Last week I did the Canadian Snowbird thing and headed to Phoenix for GITA 2010, the Oracle Spatial User Conference, and a rather lopsided NHL Game 7. I came home, not only with a souvenir ticket, but also with a solid sense that we will have plenty to do to cope with the continued stream of new data formats and the impending LiDAR tsunami.

Another Day, Another Format
Though I wasn’t able to see it myself due to scheduling issues, Peter Batty’s panel discussion was the buzz of the GITA conference (Joe Francica’s write-up of it is a must read, and *UPDATE* Peter has posted a video of the discussion on his blog).

From this panel discussion, I particularly enjoyed OSM-founder Steve Coast’s colourful question about why there is so much “goofing” around with formats in light of the next day’s announcement of a new OSM binary file format!

The announcement gives a good description of why there is so much goofing around with formats: different application scenarios impose different requirements, and hence it is nearly impossible for one format to rule them all. And yes, the announcement does seem to provide confirmation that “all of this has happened before and all of it will happen again”.

So, another format for us to track.

The LiDAR Tsunami
And speaking of formats, point clouds got mentioned almost everywhere I went. Stu Rich correctly referred to LiDAR as being a disruptive technology in his GITA writeup. A similar theme was present on the floor of last year’s Autodesk University and a lot of my hallways discussions at the Oracle conference involved the point cloud Oracle spatial data type, and managing the storage, display, and application of this data.

With the LiDAR tsunami coming, the world is suddenly awash with data. The cost of collecting LiDAR point clouds continues to drop dramatically and the data volumes that these new sensors provide is staggering. The raw sizes present serious challenges for software vendors on how to handle the sheer volume as well as to analysts on how to make sense of this very accurate data which is already seriously altering the craft of field data collection.

So I’m anxious to ask — how is LiDAR disrupting your workflows? Is it off in the future, or old hat by now, or just too darn voluminous to be useful?

Drop me a comment, I’d love to hear your perspective.

By the way, the NHL game in Phoenix was a bit lopsided, but I sure couldn't complain about my seat.

By the way, the NHL game in Phoenix was a bit lopsided, but I sure couldn't complain about my seat.

About Data Data Formats LIDAR Oracle 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!

Comments

9 Responses to “Sensing Disruption at GITA 2010”

  1. Michael says:

    Hello Dale

    This brings back a thought I had, more than 25 years ago, when I had finished University, was starting my job in GIS business and had just learned about CAD systems, vector data, digital maps, etc.

    I wondered about what would be the most computer-like way to handle GIS data. I came to the conclusion that it is NOT vector data. I think it is some kind of raster data, with a real-life cell size of ~0.1m and a collection of powerful algorithms which then extract what we need.

    The “cell” of course would not just be a color-value, height, or so. It would be a container with a coordinate which can hold a vast amount of individual information which is then filtered for the use case in question.

    Somebody could use it as terrain model, someone else to extract an aerial image, somebody else to extract the vectors we use today, which maybe borders, streets, you name it.

    Back then, ~25years ago, I could not imagine machines which are able to handle the amounts of data of such a world-wide raster approach, but it seems we are heading in this direction.

    It was a dim idea then, but is more like a … “what’s missing” now.

    Michael

  2. Annette says:

    Hi Dale,

    I spent some time today processing contour data derived from Lidar data.
    Lidar data is increasingly available here; its been a regular FME course question for the past few years from all sectors of industry.

    Annette

  3. Dale Lutz says:

    Michael — in my youth, I had the idea of doing a masters degree on applying Voxel technology to GIS problems. Forgot all about that until your comments today. Perhaps in some future alternate universe there will have to be a “Voxelator” transformer in FME that can take an irregular point cloud and populate it into a Voxel dataset.

    Annette — good confirmation to know that you’re living the LiDAR dream already. And good impetus for us to give you more to teach at future FME courses on this topic as well!

    Thanks for your comments.

  4. Michael says:

    Hello Dale (again)

    On second look (my screen at home is better than in the office) I wonder if the
    picture will make it as a splash screen for next FME version 😉 … only I couldn’t find a lizard …

    Obviously some special lighting (maybe a LiDAR beam) must have hit you exactly when the picture was taken ?

    Whoever edited this, he did it too well to be true …

    Michael

  5. Dale Lutz says:

    Michael — it is our graphic artist Jara that did this fine bit of slight of hand. Yes, she does an excellent job. But I wish she could tone down the white LiDAR beam that makes me look so pasty. I really need to get outside next time I’m in Phoenix and get a bit of a tan!

  6. Peter says:

    Hi Dale,
    I’m certainly encountering a proliferation of LiDAR and other point cloud data eg. from automated photogrammetry – collection is on the up, but utilisation is as yet limited. The datasets, as we all know, are increasingly large. Oracle’s point cloud storage is great, if a little costly for the smaller enterprise to implement. The problem from what I see is that there are few compatible clients deployed in these environments yet, so there’s little visibility of the data outside the select few that play with it in darkened rooms in the basement. That, and direct visualisation of the cloud is not what most of us want – we want some form of derived data, whether it be surfaces, contours, temporal deltas or whatever. These products are often generated by purpose specific software.

    There’s latent value in point cloud data. In the same vein as raster support, the trick for Safe will be to implement a collection of point cloud functionality that helps identify and extract that value. It’s not so much format translation as feature extraction…

  7. Dale Lutz says:

    Great points Peter. It does seem like the Holy Grail of the LiDAR world is intelligent feature extraction, which strangely our human brains seem to be able to do reasonably well, so long as we can navigate the point cloud in 3D. At Safe we won’t *start* with that as our goal, but instead follow what we did for Raster (we don’t do feature extraction there either) and see how we can add value to workflows related to the management and manipulation of point clouds. And if someday we can plug some interesting additional value added transformations into the mix….we’ll be ready.

    Dale

  8. I see the real value of LiDAR is being able to combine it with other sources of data, particularly imagery. Imagery has very important spectral information, yet automating feature extraction is still challenging as humans have an inherent ability to perceive depth in 2D images. LiDAR resolves the depth problem and we have found that it is now very cost-effective, using a combination of LiDAR and imagery, to map land cover from billions of pixels of data. This turns the remotely sensed data into usable information for decision makers.

  9. Dale Lutz says:

    Yes, as we’ve been talking with customers about their LiDAR usage patterns, one particularly innovative theme was precisely this — integrating LiDAR with other data, either vector area features, or raster, and having an extremely valuable result.

    Exciting times.

    Dale

Leave a Reply

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

Related Posts