Irena Hajnsek


Last Name

Hajnsek

First Name

Irena

ORCID

0000-0002-0926-3283

Organisational unit

03849 - Hajnsek, Irena / Hajnsek, Irena

Filters

2013 - 2019132020 - 202562
Reset filters

Search Results

Publications 1 - 10 of 75
  • Hybrid AI-Physical Modeling for Penetration Bias Correction in X-band InSAR DEMs: A Greenland Case Study
    Item type: Conference Paper
    Mansour, Islam; Fischer, Georg; Hänsch, Ronny; et al. (2025)
    2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)
    Digital elevation models derived from Interferometric Synthetic Aperture Radar (InSAR) data over glacial and snowcovered regions often exhibit systematic elevation errors, commonly termed “penetration bias.” We leverage existing physics-based models and propose an integrated correction framework that combines parametric physical modeling with machine learning. We evaluate the approach across three distinct training scenarios — each defined by a different set of acquisition parameters — to assess overall performance and the model's ability to generalize. Our experiments on Greenland's ice sheet using TanDEM-X data show that the proposed hybrid model corrections significantly reduce the mean and standard deviation of DEM errors compared to a purely physical modeling baseline. The hybrid framework also achieves significantly improved generalization than a pure ML approach when trained on data with limited diversity in acquisition parameters. 11The source code is available at https://github.com/IslamAlam/pydeepsar
  • Combining AI Techniques with Physical Models: Forest Height Inversion from TanDEM-X InSAR Data Using a Hybrid Modeling Approach
    Item type: Other Conference Item
    Mansour, Islam; Hänsch, Ronny; Papathanassiou, Konstantinos; et al. (2023)
    In the realm of artificial intelligence, specifically utilizing methodologies such as machine learning and deep learning, a conspicuous display of substantial potential across various parameter estimation problems has been demonstrated. However, such AI techniques are often employed without the incorporation of domain-specific knowledge or expertise, raising concerns about the explainability and robustness of the implemented methodologies. In contrast, physical models (PMs) offer a significantly enhanced level of deterministic robustness. However, it is imperative to recognize that these models can exhibit performance limitations owing to their inherent simplicity and/or strictness. Moreover, the accuracy of their inversion process is circumscribed by the assumptions and simplifications that underlie them, particularly those applied to the vertical reflectivity function, which are prerequisites for achieving a well-balanced inversion problem. As a result, it becomes imperative to advocate for hybrid modeling approach by the integration of AI techniques with physical models, especially in the context of forest height estimation derived from TanDEM-X coherence measurements. Accurate estimation of forest height is crucial for understanding forest structure and biomass, which in turn plays a pivotal role in climate change mitigation and ecosystem management. In this study, we propose a novel hybrid modeling approach that combines machine learning techniques and physical models to invert forest height from TanDEM-X InSAR (Interferometric Synthetic Aperture Radar) data. This approach might be relevant for the Biomass mission for understating the forest and its structures.
  • Monitoring the Subsurface of an Alpine Glacier Using Polarimetric SAR Observations at L-Band
    Item type: Conference Paper
    Parrella, Giuseppe; Farinotti, Daniel; Hajnsek, Irena; et al. (2016)
    Proceedings of EUSAR 2016: 11th European Conference on Synthetic Aperture Radar
    Polarimetric Synthetic Aperture Radar (PolSAR) sensors represent, nowadays, an established tool for the observation of glaciarized areas. At dry snow conditions, such sensors combine the penetration capability of microwaves into snow/ice and the sensitivity of polarimetry to different scattering mechanisms. Therefore, PolSAR measurements are sensitive to the surface as well as near-surface features, especially at lower frequencies. A key objective of this study is to investigate the potential to monitor the seasonal variability of surface and near-surface features of an Alpine glacier by means of multi-temporal airborne PolSAR data at L-band. The results of the polarimetric analysis are compared to ground-based Ground Penetrating Radar (GPR) and snow depth measurements performed in coordination with the SAR campaign.
  • 3-D SAR Imaging of African Forests: Results from the AfriSAR Campaign at P- and L-Band
    Item type: Conference Paper
    Hajnsek, Irena; Pardini, Matteo; Horn, Ralf; et al. (2016)
    Proceedings of EUSAR 2016: 11th European Conference on Synthetic Aperture Radar
    AfriSAR is an ESA-funded airborne P-band SAR campaign over the African tropical forests of Gabon that was carried out by ONERA (July 2015) and DLR (February 2016) in support of the development of the geophysical algorithms of the future BIOMASS mission. In addition, DLR will complement the P-band data set with L-band over the same sites. Both data sets will be constituted by properly designed multibaseline fully polarimetric acquisitions allowing the inversion of key forest vertical structure-based parameters, like e.g. forest height and high resolution reflectivity profiles. A first processing result of the acquired data and parameter inversion is shown in this paper at both frequencies.
  • Wet Snow Mapping in the Karakoram using SAR and Topographic Data
    Item type: Other Conference Item
    Li, Shiyi; Huang, Lanqing; Bernhard, Philipp; et al. (2023)
    EGUsphere
    Wet snow is a critical component of the cryosphere, and its spatial and temporal distribution has important implications for water resources, natural hazards, and the regional climate. However, mapping wet snow in alpine regions such as the Karakoram is challenging due to complex topography, harsh weather conditions, and limited in-situ observations. Previous studies have shown that synthetic aperture radar (SAR) can effectively detect wet snow surfaces using the backscattering ratio between the current and reference images (e.g. the average of summer acquisitions). However, its regional application on a large-scale and complex terrain is hampered, as the ratio value is easily affected by the land cover, local topography, surface roughness, and snow wetness. In this study, we present a new approach for mapping wet snow in the Karakoram using a combination of SAR data and topographic information. The SAR data used in the analysis were obtained from Sentinel-1, and the topographic data included a digital elevation model (DEM), slope angle, and slope aspect ratio. We first used a Gaussian Mixture Model to classify the ratio image of Sentinel-1 into wet snow (WS) and non-wet snow (NWS) classes, then transformed the two classes into a logistic function to characterize the probability of WS based on the backscattering ratio. Secondly, we categorized the image based on the topography and calculated the likelihood of WS for each topographic bin using the WS probability. The joint WS likelihood map was finally obtained by multiplying the WS probability on the backscattering ratio with the WS likelihood on topography, and a binary WS map was generated by setting a threshold on the joint likelihood map. The proposed method was validated using snow maps generated from Sentinel-2 images. Compared with the traditional method of using only the SAR backscattering ratio, our method significantly reduced false negative detections and preserved the high true positive rate, indicating an improvement of robustness and accuracy by combining SAR and topographic data for regional wet snow mapping. This study demonstrates a practical method of merging SAR backscattering features and topographic information for robust regional wet snow mapping in complex mountain ranges. It also provides new insights into the incorporation of different datasets using a probabilistic framework. With the proposed method, the operational monitoring of wet snow distribution in the Karakoram using SAR becomes feasible and reliable.
  • Model-based Tensor Decompositions for Bio- and Geophysical Parameter Retrieval
    Item type: Other Conference Item
    Basargin, Nikita; Alonso-González, Alberto; Hajnsek, Irena (2023)
    8th International Workshop on Retrieval of Bio- & Geo-physical Parameters from SAR Data for Land Applications: Book of Abstracts
  • KAPRI - the Polarimetric and Interferometric Bistatic Real- Aperture Ku-Band Radar for Environmental Information Retrieval
    Item type: Conference Poster
    Mas I Sanz, Esther; Štefko, Marcel; Hajnsek, Irena (2023)
    Radar observations that combine polarimetry and interferometry allow an enhanced characterisation of the natural environment and its processes. Fully polarimetric imaging provides data on scattering mechanisms present in the resolution cell and thus on the structure of the natural medium. Furthermore, differential interferometry allows the monitoring of displacements which occur on the line-of-sight component, focusing on changes caused by movements of the natural environment. Another key factor in radar measurements for characterization of natural processes is the revisit time: monitoring some of the Earth’s dynamic processes requires dense temporal sampling and long time series durations. For this reason, ground-based radar systems are a suitable choice for applications with high temporal sampling requirements as they can provide repetition times on the order of minutes and virtually unlimited time series [1]. In contrast, current spaceborne synthetic aperture radar (SAR) Earth Observation systems cover large areas in a single pass but their revisit time is, at best, within the region of several hours, making them unsuitable. KAPRI is a ground-based portable, frequency-modulated continuous-wave (FMCW) , Ku-band polarimetric radar interferometer. It is able to acquire both monostatic and bistatic full-polarimetric measurements with high spatial and temporal resolution. The bistatic configuration makes use of two independent KAPRI devices, one operating as the primary transmitter/receiver and the other as the secondary receiver. A new processing pipeline as well as a calibration method were developed to address these differences between bistatic and monostatic configurations [2]. The bistatic configuration is of special interest as less extensive research of its applications in the context of natural processes has been done if compared to the monostatic case, in part, due to the added challenges coming from separating the transmitter and the receiver [3]. KAPRI’s bistatic regime sets the scene for new investigations given that had not been possible by employing the monostatic mode, such as the measurement of 3-D deformation vector fields or the investigation of possible presence of unique scattering events exclusively detectable in bistatic configurations. These bistatic capabilities of KAPRI were recently leveraged for the first characterization of the coherent backscatter opposition effect (CBOE) in dry snow using a terrestrial bistatic radar system [4]. A better understanding of bistatic radar measurements, which can be efficiently provided by ground-based systems, lays the foundations for future SAR space missions and pushes forward the present knowledge on snow landscape areas and their role on climate. [1] Baffelli, S., Frey, O., Werner, C., & Hajnsek, I. (2018). Polarimetric calibration of the ku-band advanced polarimetric radar interferometer. IEEE Transactions on Geoscience and Remote Sensing, 56(4), 2295–2311. https://doi.org/10.1109/TGRS.2017.2778049 [2] Stefko, M., Frey, O., Werner, C., & Hajnsek, I. (2022). Calibration and Operation of a Bistatic Real-Aperture Polarimetric-Interferometric Ku-Band Radar. IEEE Transactions on Geoscience and Remote Sensing, 60, 1–19. https://doi.org/10.1109/TGRS.2021.3121466 [3] Moccia, G. Salzillo, M. D’Errico, G. Rufino, and G. Alberti, “Performance of spaceborne bistatic synthetic aperture radar,” IEEETrans. Aerosp. Electron. Syst., vol. 41, no. 4, pp. 1383–1395, Oct. 2005 [4] Stefko, M., Leinss, S., Frey, O., & Hajnsek, I. (2022). Coherent backscatter enhancement in bistatic Ku- and X-band radar observations of dry snow. Cryosphere, 16(7), 2859–2879. https://doi.org/10.5194/tc-16-2859-2022
  • Karakoram Glacier Elevation Change During the Recent Decade Revealed by Tandem-X Dem Series
    Item type: Conference Paper
    Li, Shiyi; Huang, Lanqing; Bernhard, Philipp; et al. (2024)
    IGARSS 2024 - 2024 IEEE International Geoscience and Remote Sensing Symposium
    The change of glacier surface elevation in Karakoram holds significant relevance to the study of global climate change and regional water resource management. In this work, we focused on understanding glacier dynamics in the Karakoram region through the analysis of Digital Elevation Models (DEMs) generated from a comprehensive 481 TanDEM-X CoSSC products spanning 2011 to 2020. An iterative approach based on the residual phase with respect to the reference DEM was employed to generate high resolution DEMs. The systematic approach offered a robust methodology for extracting precise glacier elevation changes in the Karakoram with high temporal and spatial resolution, provided valuable insights into the understanding of complex glacier dynamics.
  • A Novel Atmospheric Phase Correction Based on Kriging Incorporating Temporal Phase Evolution for Ground-Based SAR
    Item type: Conference Paper
    Izumi, Yuta; Frey, Othmar; Sato, Motoyuki; et al. (2023)
    2023 8th Asia-Pacific Conference on Synthetic Aperture Radar (APSAR)
    In ground-based synthetic aperture radar (GBSAR) interferometry applications, the generated interferograms often suffer from propagation delay in the troposphere, known as atmospheric phase screen (APS). Kriging interpolation is one of the approaches that can predict the APS in interferograms. In this presentation, we propose a novel Kriging approach for more accurate predictions of APS in multi-temporal GB-SAR data. In the proposed method, the temporal atmospheric phase evolution at each pixel is taken into account in the Kriging. The proposed method is validated with a real GB-SAR dataset, showing improved results on the prediction accuracy of APS in the spatial domain.
  • System Characterization and Polarimetric Calibration of the Ku-Band Advanced Polarimetric Interferometer
    Item type: Conference Paper
    Baffelli, Simone; Frey, Othmar; Werner, Charles; et al. (2016)
    Proceedings of EUSAR 2016: 11th European Conference on Synthetic Aperture Radar
    This paper addresses the system characterization and the polarimetric calibration of the Ku-Band Advanced Polarimetric Interferometer (KAPRI). KAPRI is an FMCW ground-based real aperture radar system that uses slotted waveguide antennas. The rotation of the antennas introduces undesired phase ramps in azimuth. We present a geometrical model to account for this phase, and propose a method to correct it. Experimental data with a set of trihedral corner reflectors (TCR) in the scene was acquired with the system. A linear phase variation of 30 degrees was observed over the TCR which was geometrically modeled and successfully corrected.
Publications 1 - 10 of 75