidar is you had any experience working with lidar acronym as far as I know I think it's light detection something well you know there's a little discussion about that there is some comfort I guess controversy as to what the Accra the returns actually stand for that's actually my floor first bullet I was under the impression until fairly recently that lidar was an abbreviated form of light detection and ranging well some sources that I can't counted recently argued the point that is actually a combination of the words light and radar I don't think we should quibble over the definition look like it changes the basic function of the of the data format but it is a format that's been around for many decades basically as you will see it's a laser derived point data set typically airborne bring up my bullets as Abraham as I talk about them typically airborne collected that increasingly were finding a lot of these data sets are actually collected on the ground to rest real models derived from lidar becoming increasingly common increasingly available in fact a few weeks ago right outside our building here in Halliwell we had a group of technicians actually collecting local lidar data from the ground on the ground so we're very eagerly looking forward to actually getting our hands on some of that data so we can see our own building here as a three-dimensional point cloud so it's a laser based system laser based remote technology system it's typically airborne most of the data we're gonna be working with today has been collected from an airborne platform typically a fixed-wing aircraft flying over an area transmitting a laser pulse or we see a series of laser pulses and determining what the nature of that surface is that laser but that laser pulse reflects off reflected laser pulses are used to calculate distance ultimately and this obviously depends on some very precise calibrated instrumentation on the aircraft that is able to calculate that precise distance and indeed the elevation of the aircraft so based on those variables lidar ultimately comes up with a series of XY and Z or x y&z points that position each point in s in a space a three-dimensional space points are obviously only relevant if they're processed as a collective in a collective form a single point obvious isn't that much of a much value and lidar you will find as we get into looking at some samples typically is comprised of very closely spaced points and the volume of points covering a certain area can get up into the millions tens of millions hundreds of millions and sometimes even billions of points obviously the higher the number of points and the closer the spacing of those points the more precise the modeling process will be but when it comes down to the very basics it's a very simple to understand format it's simply XY and Z value associated with a point now additional information may be associated with a point and we'll see that when we get into looking at global mapper as well where each point may and typically does have a classification in other words the nature of the surface that was first of all detected by the laser was reflected from that laser pulse I can be determined the intensity of the return will be a reflection on the type of surface and return number a given laser pulse may be returned several times if it's able to penetrate through one surface and hit another surface and that would typically be a leaf where a leaf obviously has a lot well a lot of certain amount of penetration of that laser transmitted laser pulse so we'll get multiple returns again we'll see that when we actually look at a point cloud lidar is the raw material for many 3d analysis processes it's typically not an end in err of itself a loanword casing you find a lot of people see the liner model itself as being a viable way of representing features of three-dimensional features but it's more typical of people using light ours means as I mentioned before the perhaps to crave train service and from that terrain surface modeling process perhaps to generate contours do some volume calculation perhaps even do run one of the analysis for instance water see an analysis or view share analysis so lidar is certainly a very important stepping stone into getting into those analysis tools and we will as I said be covering some of those as we go through today so with that let's take a look at the application itself nice you can see I've got several windows open here first one I'm gonna bring up is actually a blank Google mapper window and and I want to point out before we actually start looking at any light our point because I want to point out the lidar module right here my toolbar I'm just dragging down the module toolbar this is what you'll see if you activate the module is this new toolbar and to make a little more clear and more visible I'll drag it down into a introductory screen here this toolbar provides many advanced functions that you can use to process your lidar data and we will get into looking at many of these throughout the course of today's presentation for instance we can change the visualization based on whatever characteristic that we want to focus on I mentioned before things like intensity or classification and we'll look at those in just a little bit so this tool part gives you a very quick and easy way to change how the lidar point cloud is being represented on your screen what data what component of the data is actually being reflected in your point cloud all other buttons allow you to automate or just simplify I should say the classification so if we want to select a group of points or select you know collection points within an area and render them as a particular class you'll get some very easy buttons for doing that right here in the module toolbar we then have a number of very powerful functions as I mentioned during my introduction for doing things like auto classifying ground points Auto classifying non ground points specifically there's going to be buildings and trees for feature extraction in other words for actually vectorizing features based on those points building footprints for instance can be vectorized as three-dimensional models we have filtering tools we have a tool for filtering your point cloud and we've told for our applying colors we'll get into looking at these a little bit later so then this toolbar will not be available out of the box in the standard virzal version of global mapper but it's certainly something that you can activate from the help menu if we go to the module extensions that extension license manager you will see on the left-hand side our built-in modules list and your setup will include the lidar module if you have not activated the checkbox will not be there obviously I have but if you want to activate the lidar module simply check the box it will prompt you with a registration screen and as with any of the registration options in global mapper you can either enter an order number I'll point to a license file of you if you've been provided with one or if you have not actually purchased the module you can enter you can request rather a 7-day trial and it will activate all of the function of I'm going to show today for that limited duration allowing you to test it out obviously if at some stage you want to purchase the lighter module on a permanent basis again it's simply a case of going into that dialog box at registration dialog box right here it's a greyed out box in my case register the module and it will be active for the duration of our firfer permanently it'll for the duration followed as long as you use the software so again write our module is going to be used today during some of our scenarios it is not required for using lidar data you can perform certain basic processes within global mapper without the liner module importing a little bit of filtering generating terrain surfaces you can do that within global mapper without activating the module and hopefully you'll see today that activation of that module provides a lot more advanced functionality and so I want to try to make sure that as we go through certain scenarios today we address those components that require the module and those components are available without the use of the module the first thing that we're gonna do the simple act of importing a point cloud does not require the modulus you want to pull this up the way just a little bit towards the top of my window here I'm going to go through a simple import process to bring in a sample point cloud as with every file format and Google map or the way to get data in it's very simple obviously the file menu open data file there's a folder button in my toolbar that will do the same thing right here in the middle of the introductory screen is our open data files button again same end result you want to get the standard import dialog box now global mapper supports many different point plat formats or different data sets that may include point x y&z data archive and Z data the common ones inland when using lidar r dot la s and dot l AZ or la Z files this is a compressed lidar file I'm just going to point to this file as an example and we're going to go ahead and bring that in now you'll notice there's no print no specific entry point for lidar files it comes in the same front door as every other format that's within the data don't have to predefined them to feature the file type before you import it's just a standard import function what is different about lidar is the dialog box that will appear after you initiate the import because within this dialog box you can perform certain filters certain you know apply certain constraints to the data based on what you know about your point cloud now before I get into showing you some of the examples of what we have here let me stress that in a typical situation you'll probably not want to filter your lidar point cloud during import we can obviously do that your after it's been imported but during import I would typically say that most people would not know enough about the data to make those decisions in other words do you know that your data specifically has two different types of classes do you know that information enough to be able to make the decision to filter out what you don't need at this stage I would suggest especially if you're working with data you've never seen before go ahead and import everything but you do have options and during this import if you need or if you want to pre-filter that data a filtering can be done based on class you'll notice a lot of these classes fairly self-explanatory unclassified is something we want to address at length when we actually looking at some of our data ground points probably the most common classification that people are interested in and then we have others vegetation building etc etc etcetera and this list goes on and on and on it's a very long list now the availability of these classes in your data set is obviously going to reflect where your data's come from Morpheus process to your data but you will find that most data will have at least a certain number of these classes we assigned if it doesn't don't worry we're going to address that towards the end if your data is entirely unclassified you don't know it can't distinguish your building from a tree from the ground we'll address that a little bit later we can also filter by return type your turn types reflect the number of reflected signals for every laser pulse if there's more than one you can filter those on it as necessary we can also filter geographically based on whatever bounds we specify and this bounding box is the same bounding box that you'll see in many different components of global map orbit allowing you to filter again based on whatever Geographic constraints probably the most common of which would be to crop to a selected area perhaps this is going to reflect a jurisdictional boundary and obviously you can limit just the point cloud rendering to what falls within that boundary if that's what's relevant to you and we can also preview to create a subset of our data delete files that are W points I'm sorry there are over a certain number of standard deviations and you can specify that number this all allow you to remove those obviously erroneous points that are going to really extend the elevation extent but are obviously get completely incorrect so you can filter it down to a known range if unlikely and if you have data of a lidar data that reflects depth on the lowest bathymetric lidar data which is becoming increasingly common the values themselves may be positive global mapper will see those as positive and assume that they're above the surface but you can treat those as negatives by simply checking this box so those increased values will be depths rather than Heights in that case so against just on some of these filtering options we have available during import another option you'll see is to go ahead and immediately create an elevation grid by passing the rendering of the point cloud and create a terrain surface or a a raster surface model again on a typical scenario I would not suggest I'm not recommend doing this I would suggest you actually look at the point cloud first before you actually import the data but it is an option during the import process and the elevation grid options here would reflect that selection so in our case we're just simply going to import our point cloud and this example is a relatively small area it's just up the road from where we're sitting today in Augusta about three or four miles up the road this very distinct structure you see in the middle is actually our state capital we may showing you this in previous webinars and we're gonna be looking a little bit at this point cloud and some of the information that could be derived from this point cloud today is there a size limit for files that can be brought in to global mapper for later that's a very good question and one that we do get a lot - yeah that's that's I'm not surprised that one came in early you'll notice this point cloud by the way is eight hundred and seventy thousand just over eight hundred seventy thousand points we're gonna look in a little more detail at some of the information that we know about this point cloud but off the bat we can see it's under a million points this is actually a very small point cloud and we use this small one obviously because we want to you know see some of the processes working very quickly we have experimented and worked with point cloud data up to a billion and actually over a billion points this is a very much a moving target we're encountering users who are wanting to process five billion ten billion points at a time global mapper has no physical constraint but when you get into point clouds of that size you're gonna find that it's a hardware issue rather than this the capability of software but there is no defined upper limit in terms of the number of points in the global mapper can process so and it is something where in terms of the processing speed obviously the larger the files and more intensive the processing requirements we are making sure we work on that and making sure global mapper is optimized for working with those large point clouds so that was just a simple import process very straightforward I should also address very quickly I'm not gonna spend a lot of time on this but we we kind of make the assumption that you have access to data you know I obviously have a sample file here I was able to import it but very often the question that pre precedes this is where I get lidar and we've actually had that question come in quite a lot even recently working a particular area we've introduced the idea of working with point clouds but you know if there's nothing available in your area obviously you don't have access to the tools so the question comes up where do I get lidar well obviously there may be data available local to you certain states have archives of data that you can actually download directly also it's worth considering a worth noting that in blue and global mappers online data lists where we essentially stream data and you may have used this for streaming imagery for instance but there is a lidar section in here now this lidar section provides a number of links should stress that these are not streaming services per se we've got a NOAA link here with the USGS link which is a very useful one we've also got our open topography lidar portal site which is a great source that it's a collaborative effort to pull together data from multiple sources these actually are not streaming services but rather what will happen if I select this is I will open up a web browser go to that site and explore what data sets may be available and so there are some links in here that will provide at least a starting point for trying to find some data for an area that you you're working in a good source as well again it's calling your local GIS office and asking the question they may need they may know of a collection process that's going on and may be able to direct you to a good source so but I just want to address the fact that we do have some links in there in regard to accessing liner now importing our point cloud very simple process I'm going to talk about visualization in just one minute I also want to very quickly bring up another possible source for lighter and this is being a little creative in our processing here I'm going to use the standard control you to remove that file I just import it I'm going to initiate another open and this time I'm going to point to a text file this is a simple X Y Z text file I'm not going to protect the time to preview it take my word for it it's X Y Z and Z's and a simple comma delimited format and when I click open on this this is going to bring up our standard ASCII import dialog box this is how you'll typically import text for creating point array or creating lines or even creating areas from those points one of the options that you'll notice here is to automatically take those points and generate a point cloud so as well as defining the characteristics of the file itself you can have global mapper believe that it's actually a lidar file now that will allow you to process it in the same way you would process the standard lidar data that we're going to be working with today so your source data when working with Leiter may simply be an X Y and and Zed file I'm not actually going to bring this in but it is an option that you know if you have again very simple XYZ files you can use this ASCII import and by checking this option in the import box you can build a fake global mapper into believing it's actually on light our point cloud so let's go ahead and open up a workspace now this workspaces got the same data that we looked at just a minute ago except that in this case I've actually brought in another layer you'll see I have some underlying imagery as well we're gonna bring that into play in just a little bit we've got some tiles of ortho imagery what I want to talk about now is point cloud visualization this is one of the bullets in my my introduction the initial display of your point cloud it's a fairly unique format in this regard it is a vector format but the initial display of your point cloud reflects the inherent elevation value that's associated with each point if I quickly choose my feature info button and I select a point at random and just bring up the standard feature info dialog box we adjust its size so we can see little more clearly you'll see for that specific point you'll see a number of variables that have been associated the elevation right down to centimeter level this is a meter value so point four one meters obviously down to the centimeter level in terms of precision we also have intensity we have scan angle we have classification what is that point well in that case that point is unclassified and a number of other pieces of information are also available there the initial display by default of the data reflects this elevation variable and the display actually the actual color is associated with whatever the current shader that has been selected in my elevation drop down list here if I wanted to change that to for instance global shader you'll see that will automatically reflected as well this is a different shade or pattern obviously within the context of this area there's not a lot of elevation range so I'm just seeing basically one color so whatever shader pattern has been selected here will be reflected or by default in terms of the display and it was dynamic it can be changed you can apply your own custom shader if necessary no but you can build a shader pattern or flex colors that you want to apply to certain elevation ranges that's an option that's available in the configuration dialog box if you want to perform that one now as I mentioned previously there are options in this drop-down list that will change the visualization from elevation to some other variable intensity intensity is an interesting one if you want to zoom in just a little closer here because intensity depending on the zoom level think this is probably a good zoom level to look at this actually looks like a black-and-white photograph the darker colors are in this case of reflecting vegetation the lighter colors grounds and you can see it almost looks like a image if you like it is derived from closely spaced points but it looks like a black-and-white image and one of the options we'll just introduced with the lidar module as actually to create a surface based on this intensity value rather than simply basing on the elevation you can base it on intensity value so again just another visualization option I'll point to a couple more I'm going to take the time to go through each one of these return the return number is going to reflect in a color pattern the return the which how many returns were generated from one laser pulse in this case the blue indicates that there was one return or either that or it was the first return the red indicates second return and if we go in closer you'll see some other colors starting to appear here you'll see yellows as well so this color pattern is a great indication of where there's likely to be vegetation because these vegetation typically is going to correlate with where there's multiple returns from one pulse obviously if a pulse it's a solid surface it's likely it's only gonna have one return if it's a a a vegetative surface and leaf for instance it's likely it may penetrate and that's what we're seeing here in this in this color pattern you may also notice at the top of this list we have color by RGB elevation and that's actually where we started the default the appearance if we choose this option is to have the actual colors of the points themselves reflected well obviously in this case our point cloud doesn't have any RGB values we don't have any colors associated with these points but one of the things we can do with the global mapper module and which is the final button over here in the toolbar is actually apply a color from whatever raster layer you happen to have displayed which in my case is this ortho photo we can actually apply the color of the pixel underneath each point to that particular point and it's simply clicking the button that does that that's going to go through each of the 807 points and initially it appears that the point cloud has disappeared but it actually is now rendered with the colors that are associated with that image and easiest way to visualize that is actually to look at the points themselves again now you'll see the reflect they were points are now reflecting the actual colors the most common use of this is obviously gonna be with the use of aerial imagery but you can be very creative in this process you can apply colors from for instance a topographic map or maybe it's a map that you scanned that reflects some of the data that you're working with perhaps a pipeline map those pipeline maps whatever colors are associate them can then be assigned associated with your lidar point cloud so again you can be very creative in this process now at this stage is probably a good a good time to take a look at another visualization option with our point cloud to this point we've been looking at our point cloud from a top-down perspective we can now activate our 3d window and obviously this is not a new component of global mapper but it certainly takes on a considerable amount of significance when we come to actually working with these two three-dimensional point cloud data sets those of you who did not see are what's new in version 16 webinar I think over the last month we did that or the month before I guess we introduced the sky box which is this kind of realistic backdrop to our 3d view now we can see our point cloud as a 3d layer as H point obviously suspended and it's a three dimensional plane and don't we see the skybox indicating it's a very sunny day behind the scenes so this is a great way to visualize your point cloud especially knowing that we have associated each point with the appropriate RGB it looks like a photorealistic model obviously whatever selection you make it's important to know what's being represented and that's very true when they choose this option which I've bypassed which is classification classification is going to reflect what each point actually represents and this is perhaps one of the more important visualization options from this list in the case of this particular point cloud we only have two classifications we've got gray points and we've got brown points in order to help us determine what those points are we're going to actually activate the legend and legend is triggered by the over the map layout button which is fourth button and in the toolbar and we're going to choose to display a legend based on are loaded vector types now we could customize the legend with whatever settings that we want in here changed its position etc I'm just going to go with the defaults and click OK and you're gonna find its displayed right down here in the corner and as you can see confirming what I said previously we have two types of points in this layer ground points and on classified points is it possible to change the color of the classes in this you don't like my brown points I'm saying that the grass is usually it's sort of greenish so it might be nice to have the good point good point well-taken yes the answer to the question is absolutely you kind obviously the the initial Association we try to make it as relevant as possible brown dirt ground grey we don't really know what it is so that's kind of left as an unclassified but yeah the colors can be assigned based on whatever whatever colors you want and that's actually triggered in the configuration dialog box from my toolbar these are points so we're gonna go under point styles and if you scroll through this list of the currently available Styles you will encounter lidar ground shots unclassified etc and it's just a case of defining the colors right down here at the bottom you could you can make them whatever shape whatever size you want so the sky is the limit quite literally as far as the visibility is concerned so yeah that is certainly an option now we are going to talk a little bit more specifically with this dataset about the idea of this data being unclassified and I think actually may actually bring in a different dataset for some of our scenarios as well obviously whoever collected this particular point cloud was mostly concerned with ground that's why that's the only visible classification buildings trees in other words everything that's not a ground point it's through under the unclassified pile and if I was intending to use this lidar point cloud for instance for vegetation analysis or perhaps we're doing some building building modeling I would have to do some processing on this data and that's what I'm going to introduce a little bit later a very very powerful tool that we now have access to that will identify where there are buildings and where there are trees etcetera we'll get to that just a little bit later before we get there a couple of other things to show you by way of some of the initial visibility options this again I have to stress this is only available if you have the lidar module the tool that I'm about to show you is in the standard version of global mapper but it only will work with a point cloud if you have the module activated and that is my profile tool right here in the tool bar 3d path profile you can see the text hovering over my cursor if I select that option I can generate a cross sectional view of my point cloud I'm looking at my data and from a lateral perspective now this again is not a new tool in global mapper it's being it has been available for many generations essentially it creates a cutaway view in the context of this data the fact that these are points we need to do a little more with just the profiling tool rather than defining the profile as a line because obviously if we drew a line through through a certain area it's very unlikely actually would intersect a lot of points so what we need to do is define a a buffer if you like within which the points will be displayed so the first thing you can do when you activate this tool or when you're ready to create a profile view and I want to initiate this through a right-click just simply right-click on the map is I can specify the corridor within which the points will appear it brings up a dialog box lets you put in a unit of measure just for the sake of it I'm gonna bring this down to about 25 meters now this is a buffer distance so this is gonna be either side of the buffer law or the profile line then I'm gonna create and you'll see this in just a second how this is manifest so if I begin the process of creating a cross sectional view I'm simply using my left mouse button and then when I'm done I use my right mouse button again you'll notice I've created a line that the yellow object is the linear path of my profile the pinkish area defines the extent within which the points will be displayed so in other words it's not just along the line but either side of the line and to finish the line as with any digitizing function in global mapper it's simply a right-click option and as you'll see now we're looking at that point cloud from the side if you like cutaway view lateral perspective and the fact that our buffer area encompasses the full extent of the capital building in this case we can obviously see that displayed nicely in this view come back and look at this in a different context a little bit later but initially for initial visibility it's a great tool again isolating date points in that a vertical plane if you like okay couple other things I want to look at but while we have this point cloud visible I'm looking at my notes you know obviously the visualization of the point cloud will help you to discern patterns in the data we can see through the selection of classification we can see that there are ground points and they're unclassified points but if we want a little more detail if we want information about how many points what's the distribution of points within my data what we're gonna talk about in are some of the analysis functions that we can apply to a point cloud if I select the point cloud layer and my overlay control center and I can clump metadata button and metadata that will list a number of characteristics for this file and the number of points this is confirming what we saw in the overlay control center they might have memory that it's using come to a point cloud memory this is a very important one the density of the points and this is an averaged value so it's just over two and a half points for every square meter from this point cloud the bounding boxes are also noted here in various formats and also the projection information is nearly at the bottom you'll notice as well we haven't been max elevation we have a number of points you're obviously that there must be some problem there and buying these or below sea level will address that in just a little bit but you can see a summary of the point cloud right next door to my metadata tab there's a statistics tab and this is extremely useful because this provides a more broad analysis of the contents of this file we talked about returns numbers we can now do a a quick cursory glance at which points our first point returns which our second first of many etc and the percentage breakdown of each so of the all returns are a hundred percent those that are considered first while 80 cent 86 percent of my points are first returns obviously that leaves what 14 my math right there 1014 that would be a second third or fourth um in other words a second return so again it's just quick visual analysis of those percent break the attribute values and min/max within each attribute is also noted here perhaps most useful and most important are the classifications that are associated with each point if you look at the percentage breakdown almost half of my points actually over half of my points I should say are unclassified I see that as a problem in this case because the half of the points half of those 870,000 points are completely useless in their current form and we would need to address that as far as making the data more useful as concern so metadata will give you the information necessary to see what classes are available in your point cloud to see the distribution of those that will allow you then to perform the filters that we encountered first during the import process we can now if necessary we re address that in filtering process using another button which is in our lidar module tool bar and again we've got this we have now have the same filtering options now that I know have ground and unclassified points I can simply disable all and perhaps enable my ground points or whatever combination of classes we want to display or we want to work with so now we have that information where we feel better qualified better informed to make those decisions the time being we'll just keep all of those all of those on call of other options to look for and options being the operative word we're going to take a look at the options for a point cloud and we can specify override the default visual characteristics and specify for this particular layer a shading pattern if necessary so that's over and above the the module the toolbar setting we can specify for in individual layer and this will be useful if you've got multiple point cloud layers in the overlay control center and you want to distinguish them in different ways you have that option at the layer level redefine the unit's meters is the default in this case we can alter our elevation values we're getting into a little more of the editing capability which I'm going to touch on own next but we can specify scale factor power or offset to adjust the point cloud elevations themselves universally or filter the point cloud based on known constraints so for instance if I knew that my negative numbers are obviously incorrect I can specify a cutoff in this case of zero which will remove those that are negative so we can define very precisely here the maximum values I'll show you in a little bit how we can go into even more depth as far as those elevation filters are concerned but at this level very simple process just to define the extent those points that are not within the range that I define will be removed or there will be rendered non active any process I applied will not be applied to those points next one I'm going to show you again another new tool in global mapper but it comes in two it has a certain amount of significance when working with point clouds and that's our search function I'm triggered with the binoculars tool in the tool bar if I click on the search dialog box a couple of things that this this will allow me to do first of all it will allow me to visualize the data the point cloud data in a tabular context and you can see now all 870,000 points are now listed in a table view we can also query at this date and I'm not actually gonna do this right now in the interest of time but it's fairly intuitive any of the available attributes elevation perhaps being the most common we can define and build a query based on whatever operator we want on whatever numeric value we want as well and in that case will be a numeric value so you can build queries a building new search and then within those results create a further level filtering so that the search dialog box has search your query function within global mapper certainly has its relevance when working with the tabular data that's associated with our point cloud I think perhaps even more useful function in this dialog box is simply the ability to sort I've selected the elevation column here and because it's almost a million points it might take a few seconds here to respond but when it does and when it refreshes I will have my point cloud ordered essentially in terms of elevation I'm confirming what I suspected there are a number of points here that are in negative territory I would be very confident in simply selecting my hell down my shift button and using that delete selected at the bottom to remove those points so this is a manual visual cue a process where we can go down the list and see you know what's likely to be an erroneous point and physically remove that based on simply sorting in this case going to the other extreme again we can see the upper extent and again if there's a point there that's way beyond our expected level as far as this area is concerned very easy simply select it and delete and at that stage I mean simply using global mapper for processing lidar in that way doing a little bit of filtering in this way you could quickly turn around and export those files if that's your workflow the importing glass exporting glass is very very simple oh I should have also mentioned attic we had a conversation recently with somebody who actually had done this but that tabular view of the data also allows exporter basically you can copy a group of points if you want to filter the points based on a certain area certain elevation range right click copy to clipboard and you can paste those into Excel if you want to generate a report and granted some of those the length the the amount of points is will be very large you might want to be careful working with that volume of data but it is possible to essentially export or capture that data in a tabular form and have that as a report or generate a spreadsheet from that data another function that's activated only with the lidar module and after this I'm actually gonna select a group of points is the ability to perform more manual filters again based on what has been selected we could have selected all of the points in the point cloud for this purpose as well or you could constrain the selection to the extent of a geographic area that's another option but if I right-click after some light our points have been selected and by the way I should mention the fact that this is a red box is merely an indication that we have lots of little red outlines indicating lots of little points have been selected in fact you can see at the bottom of my my task my status bar bottom of the screen below my scale bar down here 349 points selected within that area but the tool I want to show you is actually under advanced selection options this is again only available if the module act is active in your setup of global mapper 16 filter selected lidar points by you'll notice a number of options available in here again based within our selected area we can filter based on elevation range on the elevation ranges noted here we represent the minimum of a maximum within the extent of what we've selected and we can filter by color we know there's RGB s now associated with these points if we want to limit those to a pretty go to color range we can do that as well and defining the fuzziness value in other words how close to the selected color do they need to be scan angle or classification once again so it's gonna for the purpose of illustration there's not a great deal of elevation range here so I'll have to be very careful with what I do but I'm gonna change this minimum range from 38 point one four two let's say 39 and hopefully the results of this as you will see will be a slightly smaller subset of my my data is now selected now again all we're doing here is filtering this selection we haven't actually removed the points that we don't need we've just merely a lot that merely use this for a filtered selection group with points selected there's a lot we can do with those now we can edit them right click and edit and all of those remaining points we can assign whatever classification we want to those points we can apply an elevation if that's applicable I'm not sure that that would necessarily be a good idea because you're overriding the default flag them with one of the available flags or we might just simply copy and paste them into a new layer if that's relevant control see you can't see what I'm doing on our keyboard ctrl C control V and then I can paste them into a new layer creating a subset of my overall point cloud based on the filters that I've applied okay next one is an interesting one as well again we're dealing with a little bit more of the analysis of our point cloud obviously from a kind of a wide-area perspective it would appear that my point cloud is fairly evenly distributed except for the areas over here towards the water this is actually the river but if I wanted to get an idea as to the level of clustering of my points we have a tool and global mapper that will allow you to do this and again I should stress this is not a new tool this is a tool that we've actually demonstrated in different contexts but with point clouds that comes in again it has a relevance for working with this data and what I've done is open the overlay control center I've right clicked on my lidar layer my augusta lidar point cloud and you'll notice the option to create a density grid this is essentially a heat map tool can be used with any data point our typical more typically the data format will be used if I select this option I'm just gonna go with the default options here and what is happening is it's building a grid now I should stress while the status bar is moving along here this is not an elevation good we're not defining height values within our point cloud that what we're doing is we're mapping the relative clustering of points within this point cloud easiest way to see this more clearly is to turn off the original point cloud and what's left is a heat map indicating the number of points per square meter you'll see going from zero around the periphery up through one two by seven points per square meter so this is a a clustering or heat map so we're getting a completely different perspective on our point cloud now you'll notice a heavy concentration of points within this area right on my point cloud again difficult to see but that actually corresponds if we go a little closer I think I need to use my zoom tool with a very defined band through the middle this is actually an overlap band between two flight lines so you'll see there actually is a heavier concentration that corresponds with that overlap and we can see that very clearly here also kind of a peak here this actually corresponds with weathers vegetation again we know that there's multiple returns from leaves etc so in this case we have a heavier concentration where there is a combination of the overlap and vegetation so another way of visualizing your data final thing I'm going to do as far as analysis is concerned I'm just going to unload this data control you as our keyboard shortcut I'm going to bring up another workspace and this is a very small point cloud this might represent a small area perhaps a field you know a localized area I'm gonna change the classification or GPS um we've imported a point cloud the number of the variables that we saw with this point cloud include as we looked at the metadata we can see the number of points the average concentration of the points across the point cloud the bounding coordinates what if you wanted to determine the precise area that's covered by our point cloud this is another function it's not new in global mapper but again has applicability for working with lidar again it's a white click option on the overlay control center and the B boxes or coverages is the option you're gonna choose and if I choose that I have the option to choose a rectangular area in other words I'm MBR for this point cloud or a polygon area and I'm going to choose no and also decide how course or how rough my I'm gonna create the speech in other words do I want to cut the corners or to want to match precisely where these point is in the bounds of my point cloud I'm gonna keep it rough which will give me a more precise measurement and we'll click OK now it's not initially visible but if I turn off my point cloud you'll now notice I have a bounding box it defines the interpolated extent of the point cloud and if I use my feature info button and select this area feature that's been created you'll notice I get all of the information pertaining to that area including the actual physical bounds you'll see the area and square metres because that's how my system is configured to be square feet if that's the case you can see very clearly how much area we now have covered by this little sample point card well talk a little bit now about some of the editing functions we've looked at the analysis we've looked at how can we determine the characteristics of our point cloud visually how can we see how many points are in there the distribution of points well talk about editing the point cloud now move it to this just a little bit unload this data once again I'm gonna bring in the same file were looking at previously again this is to write from a workspace so let's bring in the workspace file now I showed you this very briefly in passing but editing might begin at the individual point level select a point with my digitizer and right click and edit and you have the option to change the classification for that selected point neither write like the hood of you performing edits at the point level within the point cloud is pretty minimal a lot of points to try to indeed the word 800 so what are you doing for the rest afternoon to maybe she was busy doing some editing your view point how many points per second well I'm still here no I keep you busy but yeah certainly doing individual level not very practical but is certainly doable so more likely you're gonna do this based on either selected group of points or perhaps based on a query that you have run to filter out certain points again bringing up my search dialog box X actually gonna display all of the data but we can then synchronize so our selection is Karl correlates with what we've selected on the map and then we can search within those selected features and perhaps delete perhaps edit whatever so again selecting a point or more than one point in our search dialog box again brings up the same dialog box so we have the ability at the point level or collective point level to make modifications to our point cloud what's probably more useful is the ability to perform more automated - classification processes and this is where we're really going to bring into place some of these advanced tools that are available in the toolbar I'm going to visualize my data by classification once again and once again we're gonna focus in a zoom in a little closer it'll become clear we're going to look at this overlap band here now in the case of this particular flight this lidar flight you'll see that these overlap points there was a number of them were thrown into the unclassified bidding if you like in other words they're not useful at all we've got a lot of points that obviously represent ground we'll get some that represent vegetation and buildings but we're we're not confident enough at this stage to know what they are global mapper can apply a very precise mathematical computation to determine which of these points are likely to be grown points and I'm gonna initiate that process right now simply by clicking the auto classify ground points button in my toolbar and by the way my data to begin with could have been entirely unclassified I could have started with no ground points at all and global mapper would have identified patterns within your point cloud planes in other words that represent possible surface that are not elevated those typically be buildings and based on that algorithm would assume that those our points are our ground points I'm sorry but we'll go with what we have as far as ground points are concerned now I'll go with the defaults here we can basically define our sampling area and the offset from a known ground point that's likely to be another ground point in other words we're dealing with a three Mueller threshold within our defined area those points that are within parks are 0.3 meters I'm sorry point 3 meters we're gonna assume they're also ground points you can be more precise by bringing this number down or be a little more general or a little more coarse I guess by increasing this value simply clicking okay at this stage hopefully we'll show you the results quite clearly on this example that I'm showing you and again I should stress it this could have been performed against a point cloud that was completely unclassified to begin with so going through analyzing all of our points and because of our visualization they automatic the colors automatically change you'll notice it's retained quite a few still unclassified but the vast majority of the points that were within this area and indeed many other points that were unclassified of throughout the point cloud are now identified as ground points the buildings are still unclassified we're gonna address that in just a second but we now have a tool that will again I use the term apply intelligence it will give your point cloud a lot more value by by allowing you to generate more points reflecting ground in this case and if we will generate a ground generating a ground model from this we could filter out the unclassified leaving the ground points and because we now have a denser point cloud with ground points give us a some much more precise ground model and will change to another point cloud and we're going to take a little trip down the road from where we are right now I'm going to load up another workspace this is just down the road in Portland another fairly small point cloud in this case we're about 1.6 million and again in this case we have no one grow all points are unclassified confirmed by my legend at the bottom of the screen now in this case we need to do a little processing to get some intelligence in our data I'm going to run the auto classify ground once again in this case we've done it before we're just going to quickly go through that again knowing that we don't have any pre-existing ground points we're first going to identify through that procedure we followed before which of these points are likely to be ground it will render them as brown points or if taya had our way there were to be green points but we'll give them as brown for the time being and based on what's left over we're gonna apply another automated reclassification process so you can see it's status bars removing our non ground points first and eventually it will hopefully render the new points on them up again going through and analyzing each of these as they go through so I see how any other questions coming through here while we're waiting for our status bar I guess or somebody's asking about sources of data I think we've covered that one okay here we go what better got our point cloud um so what we did again same as we've done in our previous example basically identify where there likely to be ground points you'll notice very distinct rectangular features here doesn't take a rocket scientist to identify these are likely gonna be buildings these are large where high style buildings next to them what we're gonna show you here is an auto classify Auto classify now non ground points that's like a tongue-twister Auto classifying non ground points essentially it's going to do what we did with our ground points except look for patterns that define what's likely to be a build building now these are typically going to be elevated we can define the elevation in other words how high above a known ground surface do these need to be to be considered knowing ground if you like again how close do we want to analyze based either on points the facings or meters in other words what level of resolution do we want as far as this analysis is concerned the offset distance on a given plane buildings would typically be a relatively flat surface whether a horizontal surface or inclined and we can do an analysis or we can perform this analysis based on threshold values defining the distance off that plane surface if you like so we can perform certain filters here supply certain settings I'm just going to go with these defaults as they are and what we're going to do as a result of this is reassign some of these unclassified points to be buildings hopefully this will work and also analyze vegetation or vegetation would typically be elevated points that don't follow a single plane but are offset obviously leaves do not exist on a single plane a given tree has leaves you know obviously all over the tree so that is gonna be analyzed as well and those patterns are going to be used to determine what's likely to be a vegetation point well go ahead and click that again it's going to go through various status bars here and hopefully give us the results that we need I've got another question coming in can we take these points now they've been classified and create 3d vector features probably you somebody's been listening to my somebody listen to my bullets yeah we're actually gonna do that as you can see with the results of this process we have reassigned certain points to be more appropriately classified and no manual imprinted this was an automated process you'll see vegetation these green points high vegetation the threshold value if you recall was 2 meters so we've only identified those vegetation points that are likely to be trees all over the over 2 meters tall and we're not dealing with ground vegetation in this case and the building is fairly well defined here you can see these are flat surfaces elevated from the surrounding ground identified as buildings it is not a perfect science you will notice anomalies I know one for a fact in this case right here is none of you noticing where my cursors position this is a highway bridge is a highway and there's a bridge and there's another road underneath well global mapper looked at that as an elevated horizontal plane if you like I thought it was a building but we couldn't manually change those obviously we can manually override some of those anomalies similarly this very large building with so large global mapper Seoul is a ground surface in and of itself again we can use that cross-sectional profiling tool I showed you previously and edit that we can just draw a cross-section i'm not going to encompass the whole building here we'll just do a small section and simply select all of these points in this window and redefine them collectively again this is not gonna be a complete I didn't extend the path enough but you can see very easy for me to do that if I define the width a little more precisely and reclassify those in a manual way so the initial reclassification obviously it does a very good job of identifying where the likely buildings but you can also use some of the manual tools too to clean up the process if you like so what we've done manually editing points reassigning points based on a selection process or perhaps based on a query identifying ground points a very very simple process and then identifying non-growing points in this case buildings and vegetation so there's a number of very very very powerful editing tools that are available to you within global mapper and the reclassification tools specifically within the module the final section of my the process to the workflow today is to derive meaningful data from my point cloud and this is going to give us a ability to essentially vectorize some of these features or perhaps to generate a surfer so we're gonna do that very quickly in just a minute let's look first well we have this layer on our screen at the building the final button we have not addressed in the lidar module toolbar is the ability to extract vector features and if I go ahead and click that button there are two sections here one in which allows us to outline buildings and one which allows us to extract tree points now my tree points will be defined as a clustering of lidar points that follow a specific pattern and in terms of their arrangement that global mapper will identify as an individual tree and you'll see a lot of information that can be garnered from that tree we can also have necessarily approximate the coverage area for each tree in other words draw a little polygon that is an option it does create kind of a messy map to be honest because you're gonna get every tree outline and obviously if there's overlap so those polygons will be overlapped as well so I'm not going to do that right now in this case but it's another option that's available so again two different procedures happening concurrently here and as a result of this we're actually going to generate two new layers in our overlay control center and we'll go ahead and turn off the point cloud when this put this finishes so we can see all the more clearly what we generated here and let's first of all look at the trees this is fun and this I should mention is a work in progress there are technicians as we speak that are developing the next generation of this function which is actually allows to model these trees in 3d themselves right now there are simply points but with these points we can select each one individually and we can identify the characteristics of each tree the elevation represents the elevation above sea level the height value is the height above known ground so we can see that's a 10 meter tree and we've averaged spread and maximum spread as well another resistor tells us the size of the tree you know foresters and arborists will tell you this information is extremely useful the ability to query this and to search this data I'm going to bring up my search dialog box again we can once again run a search on the height value and you can see I'm just running a sword in this case obviously we had our cutoff threshold at 4 meters so that's a minimum value of our height but we can go right up to the other extent and look at the highest tree in this area's 33 meters if you wanted to quickly find out where the the tallest trees were I'm going to manually select these just so we can see the results you see a few down here in the corner selected so very quick and very easy way to analyze where the likely highest vegetation areas are the other layer we generated was our building layer now this is a three-dimensional layer again this is a work in progress we're working at fine-tuning this so that we enforce rectangular buildings perhaps you know you can see this one a fairly well-defined building in fact if we look at it in 3d you can see it's actually a a three-dimensional structure so we actually generated a three dimensional building footprint again the elevation is defined in here precise elevation I saw elevation above sea level and the precise height of that building as well as also noted so very very powerful tool for doing feature extraction from what essentially began just a few minutes ago as an unclassified unintelligent the three-dimensional point for the x y&z file we're able to identify through these algorithms where they were likely to be buildings now the final process once again on loading this map again we're talking about getting intelligent data are getting useful data from a point cloud final works best I'm gonna load up is actually a very simple pre-filtered area of ground points you can see my legend now implying there's only ground points here I did this by way of preparation I filtered out the non-growing points I left just ground points as I mentioned at the start of my presentation lidar more often than not is a means to an end rather than an end and of itself and very often the end is a terrain model here we have our x y and z point cloud i'll actually change back to looking at this by elevation I want to generate a surface now very simple process in global mapper from the analysis menu create elevation grid from 3d vector data I want to go through this procedure there are a couple of options I can choose in terms of how that surface is created triangulation is the default this will go through a process of creating at in essentially a series of triangles connecting each one of the points generating a surface from that but with the availability of the module they go from lidar module you can also run through a bidding process bidding allows you to generate a deep I am digital terrain model which takes the minimum value within a defined area or a DSM digital surface model which is the maximum value within a defined area so two different variations or the average is another one as well so bidding is gonna take a lot less time to process give you a much smoother model at the end and obviously let you define whether it's a DTM or DSM I'll use the triangulation method and method in this case simply click OK so we can see the end result it's creating a triangulated model for all it from all of the points in this point cloud sample again we'll open the overlay control center turn off our lidar point cloud and you'll see now we have a very precise accurate surface model and I'll just zoom in just a little closer here and we're running out of time very quickly here appreciate your patience today and the 3d now this is just a precursor for where we're going next week because we're gonna talk a lot more detail about what we can do with these terrain surfaces I'll talk about volume calculation contour generation which I hope to get to today but unfortunately time is not on our side so many other analysis tools like watershed analysis and view share analysis but the raw material for all of those processes is a terrain surface and we were able to generate a very precise and accurate terrain surface from our line or point cloud final thing to talk about that keep saying this final thing but I promise this is the final thing again the start of my presentation I talked about the two bookends one is importing data the other bookend is exporting let's assume we had performed certain edits to our data perhaps reclassified it or applied classifications that were more appropriate getting point-cloud data out of global mapper is as simple as getting it in from the file menu export vector lidar format we even added lidar as a subset of our vector data the supported formats include the standard AAS which is the standard point cloud that lidar format but we also support the LA Z the zip to the compressed version of line R as well both of those file formats can be imported both of those file formats can be exported so what we're gonna get from this process is simply a series of points based on the points that are their own race with import export of point cloud data I'm again I'm not gonna go through with it but it's simply a case of kicking ok and your export will be your exported file will be generated the other thing that we can do is we can take a terrain surface if your raw material is a raster surface model such as we have on the screen right now we can actually export that as a last file as well now in this case we're exporting our elevation grid format data so we select that from the export submenu and once again even though this is typically used to generate things like a de M file or some sort of grid file in this case we can also generate a last or last file because last allows our vector point files one of the things that we need to decide during this export is the spacing of the greatest going to generate a regular array of points based on whatever spacing that you assign here and obviously it's gonna like the elevation of the original model so somebody provided you with that a one meter resolution de m for instance you can generate a one meter point cloud obviously the more the more precise the high-resolution the original data the better the potential quality of your point cloud and there's little point in actually increasing this beyond the values that you're seeing here because this will just interpolate and add additional points or perhaps they're not necessary but you again can can generate a last file from a raster surface model so importing exporting and all the stuff in between all the processing certainly part of the global mapper package now it looks like we're about out of time hopefully we've been able to get to most of the questions um we got to today I've seen people answering fast and furious well they've been talking here if not we'll follow up as soon as we can on your screen you'll also see the support email address that you can use if you have any burning questions that you have an ant's asked ya and you need want to ask one of our technicians after yeah and also the global mapper forum I mentioned before here we have the URL for the forum I didn't remember to point out in the help menu it's also listed in there and as Ted mentioned at the start we've hit the record button on this at the start of that we did remember to record it didn't we tell hope we did I really hope so so the record button has been hit so this will be processed hopefully within the next couple of days it will be posted to our website and if you're new to global mapper and obviously been dealing very specifically with light art today if you know a little more context if you like it will be added to the ever increasing list of other webinars some of which are very basic introduction to some of the basic functions of the software some more detailed webinars but yeah again from the help menu you can actually see a link to pre-recorded webinars this will be added to the list but feel free while you're there to look at some of the others you can also download the files I believe so if you want to look at those files locally maybe store them locally on your network if you have other folks in your company using global mapper you can use those as your own informational library so thank you for attending today thank you for your patience I apologies we overshot our time just a little bit and taya thank you very much for your help in keeping those questions coming you're welcome and thank you David and thank you all for attending well look forward to speaking with you next month at next month's webinar which is 3d analysis we're looking forward to that one thank you everyone

Channel: Blue Marble Geographics

Views: 32,753

Keywords: LiDAR, GIS, Global Mapper, Global, Mapper, Terrain, 3D, Point Cloud

Id: U1ohTX2YPkU

Length: 69min 37sec (4177 seconds)

Published: Mon Dec 22 2014