One year ago on this blog I wrote about putting LiDAR in its place with FME 2011. One year (and a whole bunch of hard work) later we’re back with FME 2012, and we’ve got LiDAR files everywhere scurrying for cover. They aren’t going to know what hit them.

In the past year we’ve been repeatedly impressed with what we’ve seen our users do with the point cloud and LiDAR support that was part of FME 2011. But as much as we’ve been impressed, we have also been challenged with the increasing data volumes and scenarios our users put before us. FME 2012 is our answer.

A Billion Point LAS File? No problem.

The first thing we did was address performance in the case of astronomically large input. So if you were using FME 2011 to clip or tile your billion point LAS files, you are going to be very pleasantly surprised when you try FME 2012. To put it in perspective, clipping or tiling moderately sized LAS files on a standard computer is about 4-5x faster in FME 2012 as compared to FME 2011 (which was already no slouch at this).

Support for LAZ – Compressed LAS Files

For users looking to archive or exchange LiDAR data, the incorporation of Martin Isenburg’s most excellent LASzip compression technology allows FME 2012 to really cut these files down to size. Thanks Martin for making this available for us all to use! We typically see LAZsip’ed LiDAR files being 30% the size the original LAS, and others have reported reductions down to 10% of original. If you have the need to archive or share any amount of LiDAR data, these space reductions can really begin to add up and save significant storage space. All you need to do to use this is flip on the “compression” setting of the LAS writer in FME 2012, and you’re golden. The standard LAS reader in FME 2012 can just read these as if they were LAS files to begin with.


Geocentric Coordinate Systems

Paul wrote about the importance of Geocentric Coordinate Systems a while back, and FME 2012 is our first release with this support. So bring on those files that store their position as a distance from the center of the earth, and we can turn them into something a bit easier to deal with on old terra firma. We can also go back the other way, taking projected or geodetic coordinates and making geocentric, which in conjunction with 3D output formats like PDF can cast your data in a whole new light.

Oracle Spatial Point Clouds

While we’ve partially supported using FME to load Oracle Spatial Point Clouds for some time now, in FME 2012 we’ve taken it to a whole new level. Got a billion points you need loaded? We’ll load it and block it for you in under 18 hours (on a 3 year old desktop computer, on less). So if you are doing a lot of Oracle Spatial work with point clouds, you really need to check that out.

Support for .lasd (LAS Dataset) in ArcGIS 10.01

ArcGIS .lasdLooking forward to ArcGIS 10.1’s support of LAS files? But not so much the need to have a .lasd (LAS Dataset) file first? Don’t worry, FME 2012 is ready when ArcGIS 10.1 is and will create .lasd files at the same time as it writes out LAS files from whatever input source. Just make sure you have ArcGIS 10.1 installed, flip on the setting, and go straight into ArcCatalog to start using the results.

LiDAR Ninja Tricks – Colorizing Point Clouds

My personal favorite LiDAR Ninja trick is the ability to take a point cloud that has no color information on it, and overlay it into an orthophoto to produce a colorized point cloud. As more software becomes available for visualizing massive point clouds, the volumes of data collected without color can be made that much more interesting and useful by colorizing them.

Colorized Point Cloud

We’re really looking forward to meeting you and showing off the point cloud prowess of FME 2012 on this year’s FME World Tour, so check out the dates and sign up if we’re coming to your neighbourhood. We’ve also got an FME & LiDAR focused webinar coming up in March (actually, two in one day – if you like the first showing you can tune in again for the second!) where we’ll show you how to do all this and more. So check one or more of these opportunites out – and I’m sure you’ll agree that using FME to tame the LiDAR beast is fast, productive, and, dare I say, fun.

We wouldn’t endorse using FME 2012 to colorize the Three Stooges, but if you had a point cloud and an ortho, we wouldn’t stop you either.

About Data FME FME 2012 LIDAR Point Clouds Spatial Data Interoperability

Dale Lutz

Dale is the co-founder and VP of Development at Safe Software. After starting his career working spatial data (ranging from icebergs to forest stands) for many years, he and other co-founder, Don Murray, realized the need for a data integration platform like FME. His favourite TV show is Star Trek, which inspired the names for most of the meeting rooms and common areas in the Safe Software office. Dale is always looking to learn more about the data industry and FME users. Find him on Twitter to learn more about what his recent discoveries are!


8 Responses to “Beating LiDAR into Submission with FME 2012”

  1. The LiDAR points in the LAS file need to be more or less in acquisition order (e.g. in flight line order) to get the lossless LASzip compression rates of 1:6 up to 1:12 reported in the paper or the video that are linked here The worst possible point order is when they are sorted along a coordinate axis, since that completes destroys the coherence of all other attributes of subsequent points (except the one that was sorted by). If the points in the LAS files have GPS-time, one can use lassort.exe from to get them back into flight line order. For incoherent point orderings the compression gains can be significant. You find 40+ counties of Minnesota LiDAR with ~ 1:12 compression ratio of free LAZ files in flight-line order here:

  2. Dale Lutz says:

    Thanks Martin for chiming in, it is very good to know how this works underneath, and indeed the situations you are citing which exhibit better compression ratios are indeed much more likely and probable in the field.

    Let me also take the opportunity to thank you, on behalf of software vendors and more importantly LiDAR data custodians and users, for making your work available for us all to use. I’d like to be so bold as to say that the world is a better place because of it.

    Thank you very much indeed.

  3. Hello,

    the reason that Dale only saw compression rates of 30% (instead of the 7% to 20% we report in the paper and the video of LASzip is that FME 2012 rescales the LAS / LAZ files on output to make full use of the 32 bit integer range. This introduces “precision fluff” (additional low-order bits that contain no actual information) in the x/y/z coordinates that adversely affects compression. Apparently this will get fixed with SP1. Until then I would advise folks to be careful before converting their entire LiDAR holding with FME 2012. You can also consult with me on what to do with LAS/LAZ files that were already or are currently created. There is a (free) tool in LAStoolsthat can remove the “precision fluff”.

  4. Dale Lutz says:

    Thanks Martin for helping us find this issue. With FME 2012 SP1, which we’ll make available in the next week or two, we introduce a setting that preserves the original scaling and offsets when doing LAS->LAS so long as you aren’t doing reprojection, rotation, offset, or scaling during the transformation. So if that is important to your workflows, we will have a solution.

    Aren’t integer coordinates fun? I should write a blog article on that topic alone sometime. I”m again reminded of Louis Armstrong — “Nobody knows the trouble we’ve seen….”

  5. xinxin says:

    LiDAR data classification is very improtant,.we can’t get a satisfied result if the data haven’t been classified, such as putting out DEM, building reconstruction. And now FME seemly cann’t do such things. I want to know whether you have some plan for making FME do something about LiDAR data classification? I think it will push FME more useful in LiDAR.

  6. Dale Lutz says:

    Yes, we do get asked more and more about classification. We have some plans to build tools to help with this, but we also recognize this is an area where much research will be ongoing for the foreseeable future. FME’s ability to fuse vector and raster data with LiDAR should allow us to be come a good platform for this type of work, in time.

  7. Evan in Alaska says:

    Dale Lutz, just want to say I am excited for that day =)!

Leave a Reply

Your email address will not be published. Required fields are marked *

Related Posts