It’s all about data!

As part of the reorganisation of LSC Engineering Group, the GIS & Topo teams at Luxplan and Luxsense merged to become Geodata. The three units had in common the collection, processing and management of spatial data, but with different tools, levels of accuracy and exchange formats.

How did the new Geodata division come about?

Gilles Rock: Geodata is a technical term that refers to spatial data, with a geometric component that concerns measurement via 3D scanner, topographic survey, drone or other means, and a semantic component that concerns content: does this point represent a pole, a building, a tree? The three pre-existing teams have the same mission, which is to collect, process and make data available, but each did so using its own method: one mainly using aerial photographs, the other using traditional surveys. The degree of accuracy also differs, ranging from millimetres for topographic surveys to centimetres for orthophotos.

You brought together the expertise of the three teams to create a prototype for a continuous construction site monitoring solution called Instant-as-Built. What conclusions do you draw from this experience?

Sascha Rohner: To create Instant-as-Built, we had to capture the infrastructure as it was being laid in the trenches using a smartphone, refined with a small GPS antenna. These images were then processed to generate a 3D model. It is clear that one of the three teams alone could not have done this, because there would always have been a missing piece. We needed all the expertise: that of surveyors, geomatics engineers, geographers and engineers. Everyone brings their own know-how and vision, which is enriching.

What technologies do you use? And what do these technologies bring you?

GR: Technological development has been exponential over the last 5 to 10 years, and technology allows us to improve our performance, speed and accuracy. Whereas in the past we had to send a surveyor into the field to take measurements when we had any uncertainty, today we are able to scan millions of points. We have all the equipment we need to do this, from the simplest to the most sophisticated: theodolites, 3D scanners, drones, smartphones, tablets and GPS antennas. This allows us to choose the most appropriate method for each situation. We then extract the standard geometric data for roads, poles, tunnels, etc. And if, afterwards, we still need to know the diameter of a tree or the height of a hedge, we can find this information in the stored data.

This minimises the time spent on site, but increases the time spent in the office, and we have to constantly invest in user training, hardware, software and improving routines to be able to manage this large amount of data. The hardware is beyond our control, but we make sure we stay at the forefront of the tools available and make full use of their features. Standard tools cover 80% of our needs, but we go further by automating the processing of data collected on site. These automated processes, set up by our developers, represent not hours or days, but months of work, back-and-forth communication and fine-tuning to ensure that the data captured via a smartphone reaches the platform via the servers with the desired accuracy and quality.

SR: I would add that the subjects covered are very complex and the amount of data extremely large, but all our solutions must remain manageable and easy to use for our customers. I like to say that anyone who can post a photo on social media is capable of using Instant-as-Built: all they have to do is take a video with their smartphone and place it on the platform where it is processed by the algorithms we have put in place, then they can move around the 3D model, take measurements, reopen a trench virtually, visualise how the networks have been laid, all in a very user-friendly way. This tool also helps to clarify certain situations and, ultimately, makes the construction site run more smoothly. The point cloud and/or video of a section guarantee a certain degree of data accuracy: if 25 metres of kerbs are measured, all parties involved agree that it is not 22 or 30 metres.

The same applies to drone surveys and municipal spatial data (heritage, infrastructure, roads, cemeteries, forests, etc.). We work with 62 of Luxembourg’s 100 municipalities, to whom we provide the SIGcom platform. Municipal technicians have direct access to the data they need: if they want to analyse a network, for example, they can do so in just a few clicks without having to download a map or take measurements. This platform functions as a geoportal containing specific municipal data collected and/or integrated by our services: its PAPs and PAGs, its infrastructure, its road signs, its bus stops, its cemeteries, etc. We are fortunate to have a large team of GIS experts within the group, which also allows us to develop additional modules on demand. Smartphones and tablets are no longer just tools for consultation, but have become tools for collecting data in the field: if the gardening department has tended to a tree, it can document this on the platform, and the same applies to the lighting department if it has repaired a street lamp. All the information is thus accessible in real time.

Based on reality, we can create digital twins of the most successful municipalities and construction projects in Luxembourg. These twins bring together the existing and the planned. It is possible to play with certain parameters – adding a floor to a building, for example – and to integrate BIM data. We can simulate torrential rain in a municipality with several variants – 100 mm, 200 mm, etc. – and measure the consequences, test different measures – installing a barrier, creating a retention basin – and see the impact of the different scenarios. The digital twin also makes it possible to simulate the shadow created by different elements on a building depending on the time of day, or to assess the radius and locations from which a water tower is visible at the touch of a button. There are so many possibilities that were still reserved for experts five years ago and are now accessible to everyone!

In short, new technologies and artificial intelligence have become indispensable, but human intelligence remains essential.

SR: Today, we deal with such large volumes of data that it has become impossible to analyse them “manually”. That’s why we set up routines and use artificial intelligence to extract the data we’re interested in. For example, to quickly find out how many pedestrian crossings, public trees, bridges or bus stops a municipality has, all you need to do is choose the right method to obtain a data set that is 80% reliable. Artificial intelligence is very effective at analysing spatial data from aerial images. For example, it can be used to estimate the stock of materials on a construction site or to count the trees in a forest, differentiate between species, measure each tree’s diameter and height, and calculate the total volume of wood available. However, in order to achieve a certain level of quality, extensive experience in the field of data and the expertise of an interdisciplinary team are essential. We are delighted with the possibilities that are emerging with all these new technologies and we are very pleased to be able to respond to all kinds of challenges in the field of spatial data with the new Geodata team.

Excerpt from the article published in NEOmag.