Merged & Decimated vs Separate & Full. Undet ToolBox Philosophy

Merged & Decimated vs Separate & Full. Undet ToolBox Philosophy

Hello everyone! I hope you had great Easters, relaxing Worker’s day, lovely Mother’s day and other holidays that appeared around the globe in the past weeks. I didn’t post anything during this period and that’s thanks to you! After the last post regarding our beta release of the new Undet ToolBox version we received lots of request for beta trial evaluation. So all of the time I’ve been providing one on one live demos to all of you and I really appreciate your interest, spent time and received feedback on our software.

I know I promised to post some recorded videos regarding each of the new functionalities of the 2nd release and I’ll do it very soon! But during this marathon of online demos I came up with an idea to prepare a post which should clarify many uncertainties that appeared during our discussions. I named it ‘Undet ToolBox Philosophy’. My goal of this post is to clearly express the way creators of the software think, focus their developments and how can you benefit from it in your point cloud processing workflows. And believe me, you sure will.


In the past years 3D laser scanners improved dramatically both technologically and in terms of prices. If the old ones time-of-flight based scanners seven years ago captured something like thousands of points/sec and the phase based ones something like hundreds of thousands of points/sec, so now time-of-flight based scanners capture 10 times more and the phased based scanners reached million points/sec. As the market ‘Mercedes’ such like Riegl or Z+F kept sustaining their innovations and forced to move and improve rest of the players, Faro came in with a disruptive innovation of several time lover price of Focus 3D scanner, making this technology more available for masses.

But what about software? Have the desktop applications for point cloud processing improved as much as laser scanners? Let’s say you spent one day in the field scanning your object with a fancy hardware which captures 1million points/sec, spent one more day for adjustment and registration. If your final deliverable is 2D documentation or 3D model you’d probably transfer your point cloud into CAD. And here is the annoying part – you know that CAD will struggle dealing with all this huge and rich data, so what you usually do? Let me answer this, you decimate your point cloud to half of the points or even less to increase CAD performance and merge your scanning positions to make your point cloud more manageable. So here is my open question:


Do you use only half or even less of your fancy scanner abilities?

It would be really interesting to know your thoughts regarding this. I hope we can make some discussion so do not hesitate to drop down your comments over here. Anyway, let’s move further and take a look to some Pros and Cons regarding Merged & Decimated vs Separate & Full point cloud.

+ As already mentioned, merged and decimated point cloud is lighter, which increases your CAD performance and makes your point cloud more manageable.

– But if scanner captures millions at a time and later you decimate point cloud, you end up in losing information of details and unclear geometry of small objects.

– You already have initial scanning error + adjustment error and if you merge decimated point clouds some captured elements might disappear at all or duplicate. In such case you lose ability to check if this is an adjustment error or a real as built geometry.

– Merged point cloud = overlapping intensities values. Overlapping intensity values disturbs to clearly identify objects. Ability to visualize intensity only from one scanning position allows to see true values.

– Merged point cloud = overlapping RGB values. The same situation is with natural colors. Different scanning positions end up in different angles, different lightning and different time of the day.

– In merged point cloud you lose scanning position and ability to see objects just as if you were standing right where the laser scanner captured the scan data.

Resume, you’ll win some performance in CAD software but lose confidence and ability to accurately and precisely determine as-built objects geometry. If all these issues affect human being abilities to recognize structures in point clouds, don’t you think that at the same time it handicaps automatic recognition algorithms? It sure does…


Is it possible to have performance and easy manageable point cloud in your CAD without any loss of captured information?

The answer is yes. In fact it is one of two cornerstones of Undet ToolBox philosophy.

Cornerstone 1. Undet ToolBox connects best from laser scanners and CAD.

The idea behind this is that we want our users to import full information from laser scanners – registered but undecimated and unmerged scanning positions. Undet ToolBox itself will decimate point cloud providing you with high performance. But our decimation does not mean that you lose information because anytime you can control your level of visible points and if you need to go into details and visualize all scanned points you can do it just with one click. Second thing is that in our software you are able easily and interactively control different scanning positions, turn them on/off separately and select different coloring options.

This corner stone made foundations for our primary Undet ToolBox release which provides best value for users that want to work in traditional AutoCAD workflows with enhanced point cloud processing capabilities for visualization, management and high performance without any loss of as-built details.

Cornerstone 2. Undet ToolBox is horizontal but vertical.

I know it sound like a mistake but it’s not. We assume that users already know how to work with their standard CAD. Either user is using simple AutoCAD or vertical solutions like AutoCAD Architecture, Civil 3D, Plant 3D one already has one’s style and workflow. We do not want to change it. We concentrate our developments only on point cloud side. We do not want to reinvent the wheel that some players try to do. For example, we do not create any architectural components or libraries for pipes because AutoCAD Architecture and Plant 3D have this already. We are focused on extraction of critical information from point clouds which is necessary for efficient workflows in AutoCAD and its verticals when dealing with captured data.

This corner stone made foundations for our second Undet ToolBox release which provides best value for users that want to work efficiently with point clouds in AutoCAD and speed up point clouds vectorization into 2D drawings and 3D models. Therefore, it has a bunch of great features which are applicable in many vertical industries like Architecture & Construction, Cultural Heritage & Archaeology, Civil Engineering & Land Surveying, Oil & Gas.


Would you like to use all your scanner abilities now?