Summary: Ross Winans observes, "recent advances in computer hardware and GIS software have created the opportunity for even GIS beginners to easily work with lidar point data." He invites readers to see what an LAS point cloud has to offer.
3-D is cool. I don’t have any hard evidence to back up this hypothesis, but I do see the kids tweeting about 3-D movies and 3-D printers so I am led to believe 3-D stuff is “in.” I will further hypothesize that GIS is also cool since the industry is trending towards the third dimension, just like the entertainment and design industries. Most (cool) GIS users have probably worked with 3-D data in the form of an elevation product such as contour lines or a triangulated irregular network (TIN). Elevation products are nice because they are supported by virtually all GIS software and are smaller in size than raw lidar data. However, recent advances in computer hardware and GIS software have created the opportunity for even GIS beginners to easily work with lidar point data. If you aren’t familiar with lidar point data, then it might be time to see what an LAS point cloud has to offer. Come take a look—you might just like what you see.
So What Kind of Stuff Can I See in an LAS Point Cloud?
LAS files are a collection of points, each with horizontal coordinates (X and Y) and a vertical elevation (Z) value. In addition to elevation values, an LAS file also provides a common format for storing additional information such as laser intensity, scan angle, return information, etc. Some of this additional information, such as intensity, can be very useful for visualization. Intensity is how much energy from the laser pulse was returned to the sensor. There are a lot of factors that affect the amount of energy returned (such as surface roughness, atmospheric affects, near infrared energy absorption, etc.), but lidar intensity can create a decent black and white “pseudoimage” from the exact time of lidar acquisition. I like to use intensity information when creating breaklines in high-energy coastal environments where the available imagery is out-of-date with the lidar data. When working with intensity values, always remember that intensity is typically uncalibrated, so this data is only suitable for a qualitative visual analysis.
Left: NAIP 8-bit aerial imagery of our study area located along the Hoh River in Washington (data provided by USDA), right: An intensity image generated by NOAA Digital Coast’s DAV from lidar data (data provided by the Puget Sound Lidar Consortium) Click for larger image.
To view intensity information in the point cloud, you can use free lidar viewing software such as FugroViewer or QuickTerrain Reader. There are tons of commercial and open source software options to choose from, these are just two that I am personally familiar with. Another useful way to use intensity information is to create a raster image of intensity values (it will be a smaller file that is more easily used across various GIS software). The Digital Coast Data Access Viewer (DAV) will create intensity images for you (provided intensity is available for your lidar data set and you “checked” the box in the Checkout Menu). However, if you would like to learn to create intensity images manually, there are several tools available. One such tool is lasgrid from the LAStools download. LAStools software is not free for commercial or government use, however it may “be used freely for non-profit personal, non-profit educational, or non-profit humanitarian purposes assuming that they are strictly non-military.” So if your intentions fall under that description, then go grab LAStools. The LAStools user interface can be tricky to use at first, but there is documentation for lasgrid included in the download.
We’re Not in Kansas Anymore
Intensity is very useful in certain circumstances, but a black and white point cloud is pretty boring. Another piece of potential information stored in LAS that we have not discussed yet are Red-Green-Blue (RGB) values. Some lidar point clouds come with RGB values encoded in the LAS file from the vendor but most do not. Luckily, it is easy to add RGB information to a point cloud.
To obtain RGB values in your point cloud, you need three things: color aerial imagery, a lidar point cloud in LAS version 1.2 or higher, and some capable software. Most commercial lidar software has the ability to merge lidar and imagery so if you have a favorite, shout it out in the comments below. If you don’t have a favorite yet, I would recommend using either the USDA Forestry Service’s FUSION soft-ware or LAStools’ lascolor tool. FUSION is totally free to use, however you cannot export an RGB en-coded LAS file. LAStools’ lascolor allows for export of RGB point clouds but follows the same license restrictions as lasgrid. Both of these software packages will include documentation to help guide first time users. Below you can see the RGB encoded point cloud I created using lascolor software and displayed using Esri’s ArcScene. I also included a TIN created in ArcScene with the NAIP imagery overlaid. Take a moment to compare the two products (note that they are created using the exact same data sets).
An RGB encoded LAS point cloud, click for lager image
A lidar TIN with imagery overlaid. (Lidar provided by the Puget Sound Lidar Consortium, Imagery provided by USDA), click for larger image
Pros and Cons
There are advantages and disadvantages to each type of visualization. Features like bridges and tree canopies are much better represented in a point cloud than in a TIN. In TIN visualizations, a bridge tends to form an artificial dam and trees tend to form cones, neither of which reflects reality. Another advantage to point cloud visualization is these products require less processing and quality control. If you look at the TIN above, you will see a break in the road that was introduced when the TIN was created (right of image center). This break in the road does not exist in the LAS point cloud.
Point cloud visualization does have its disadvantages. For instance, if you look closely you will see the point cloud has a lot of data voids under thick tree canopy and over water bodies. These voids are due to a lack of lidar returns in these areas. Unfortunately, there is no way to interpolate and fill these data voids in a point cloud. For this reason, I prefer to only use high resolution point clouds. The data used above had 8 lidar return points per square meter. Check the chart below for a brief comparison of the pros and cons of 3-D visualization products:
Click for larger image.
So now it’s your turn to go grab some lidar and imagery from the Digital Coast DAV and create some awesome 3-D point clouds… all the cool kids are doing it!
Please note that no endorsement of software or data is expressed or implied by NOAA or me. Any resemblance to persons, living or dead, is purely coincidental. Objects in mirror are closer than they appear.
Reprinted from NOAA Coastal Services Center Geozone Blog.