To introduce myself, I’m in the first year of my PhD (having started in October 2013) at Northumbria University and I’m looking at quantifying the magnitude, frequency and spatial distribution of landslides that occur above thinning and retreating glaciers in the European Alps. Prior to starting my PhD, I graduated with a degree in Geography and a master’s degree in Polar and Alpine Change. Part of my PhD studentship comes from an artist (Dan Holdsworth) who is interested in looking at the way alpine landscapes have been perceived and represented in both science and art, and how this has evolved through time. This is where I come in; I’ll be collecting contemporary 3D data from various locations throughout the summers of 2014, ’15 and ’16 and comparing this with historical photographic imagery and artwork to understand how the landscape has changed on a much greater timescale to that permitted by contemporary methods alone. The research aims to further our understanding of how alpine landscapes are evolving in response to changing glacial conditions, but also explore the use of innovative techniques to reconstruct landscapes in 3D and the possibilities that this may open up.
To start, let’s look at some of the techniques and methods I’ll be employing for the contemporary data; I’ll be using a rotary Unmanned Aerial Vehicle (UAV) to capture imagery of slopes and rockwalls from a variety of locations above glaciers across the Alps that show evidence of landslide activity. Our UAV/hexacopter (DJI S800 EVO) will be fitted with a high-resolution digital camera (Sony NEX 7) that is controlled remotely on a different (self-dependant) system to that of the UAV and can provide a 360° view. The camera will be set to capture an image on a relatively short time-lapse interval to ensure that there is significant overlap between each of the images for the 3D reconstruction process later (more on that shortly). Whoever controls the camera will be using a first-person-view (FPV) screen or goggles to ensure the camera is face-on to the slope, enabling the pilot of the UAV to give 100% of their attention to manoeuvring and landing the hexacopter safely. UAVs are becoming increasingly popular within Geomorphology (and the wider geosciences) and their applications offer a low-cost solution to capturing high-resolution topographic data in areas that would otherwise be difficult to survey. The ability to carry a relatively low-weight UAV into mountainous terrain instead of lugging a 30kg terrestrial laser scanner (that’s the weight before the tripod and batteries etc.!) has its obvious advantages as well…
So… I have thousands of images from across many different slopes, what now? A process that some of you may be familiar with or at least will have heard or read something about is Structure-from-Motion (SfM) photogrammetry. SfM has been around for some time within the computer vision sciences but is still comparatively new to the geosciences with more and more researchers realising its potential for generating accurate 3D topographic data with relative ease-of-use. It operates under a similar tenet to that of traditional stereoscopic photogrammetry in that 3D structure can be attained from a series of offset, overlapping images with common features appearing in multiple images from slightly different perspectives. It differs however in that the camera geometry and position are solved automatically without the need to specify image parameters or targets of known positions. As I said, SfM is still an emerging technique and as such is yet to have received thorough testing across many studies, yet those who have used it have demonstrated that the technique can yield data quality and resolutions that are comparable with LiDAR and traditional photogrammetry with unprecedented ease of use. Is this true? Well, as part of my pre-fieldwork testing, I am currently working through LiDAR data collected using TLS and 3D data created from SfM of the same features with known dimensions, captured from the same distances under the same conditions and hope to test the two against each other for reliability and accuracy. I’m doing the same for change detection and removing different objects of known dimensions from the field of view to test the accuracy of measured change. The European Alps is rife with heritage, playing subject to some of the world’s earliest explorers, artists and scientists who throughout history have paid homage to the enormity, grandeur and magnificence of the mountainous environments. This has resulted in an enormous and spatially extensive archive of imagery from across many different locations across the Alps. As part of my research, I’ll trawl through the archives of institutions such as the RGS, the British Museum, the Alpine Club and many more, searching for images of slopes and rockwalls above glaciers from across the European Alps. This is a rich and largely unused resource that I’ll revisit and, where possible, reprocess with the contemporary data to identify and quantify differences over a significant timescale.
I have been lucky enough to receive the Royal Geographical Society (with IBG) Land Rover Bursary for 2014 with a project titled the ‘Grand Alpine Tour’. It’s a play on the original GrandTour which was an educational rite of passage for young gentlemen of the 17th and 18th centuries when they travelled across Europe. Most of the travellers were captivated by the Alps having never seen anything quite like them and spent a considerable amount of time capturing their thoughts and views through artistic works. I’ll be travelling the length of the European Alps over three months this summer starting in Trieste in Italy and finishing in Nice, France.
If you’d like to follow me as I progress through my fieldwork, you can do so using the details below:
Recommended Reading for SfM: Fonstad, M., Dietrich, J., Courville, B., Jensen, J., Carbonneau, P., 2013. Topographic structure from motion: a new development in photogrammetric measurement. Earth Surf. Process. Landforms 38, 421–430. Westoby, M., Brasington, J., Glasser, N., Hambrey, M., Reynolds, J., 2012. “Structure-from-Motion” photogrammetry: A low-cost, effective tool for geoscience applications. Geomorphology 179, 300-314.