Leica Cyclone 3DR – Smart Automation

Some years ago, creating a point with a total station would take a lot of time as you measure each point individually. With a total station, every point you measure is measured for a reason, as you only capture the points that allow you to extract the information you need. With the adoption of 3D laser scanning technology on the rise, millions of points can be measured in a fraction of the time and the speed of this data capture will continue to increase as technology advances in the future. With this manifold increase in the volume of data comes exciting new opportunities for automating its processing and extracting usable information and deliverables.

A first step that was addressed some years ago by Leica Geosystems was the possibility to publish and share this captured data via collaborative software platforms like Leica JetStream and TruView for example. Sharing data allows users to see the collected information in one place, making it quickly and easily available to someone to visualise and work productively with the point cloud. In other words, these collaborative platforms already allow users to access the collected data they want to see, when they want to see it - anytime, anywhere.

Magna – User case study

During a recent talk from MAGNA at HxGN LIVE 2019, Magna compared two data collection methods in a case study where they were tasked to answer a simple question: is there enough space in an old tunnel for a modern train? This question was asked for the following reasons: no original drawings of the tunnel’s surface were available and it had already been modified multiple times after it was built to support different track gauges.

The methods they compared were conventional topography on one side (classic single-point measurement approach instead of the 3D scanning approach on the instrument) and 3D laser scanning (millions of points approach) on the other side.

The data from the conventional approach was first registered in Leica Infinity and then processed in Leica Cyclone 3DR (formerly 3DReshaper) was used. Magna reported that they were able to use the algorithms available out of the box within the software in order to efficiently extract the information they were looking for within the point cloud.

In Cyclone 3DR, the tools they used were:

• Noise reduction in order to clean up the point cloud
• Removing vegetation in the point cloud with the DTM tool
• Creating a surface in the tunnel with meshing tools
• Simulating train envelope using extrusion tools
• Generating reports

In this use case, the DTM tool, which is looking for the lowest points in a point cloud was critical. For this application, the vegetation was not critical to the analysis and thus it can be removed. What is critical is the ground and the cliff. In order to distinguish the cliff from the vegetation, Magna utilised the flexible DTM tool to define the vertical direction. Defining the vertical direction as perpendicular to the cliff allowed them to extract the surface of the cliff even though it is a vertical surface.

The tunnel itself was meshed using the 3D meshing algorithms as well as all the tools allowing to fill holes, smooth or generally improve the mesh. This triangulated surface was used as reference surface for all collision avoidance analysis further on. The comparison of the two approaches included both an accuracy comparison and a cost comparison.

The conclusions were as follows:

• In terms of time spent, the conventional approach with the total station required more than 2 weeks whereas the laser scanning approach delivered a result in 2 days: 1 day on site to capture the data and 1 day in the office to process the data (i.e. 5 times quicker than the conventional approach with the total station).
• In terms of accuracy, the data provided by the laser scanning approach was reported as more accurate, more detailed and in the end more reliable since human error was all but removed from the process by eliminating the need for the operator to determine which points should be selected for surveying. Both accuracy and density of data were highlighted as benefits of laser scanning. Each individual stone on site which were captured in the point cloud through laser scanning rather than the subjective determination of what was important by the field capture teams when working with total station data.

The laser scanning approach was therefore validated by Magna as it allowed them to improve both productivity and data quality. They reported that the key component was the Cyclone 3DR software which was used to process the data. Thanks to the laser scanning approach, Magna were able to deliver the right information at the right time. After careful analysis, the conclusion was that a modern train can go through an old tunnel without expensive modifications.

Components of smart automation

This analysis reported by Magna, is exactly what we want to deliver - accurate and reliable information channelled through an efficient data processing workflow to support real-world decision making.

The laser scanning workflow includes six components that are fundamental to increase productivity and efficiency. These 6 components allow users to extract the right information through smart automation:

1. Storing data
   With the increase of the number of points you can capture with 3D laser scanning and the ease-of-use of the new technology, the scale of 3D point clouds has increased significantly and will continue to grow in the years to come. You can’t just use hard drives to store the data in order to access it anymore for example. The database approach offered within JetStream serves as one single source of truth with many users being able to access it remotely and simultaneously and work with the same project data in different CAD environments while syncing their changes back to the central database.
   
   
2. Quickly access and share data
   A central and reliable database is important. Being able to quickly access this single source of truth and share it with your project members or clients is key. With big data sets, all the points captured should be available all the time and rendered in your chosen platform so that you get the information you need where you need it, with no frustrating lag times, despite the size of the data.
   
   
3. Interoperability
   In our 3D data consumer world, the amount of different file formats out there is just enormous. Some being open and specified by consortiums and others proprietary binary formats—and not one contained a complete project record from points to imagery to tags, measurements and metadata associated with the project. Any given project may involve several different software for data preparation, inspection, modelling and QA/QC or sharing. Supporting simple data flow into and out of Cyclone 3DR is critical.
   
   These efforts led us to propose different approaches: First, the LGS file formats, available since 2018, is Leica Geosystems universal digital reality project file. LGS files are designed to provide the power of JetStream, in a transferable file that can be consumed in many different CAD environments via CloudWorx plugins. Second: direct Sent to CAD functionality. This allows Cyclone 3DR to send data directly to another 3D software, for instance AutoCAD and MinePlan from Hexagon Mining. third, support for industry standard formats. Within Cyclone 3DR. Up to 15 different files formats are supported for import in Cyclone 3DR and recent efforts have been made to support AEC file formats (IFC and RVT as well as export to 3D PDF).
   
   
4. Algorithms
   Algorithms are a key component to smart automation. Different techniques for the extraction of information from point clouds help to automate the data processing by consolidating all of the steps into a few guided steps.
   
   In the use case reported by Magna, three algorithms were identified as key: noise removal, meshing and DTM extraction. These algorithms are the fruit of intense development efforts over the last years. But of course, there are many others available in Cyclone 3DR which have proven their efficiency, for example creating clusters of points according to their spatial distribution or the extraction of breaklines.
   
   We strive to make use of the latest techniques. Often, tools and techniques developed for one domain might fit very well within the workflow of another domain, as is the case of the Digital Terrain Model tool in this study. While DTM was initially conceived of as a Survey tool, it was applied with great success to this specialised extraction workflow—an application far removed from its original design.
   
   
5. Dedicated workflow
   We have touched upon the algorithms, data interoperability and quick access to a large centralised database. These features serve as a toolbox of great flexibility that can support a large variety of applications. For users facing a new, niche application, this assortment of tools is great. However, in order to automate the workflow for a common application, dedicated workflows tools and customised workflows are necessary.
   
   Industry and application-specific workflows allow users to start from the import of their point cloud data and guide them all the way through to the generation of the deliverable, including each intermediate step. With the transition from 3DReshaper to Cyclone 3DR, the flexible toolbox was re-configured into guided workflows for different applications such as cross section inspection or tank inspection. Likewise, Cyclone REGISTER 360 was also designed around a guided workflow, in order to allow beginners or novice users to process simple projects.
   
   
6. Flexibility
   The last component that requires attention here is flexibility. How can we make a product that is flexible enough to adapt to new, unknown applications and uses rather restricted to addressing the uses it was initially planned for? In Cyclone 3DR, scripting offers ultimate flexibility for users. Scripting allows users to leverage the entire toolbox of 3DR to automate a succession of recurring tasks. For instance, always loading a point cloud, removing the noise, aligning with a CAD data, do a comparison and generate a PDF report – all of this can be done automatically, with or without user input.
   
   But scripting in Cyclone 3DR is much more than just repeating a succession of recurring tasks that already exist within the Cyclone 3DR toolbars. Thanks to access to low level functions such as getting all the points within a sphere or creating the best fit line from this subset of points, users can really develop additional algorithms and extract the information they are looking for. Such an approach was proven in a script dedicated to extract curbs in a street. In this script for example, six low level functions are used:
• Separate points at a given distance to a line
• Point cloud reduction
• 3D meshing
• 3D best fit of 2 meshes
• Applying a matrix transformation
• Best Plane

The access to these low level functions make it possible to apply the script to data coming to multiple kind of sensors: imaging / handheld laser scanner (BLK360 or BLK2GO for example), high density scanners (RTC or P-Series for example), Mobile mapping systems (Pegasus:Two and Pegasus:Backpack). This script will extract kilometres of curbs and will stop only when no points were measured on the curb!

The right information at the right time

The right information at the right time While, we pride ourselves in providing tools to customers that provide them with the right information at the right time, we also welcome information from customers to improve our products.

It is through the support of customers and their feedback from the field that we take guidance in developing our roadmap. If you have any product feedback, we encourage you to contact us through this form.

For more information on Cyclone 3DR, please visit our product page or contact us.


Gilles Monnier
General Manager, Technodigit
Reality Capture Division