Photogrammetry in sports

Date 2023-08-08

3D modeling of terrain chambers and sports facilities offers unimagined possibilities

Photogrammetry uses non-contact measurement methods and certain evaluation procedures to obtain an exact three-dimensional reconstruction of an object from large quantities of photographs of it. In the MCI Medical, Health and Sports Technology program, photogrammetry is a part of sports technology and is mainly concerned with 3D modeling of terrain chambers and sports facilities. 3D modeling of the Earth's surface is not a new topic: Google Earth took up the subject on a global scale back in the early 2000s. But since then, a lot has happened in the quality of camera sensors, the evolution of handy drones as well as in the computing power of computer image processing.

At the beginning of 2022, the Health Tech research focus area at MCI looked for ways to integrate a virtual toboggan run that is as real as possible into an application with a new project idea called "VRodel". The development project, funded by the state of Tyrol, is being carried out in cooperation with Rodel Austria and the two companies mfku and maab and is intended to generate a virtual toboggan simulator for training purposes. In order for the virtual tobogganist's experience to be as real as possible, a realistic model of the environment is absolutely crucial. Using drone-based photogrammetry, this model could be generated.

With the help of two drones and four ground-based cameras, about 5,000 images of the toboggan run in Kühtai were taken from different perspectives, which were later processed into a 3D model using image processing software. You have to look very closely to see that in image 1 the left image is a drone photo and the right image is the generated photogrammetry model. Terrain models like this one are processed and cleaned accordingly in order to be later embedded into the post-modeled far field of the Kühtai. In addition, the entire control of the virtual toboggan is programmed using the controllers of the VR goggles by means of the Game Engine Unity.

Another project has already dealt with a similar topic. As part of a master's thesis, the ice track in Igls was modeled using photogrammetry and embedded in the surroundings of the Inn Valley. The track, which was photographed in summer, was transformed into a wintry scenario using textures and the track can be ridden with a player (see image 2).

3D models not only find their place in virtual reality: Global terrain models such as those used in "Google Earth" or applications such as "Fatmap" are now standard for route planning for many mountain sports enthusiasts, especially in ski touring. These models are particularly suitable for planning ski tours in extensive terrain. For planning the ascent of narrow terrain chambers and gullies, the global terrain models reach their limits. The following extracts of terrain chambers from the Kalkkögeln show a comparison of the global terrain models from Fatmap on the left to the 3D models created by drone flight and photogrammetry on the right. The thickness of the models becomes especially visible when looking at some places in detail (see images 3 to 5).

The photogrammetry models provide structures and clear geometries where the information from satellite images and large-scale terrain surveys reach their limits. This terrain information offers, for example, the possibility for a skier to find new skiable lines in a slope and also to assess them much better with respect to hazards such as avalanches, terrain traps and rockfall. Another advantage is that the model can be rotated into the relevant perspectives for the downhill skier. This makes it much easier to put oneself in the skier's later viewing position (skiers view) and thus to keep to the planned line based on prominent points from the correct viewing perspective. With classic photos of ski slopes from opposite sides (lookers view), adherence to a planned line is an extreme challenge (see image 6 Same terrain chamber from different perspectives).

Discover the Ampferstein tour and the 3D model in the following video:

So when will we see Google Earth and other 3D maps in the quality of the photogrammetry models shown? Unfortunately, the private freerider will have to wait a little longer for that, because the image data generation is too complex and the generated 3D models are too powerful to be viewable on a large scale on smartphones. In addition, the extreme resolution of the models creates the problem that a mountain can quickly look different after a few snowfalls and the model loses reliability with regard to critical narrow points in the terrain. With regard to freeride events or in freeride-oriented ski resorts, however, there could soon be the first attempts to provide skiers with such models for planning tours.

Where does photogrammetry find further applications? If photos can be generated of an object or a terrain chamber from different perspectives, it is possible to model it exactly by means of photogrammetry. Whether the photos are taken by SLR camera, drone or microscope, and whether they are of workplaces, mountains or surgical wounds, does not play a superordinate role. What is certain is that the finished models can be used for much more than just visualizations. The models can be used for simulators and game applications, optical measurement methods, as visual input for broadcasting of sports events or for surface calculations and avalanche simulations and therefore offer exciting tasks not only in skiing.

This project is supported by the province of Tyrol within the Tyrolean cooperation funding.

 

Share article

Share LinkedInShare TwitterShare LinkedIn

MCI project staff David Mikulic und Jonas Kreiner during drone flight in Kühtai © MCI

Biosynthesis of Hydrogen from Biomass Using Dark Fermentation
Biosynthesis of Hydrogen from Biomass Using Dark Fermentation
Innovative research project at MCI investigates promising energy source
New Doctoral Student at the Center for Social & Health Innovation
New Doctoral Student at the Center for Social & Health Innovation
MCI alumna Annabelle Fiedler will work as a PhD candidate on the topic of care systems and innovation in the healthcare sector
Can AI predict natural hazards better?
Can AI predict natural hazards better?
How AI-IoT technology can make the Alpine region more resilient