Using Point Clouds with GIS part 2

In by Harriet Brewitt

The second in a series of pieces from contributing blogger Nicholas Duggan, check out part one here. Thanks again to Nicholas for this week’s contribution!


Point Cloud Visualisation and Rendering

In the previous blog I briefly skirted over how GIS can be used to put context to point cloud data. In this, the second part, my aim is to expand further on the discussion of visualising and rendering point clouds in GIS for analysis.

Generally, for quick peeks at point clouds, something like CloudCompare is great, it allows you to quickly see what is going on, the software that comes with your scanner is usually pretty good for removing and analysing errors. If you want to see the cloud in context, nothing beats using a GIS.

Yes, there are many of you out there right now, ready to shout how GIS is a generalisation of the surveyed cloud and you are right, try zooming in any more than 1:1 (1m) on a projected coordinate system and you’ll find that it won’t go any further. This is more than enough to look at point clouds like those created by the Robin as, at 2cm, you will see around 50 points per metre, good enough to take measurements and to understand how things will work.

As long as we manage our expectation, we can make full use of a GIS, which is ideal for analysis and visualisation of a point cloud. I’ve worked on a few projects which we’ve been able to use a point cloud and put it into mapped data to see how a building influences the surroundings or using the point cloud of an office and placing it correctly within a building. With GIS we get more for our money, as we can insert attributes (information) within the points, so when we click on points we can receive information like the exact position, intensity, elevation, class or even user entered information.

LiDAR capture of Nottingham Tramway

Just like many of its visualisation counterparts, most GIS also have the option to adjust which points you want to use and how you want to colourise them. In the example above I am using the Esri ArcGIS Pro software which has options to classify the colour scheme based on the information within the cloud. Where there is RGB data, I can draw up the data using this, or, as here, I am using the intensity but there are other options too…I can just view the first return, all returns or even view the cloud as a TIN.

Collage of different point cloud functions

So, it comes as no surprise that GIS is good for analysis, but it is a little slow, right? Well, not anymore…in the last few years there has been huge progress in 3D web mapping, meaning that companies like Cesium, Esri and Bentley have had to develop standards to cope with these large data. My favourite (at the moment) is the Esri i3s or Scene Layer Package, as it is called. By converting your LiDAR data over to this octree type format, you can render huge point clouds in both the desktop GIS, in 3D and also on the web as a 3D scene as it renders only what you need on screen. Of course, the Cesium and other octree type formats are just as good, but the Esri format stands out as it is the only one which works with a GIS.

GIF of Jubilee bridge in i3s format

You can see the speed of the point cloud rendering in this GIF captured in real time in ArcGIS Pro or you can view this data alongside some simple 3D buildings, by using this link.

There is further functionality which can be utilised by using the Esri javascript api and creating our own web map. By doing this we can add further tools such as the ability to control the density of the point cloud in points per metre and even the point size itself. That is as well as viewing the cloud alongside other GIS data.

Tool in action


Thanks to Nicholas Duggan in conjunction with Garsdale Design and GD3D.