CBChristoph Bauerhansl

Christoph Bauerhansl

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Christoph Bauerhansl has a total of 18 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Evaluating Tree Species Mapping: Probability Sampling Validation of Pure and Mixed Species Classes Using Convolutional Neural Networks and Sentinel-2 Time Series

    by Tobias Schadauer, Susanne Karel, Markus Loew, Ursula Knieling, Kevin Kopecky, Christoph Bauerhansl, Ambros Berger, Stephan Graeber, Lukas Winiwarter

    The accurate large-scale classification of tree species is crucial for the monitoring, protection, and management of the Earth’s invaluable forest ecosystems. Numerous previous studies have recognized the suitability of satellite imagery, particularly Sentinel-2 imagery, for this task. In this study, we utilized a dense phenology Sentinel-2 time series, which offered consistent data across multiple granules, to map tree species across the entire forested area in Austria. Aiming for the classification scheme to more accurately represent actual forest conditions, we included mixed tree species and sparsely populated classes (classes with sparse canopy cover) alongside pure tree species classes. To enhance the training data for the mixed and sparse classes, synthetic data creation was employed. Autocorrelation has significant implications for the validation of thematic maps. To investigate the impact of spatial dependency on validation data, two methods were employed at numerous split and buffer distances: spatial split validation and a validation method based on a buffered ground reference probability samples provided by the National Forest inventory (NFI). While a random training data holdout set yielded 99% accuracy, the spatial split validation resulted in 74% accuracy, emphasizing the importance of accounting for spatial autocorrelation when validating with holdout sets derived from polygon-based training data. The validation based on NFI data resulted in 55% overall accuracy, 91% post-hoc pure class accuracy, and 79% accuracy when confusions in phenological proximity were disregarded (e.g., spruce–larch confused with spruce). The significant differences in accuracy observed between spatial split and NFI validation underscore the challenge for polygon-based training data to capture ground reference forest complexity, particularly in areas with diverse forests. This hardship is further accentuated by the pure class accuracy of 91%, revealing the substantial impact of mixed stands on the accuracy of tree species maps.


  • Open Access

    Combining Remote Sensing Data for Habitat Mapping and Monitoring on a Regional Scale – the SEMONA RELOADED Project

    by Anna Iglseder, Michael Lechner, Markus Immitzer, Hannes Hoffert-Hösl, Christine Rottenbacher, Tanja Lumetsberger, Andreas Kasper, Maria-Elisabeth Schnetz, Klaus Kramer, Christoph Bauerhansl, et al.

    Green spaces, from small-scale structures such as green roofs and individual trees in cities to large grasslands and forests, fulfill climate-relevant, ecological and social functions. The protection and monitoring of these spaces as well as dissemination and awareness raising in the field of nature conservation is of socio-politically relevant concern. The project SEMONA RELOADED (funded by the Austrian Research Promotion Agency, FFG) aims to identify these functions through inventories and change detection. The classification and monitoring of areas with biodiversity worthy of protection (e.g. Natura 2000), as well as green infrastructure in settlement areas (e.g. green space monitoring of the City of Vienna - GRM) are obligatory within the framework of nature conservation laws and are also required within the framework of national and international reporting obligations. Currently, such studies are often based on expert-based mapping in the field (biotope types) and/or indices derived from individual remote sensing data.The motivation for SEMONA RELOADED is to support this labor-intensive process by linking regionally available very high spatial resolution remote sensing data such as airborne laser scanning (ALS) and aerial photography (AP) with high temporal resolution sentinel data (S1, S2). In addition to assisting with the initial identification and classification of green space, including remote sensing data in the workflow should enable constant monitoring of the areas. This builds on successful results from the feasibility study completed in 2021 (SeMoNa22). The processing of test areas in Vienna has shown that the combination of S1 and S2 as well as high-resolution AP and ALS data has high potential for the differentiation of biotope types and green infrastructure in urban areas. By training classification algorithms using combined features, different biotope types could be successfully identified in test areas. In the inner-city area, green roofs could be successfully identified as a sub-area of green infrastructure monitoring better than with previously applied methods.In the presented follow-up study, the research area is enlarged to a regional scale including the protected areas of Nationalpark Donau-Auen, the Vienna Woods Biosphere Reserve and the Natura 2000 area Wachau, the City of Krems as well as the whole area or the City of Vienna. In addition, different Stakeholders (provincial administration, national park and biosphere park administration, federal forestry office) are included in the research process to ensure the applicability of the developed methods for the applied use in mapping and monitoring. In the presented poster, the relevant outcomes of the previous feasibility study will be presented and an overview of the planned research activities of the current SEMONA RELOADED project will be given.