Laser Scanning Value

Or is everyone making money except you...

A majority of design and construction professionals receive value using traditional laser scanning methods to measure as-built environments. The items that dictate the value of laser scanning 1) Area of detail, 2) Accuracy of measurements, 3) cost, 4) time.

If memory serves me well it was 1997 in Chicago at a Miller Freeman conference called MechanCAD where I saw my first laser scan. After a number of questions about 1) cost, 2) work flow to get it into AutoCAD, 3) accuracy and usability, I concluded “Not something I can use in production”.

Fast forward to 2012 on a big factory conversion project. Management was convinced laser scanning could be converted into a usable background model of the factory. So Millions of dollars and many months later a model existed.

We were introduced to the 3D model and the process of going from laser scan to 3D model. After a few weeks of 1) watching and measuring the work flow, 2) seeing the quality of the model, 3) putting together the costs I concluded “How it this productive”? What I saw was slow, expensive, poor quality. The contractors producing the model and scanning were making money, but the trades using the data were suffering.

I’ve always been a drafter who wanted to produce a huge amount of accurate work quickly and cheaply. I began programming to do just that.

As 2012 became 2013, then 2014, it was crystal clear to me that everyone on the laser scanning program was making money except the project owner and the trade contractors trying to use it for offsite fabrication. The amounts of data were enormous, the clarity of the conditions was poor. What was wonderful was the photographic view and ability to measure inside that photo, the bubble view.

We developed a process to import our piping models into the laser scan native files. This permitted us to see conflicts by visual examination. On a small scale this was very productive. However, Navis still ruled on large use. Problem was operators were using Navis conducting visual inspections, not clashing against the scans. Very time consuming and wrought with error and omission.

By the end of 2014 I began to see value in laser scanning within a production environment when you remove the parts of the scan you don’t need. Nonetheless the “photographic” data always outclassed any software improvements introduced for laser scans point clouds. The point cloud data became usable as long as we post processed and tailored it to the detailer by “extremely competent” filtering of stray points and cutting all the useless area out. Yet real value add to the detailers was including all the photos of the data, the point cloud.

By 2015 we could use a combination of hand scanners (some structured light) to supplement the inaccessible areas and add more value to the as-built data set for the detailers. Our conditions were so congested you could not access them with a tripod mounted laser scanner.

But the real breakthrough was conversion directly into surfaces! Now the process was getting useful. I'm not referring to converting scan points to simple geometry, i.e. a scan of the pipe run into a pipe model. Any and all conversion software to date (January 28, 2017) is not suited to our production requirements. My point is bypassing the point cloud and going directly to surface meshes.

Months of effort developed a work flow that permits us to provide specific point cloud areas into 3D surfaces. We provide this hybrid data set to the detailer along with the photographs. Think of a 3D polyline snaking it’s way through a maze of pipe. The area on either side of the center-line is processed into really decently accurate surfaces and emailed to the offsite detailer in a matter of hours not days.

Another issue we have to address is the fast moving dynamic nature of the work we do. The condition we have in the morning does not last long. Our environment requires us to be able to “map out” an area and get it to a detailer in hours – not days.

All our data is propriety or I’d show you examples. However a dedicated search using google will reveal others doing various portions of what we do.