SEHAG - Historic image processing

For society and science it is important to understand the consequence of climate change on alpine geosystems as well as their sensitivity. While some changes are apparent, e.g. the retreat of the glaciers, other changes and their relation to a changing climate are barely understood. To improve the understanding, changes in the alpine geosystem shall be documented by quantitative exploitation of historic terrestrial and airborne archival images, analyzed in three time slices starting 1850, 1920, and 1980, until now. The start at 1850 coincides with the end of the Little Ice Age and the onset of touristic photography. Our aim is to develop a long time series of historic terrestrial and airborne images in alpine regions. Such a long time series has not been established before. We concentrate on three study sites in Austria and Italy, the Kaunertal, the Horlachtal, and the Martelltal, which have different altitude, exposition, and touristic exploitation. Quantitative information on changes, e.g. of the activity pattern of mass movements or the distribution of vegetation patches, is a fundamental prerequisite for describing the geomorphic and hydrological processes in the geosystem. The precise 3D reconstruction of feature lines, e.g. snow on the glacier or the location of rock fall deposits, shall be documented in close collaboration with researchers in geology, geomorphology (including the cryospheric component), hydrology, meteorology, pedology, and vegetation science. Given the large number of available and necessary images for reconstructing the developments since 1850, the automatic orientation of those images becomes indispensable. Images will be retrieved from the archives and stored with metadata (name of the valley, time of acquisition, etc.). Automatically determining the position and angular attitude of historic photos is so far an unresolved problem. Images of aerial flight campaigns, which started around 1950, will be retrieved from archives, and lead not only to the reconstruction of individual features but also to historic ortho- photos and historic terrain models. The accuracy of all these products still needs to be estimated and documented, in order to separate significant changes from random errors induced by the lower quality of historic images. A stable reference frame will be provided by high resolution airborne laser scanning data. New flight patterns arranged along the most active areas of change will be investigated, opposed to the current practice of regular, parallel flight lines. The collection of historic images shall continue beyond the first project phase of 3 years to improve the quality and completeness of the documentation. In a second project phase especially the effect of ongoing climate change onto the components of the geosystem, projected into the future based on climate scenarios, shall be mapped to the terrain and compared to current photos.

The aim of SEHAG (SEnsitivity of High-Alpine Geosystems to Climate Change since 1850) was to investigate the effects of climate change on individual geosystems (e.g.: glaciers, rivers, vegetation) and their interaction in three selected alpine regions since 1850. The focus of the research group "Photogrammetry" (TU Vienna) was mainly addressed to historical images, acquired both from the ground and by airplane. Another focus was the processing of data sets acquired by airborne laser scanning (ALS), which served as accurate reference for all further work. All datasets were integrated by our project partners into their analyses and models to explore the changes in our study areas. Only through this interdisciplinary approach it was possible to answer these fundamental questions in the context of climate change. ALS is a modern technology by which the earth`s surface is scanned with a laser scanner mounted on an aircraft. The resulting data, produced in multiple flights over the area of interest, must be processed and corrected to produce a highly accurate and error-free representation of the earth`s surface. Especially in alpine terrain with its steep slopes and large altitude differences specialized tools are needed, which could be further developed by us in the course of the project. While ALS was used to acquire highly accurate reference data representing the current state, historical aerial photographs allowed us to look back into the past, in our case until 1945. With historical aerial images it was possible to reconstruct the mapped area three- dimensionally at the respective time of acquisition. By comparing the model reconstructed from the aerial photographs with the current ALS model terrain changes become visible. However, with increasing age, historical aerial images become more challenging to process. Therefore, it was necessary to constantly adapt and extend existing approaches. Only then it was possible to process aerial images at a total of 11 different points in time. Between 1945 and 1850, the end of the last Little Ice Age and the beginning of our study period, there is a gap of almost 100 years. To close this gap, for the first time historical terrestrial images were systematically searched for in various archives. Taken by locals and tourists as memories, these images can be used by us to analyze the early state of the geosystems. Besides the difficult search for relevant images, also the processing of these images for scientific purposes needed a new approach. With the developed methods & tools more than 1000 images could be georeferenced and the image content could be included in spatial analyses with other data sources. The project was funded by the DFG (Deutsche Forschungsgemeinschaft / FOR 2793) and the FWF (Fonds zur Förderung der wissenschaftlichen Forschung / I 4062).