Combining machine learning (ML) with physical models can significantly impact retrieval algorithms designed to invert geophysical parameters from remote sensing data. Such hybrid models integrate physical knowledge with domain expertise through a joint architecture, potentially enhancing performance by increasing the efficiency and flexibility of the physical model as well as the generalization and interpretability of the ML predictions. This work introduces a hybrid model for estimating forest height using single-baseline, single-polarization TanDEM-X interferometric coherence measurements. In this model, the vertical reflectivity profile is derived as a function of input features, including topographic and acquisition geometry descriptors, using a multilayer perceptron network. This profile is then used to invert forest height by leveraging the established physical relationship connecting the vertical reflectivity profile to forest height. The developed model is applied and validated on several TanDEM-X acquisitions over tropical sites with different acquisition geometries, and its performance is assessed against reference data derived from airborne LiDAR measurements.

Deep learning models trained in a fully supervised way have shown encouraging capabilities for mapping forests with TanDEM-X interferometric data, being able to generate time-tagged forest maps at large-scale over tropical forests. These maps have been generated at 50 m resolution to reduce the computation burden. In this work, we now aim to exploit the high-resolution capabilities of the TanDEM-X interferometric dataset, processed at only 6 m resolution. In order to cope with the lack of reliable reference data at such high resolution, we focus on the investigation of self-supervised learning approaches. The availability of a reference map over Pennsylvania, USA, based on Lidar acquisitions at 1 m resolution, allows us to compare different deep learning approaches. First promising results show the possibility to extend the proposed self-supervised learning approach over areas where the lack of reference data prevent us from using fully supervised deep learning methods.

Floods represent a pervasive natural hazard with global ramifications, impacting a vast population and resulting in substantial property damage and severe mortality. Particularly worrisome is their disproportionate effect on the least developed countries, which exacerbates developmental imbalances, posing a significant obstacle to the attainment of the United Nations Sustainable Development Goals (UN SDGs). This paper introduces the AI-powered Flood Mapathon activity, co-organized by the Aerospace Information Research Institute under the Chinese Academy of Sciences, in partnership with GEOVIS Technology Co., Ltd., GEOVIS Earth Technology Co., Ltd., and IEEE GRSS IADF. The activity seeks to mobilize individuals worldwide to address the most prevalent natural hazard-floods by collaboratively mapping inundated regions through the analysis of satellite imagery. Gaining widespread attention, the activity has garnered 30,755 submissions from 310 participants across 34 countries. Through collective efforts, participants have curated a semantic segmentation dataset focusing on floods, incorporating annotations of pertinent features related to both floods and human activities. Additionally, the paper elucidates the custom crowdsourcing mapping system, which seamlessly integrates cutting-edge AI technologies to alleviate mapping complexities. The activity contributes to sustainability by drawing extensive public attention, creating a public flood dataset for academic research, and establishing an efficient and intelligent mapping system.

Computer vision algorithms are increasingly leveraged to accelerate geospatial analysis for disaster response and recovery. As the diversity of remote sensing imagery grows with optical, SAR, and other modalities, a perquisite for analytics is cross-modal image registration. There is a high potential to harness computer vision for this pre-processing requirement toward enabling downstream analytics such as heterogeneous change detection, automated feature extraction, and data fusion. Advancement in these areas has the potential to simplify data wrangling tasks and further accelerate disaster response timelines. The SpaceNet 9 challenge (launching in mid-2024) focuses on addressing the cross-modal image registration problem and demonstrating the utility of such modules on earthquake impacted scenarios. This paper describes the motivation for the SpaceNet 9 and provides a first overview of the dataset, the baseline algorithm, and implications for seeking cross-modal image registration in Earth observation. Code is available at https://github.com/SpaceNetChallenge/SpaceNet9.

Climate change has a multitude of direct and indirect effects on natural habitats, availability and accessibility of resources, human life and communities, and economy. Earth observation data plays a crucial role in informing about these effects, observing and modeling the corresponding processes, identifying corresponding drivers, and monitoring and forecasting climate-related events including natural disasters such as droughts, wildfires, and floods. Machine learning aids in processing and analyzing the multitude of available data to automate the creation of informative products that are of use to local communities and authorities and first responders to develop efficient mitigation strategies and observe the success or failure of corresponding measures. This paper provides an overview of the current state of the art of the different aspects of how the combination of Earth observation data and machine learning help addressing climate-change related issues.

Rapid changes in the Greenland Ice Sheet require precise elevation monitoring to understand ice dynamics and predict sea level rise. X-band Interferometric Synthetic Aperture Radar (InSAR) has the potential for this purpose but is limited by microwave signal penetration biases, which can be a few meters. We present a novel hybrid modeling approach that integrates machine learning (ML) with physical models to enhance the estimation of the elevation bias in InSAR data at X-band. Our method addresses the limitations of traditional physical modeling techniques by parameterizing the vertical structure function using a ML model. This approach combines machine learning as input for the physical model. The results demonstrate the improvements in correcting elevation biases, thus increasing the accuracy of X-band InSAR DEMs over Greenland. This advancement has the potential for more precise elevation estimation and ice-sheet monitoring.

Excelling in various image analysis tasks, machine learning (ML) models and especially deep convolutional networks (ConvNets) have become a cornerstone in the Remote Sensing community. However, their complexity makes their decision-making process opaque, rendering deep ConvNets as black box models. To address this issue, "Explainable AI" (XAI) methods have been proposed that aim to provide insights into the rationale behind ML generated predictions. Amongst them, perturbation-based techniques monitor changes in the prediction related to local distortions of the input. Thereby the relative importance of the altered input area for the prediction is determined that serves as an explanation for the network’s prediction. In the context of flood detection from SAR images, we investigate the impact of different parameter settings on the relevance estimation and thus on the explanation. The experimental results indicate a strong parameter dependence yielding ambiguous and partly contradicting explanations.

Accurate forest height estimates lead to improved accuracy of biomass estimation and are crucial for monitoring and conservation efforts. Interferometric synthetic aperture radar (InSAR) techniques use two synthetic aperture radar (SAR) images to measure the interferometric coherence that includes the volumetric decorrelation which is known to be related to forest canopy height. Several approximations and assumptions are made in different steps to compute volumetric decorrelation and to invert it to forest canopy height using physical models. Data-driven approaches overcome the potential bias introduced by these assumptions by directly estimating forest canopy height. However, the question of optimal representation and level of processing of the input data is often neglected. We address this gap comparing different SAR and InSAR input features such as single-look-complex (SLC) images, backscatter, coherence, and volumetric decorrelation. The resulting best model has a root-mean-squared error (RMSE) of 6.12 m with volumetric decorrelation as primary input feature. It is followed using coherence as primary input with an RMSE of 6.30 m.

Synthetic aperture radar tomography (TomoSAR) uses several coregistered images from different perspectives to reconstruct a power spectrum pattern (PSP) perpendicular to the line of sight (PLOS), enabling the estimation of a 3-D representation of the area. Classical estimators exhibit ambiguities and other undesired effects that are stronger for sparser and smaller stacks. To mitigate the limitations arising from a restricted number of acquisitions, we propose using a deep neural network (NN) to synthesize artificial tracks (i.e., images not contained in the original stack). The presented method utilizes a convolutional NN with an encoder-decoder architecture. We evaluate the proposed approach on real TomoSAR data from an airborne campaign over a forest region. The view estimation improves the tomographic results, offering robustness to scenarios affected by temporal decorrelation, which other classical methods, such as cubic convolution (CC), do not provide.

The automatic prediction of semantic maps, e.g. for land use/cover, is often addressed via supervised learning approaches that rely on large training datasets. Obtaining the required reference annotations for a large amount of data is tedious and time consuming work, in particular for sensors such as synthetic aperture radar. Self-supervised learning aims to mitigate this issue by leveraging the large amount of usually available unlabelled data by constructing a pretext task that does not require annotations. We evaluate DINO, a recent self-supervised learning approach, and show that it outperforms segmentation models that have been initialized randomly or by weights obtained from pretraining on ImageNet. Furthermore, we show that the inclusion of multimodal data during the pretraining phase improves results significantly.

The development of algorithms to assess the effects of natural disasters plays an integral role in response efforts. There is a growing opportunity to leverage remote sensing data and computer vision to quickly analyze the scale of damage and organize a humanitarian response when extreme weather events occur. By automating the process of identifying damage to roads and infrastructure, we can significantly reduce response time, directing relief efforts on a time scale of minutes or hours rather than days. The SpaceNet 8 challenge featured a complex multi-class segmentation problem in the context of flood detection from remote sensing imagery. Competitors were tasked with leveraging both pre- and post-flooding event imagery to detect buildings and roads, as well as identify which of these object instances were affected by the flooding event. We examine the outcome of the SpaceNet 8 challenge and present an overview of the competition and a deeper look at the top-performing submissions.

With artificial intelligence breakthroughs permeating the Earth Science domain, there is an immediate need to advance the data, tools, and resulting technologies to broader societal challenges. Different efforts are emerging with fragmented best practices for making Earth Observation (EO) data Artificial Intelligence (AI)-ready, availing computer vision and image analysis tools for broader reuse across the remote sensing community. This paper will revisit current best practices and outline a guideline for advancing EO data and derivative AI products for broader community use. We mainly discuss the Analysis Ready Data (ARD) essentials and aim to forge their evolution with Findable, Accessible, Interoperable, Reusable (FAIR) principles to support cross-modal/cross-sensor/cross-provider opportunities that appear to be central to solving complex EO challenges.

Accurate estimation of forest height plays a vital role in forest cover mapping and monitoring logging activities. In this work, a deep learning-based methodology is explored to measure forest canopy height from InSAR products and Single Look Complex SAR images (SLCs) to fully capture the complex information available in radar data and minimize the error in forest height estimates caused by various processing steps. The proposed deep-learning model consists of an U-net architecture that is used to regress the forest height from different input features. LiDAR LVIS measurements serve as reference data. This framework is used to assess the impact of different input features that are at different processing levels and their effect on model performance. We use TanDEM-X SAR images and LiDAR data from the AfriSAR campaign over Gabon, Africa. The results show the potential of the proposed approach in achieving more accurate forest height estimates if data is represented as complex coherence instead of as volume decorrelation or SLCs.

Agricultural field polygons within smallholder farming systems are essential to facilitate the collection of geo-spatial data useful for farmers, managers, and policymakers. However, the limited availability of training labels poses a challenge in developing supervised methods to accurately delineate field boundaries using Earth Observation (EO) data. This letter introduces an open data set for training and benchmarking machine learning methods to delineate agricultural field boundaries in polygon format. The large-scale data set consists of 439,001 field polygons divided into 62 tiles of approximately 5×5 km distributed across Vietnam and Cambodia, covering a range of fields and diverse landscape types. The field polygons have been meticulously digitized from satellite images, following a rigorous multi-step quality control process and topological consistency checks. Multi-temporal composites of Sentinel-2 (S2) images are provided to ensure cloud-free data. We conducted an experimental analysis testing a state-of-the-art Deep Learning (DL) workflow based on fully convolutional networks, contour closing, and polygonization. We anticipate that this large-scale data set will enable researchers to further enhance the delineation of agricultural fields in smallholder farms and to support the achievement of the Sustainable Development Goals (SDG). The data set can be downloaded from https://doi.org/10.17026/dans-xy6-ngg6.

Datasets play a key role in developing superior building detection approaches. However, most of the previous work focuses on accurate building masks and scale expansion, while the categories are always missing, which hinders the further analysis of urban development and cultures. Therefore, we propose a benchmark for building detection and fine-grained classification from very high-resolution (VHR) satellite imagery. An extensive annotation is performed for about 0.5 million building instances with 12 fine-grained roof types and individual polygons. The annotation of building functions of two cities in the previous version (UBCv1) [1] is also integrated. To ensure the building variety, it consists of VHR optical images of 20 unique cities worldwide with various landforms and styles of architecture. Its variety and fine-grained categories pose great challenges and meanwhile provide a foundation for the building extraction and fine-grained classification on a global scale. Besides, 17 cities are provided with finely aligned Synthetic Aperture Radar (SAR) images, which can be employed for the development and evaluation of approaches optionally based on optical, SAR, or multi-modal images. Significantly, the proposed benchmark is used as the base of the 2023 IEEE GRSS Data Fusion Contest [2]. The dataset and codes of the baseline methods are available at: https://github.com/AICyberTeam/UBC-dataset/tree/UBCv2.

In the era of deep learning, annotated datasets have become a crucial asset to the remote sensing community. In the last decade, a plethora of different datasets was published, each designed for a specific data type and with a specific task or application in mind. In the jungle of remote sensing datasets, it can be hard to keep track of what is available already. With this paper, we introduce EOD - the IEEE GRSS Earth Observation Database (EOD) - an interactive online platform for cataloguing different types of datasets leveraging remote sensing imagery.

Floods are one of the major types of natural disasters responsible for loss of life, destruction of buildings and infrastructure, erosion of arable land, and environmental hazards around the world. Climate change, increasing populations, and urbanisation of flood plains will only increase the risk of flooding in the next few years. SpaceNet 8 presents a dataset that combines building footprint detection, road network extraction, and flood detection covering 850km2, including ~32,000 buildings and ~ 1,300 km of roads, of which ~ 13% and ~ 15% are flooded, respectively.

Object recognition in synthetic aperture radar images is a well studied topic that has gained a significant amount of attention within the last decades. Modern approaches are based on machine learning, i.e. deep learning, and often show excellent performance. What is so far missing in the literature is a study dedicated to the generalization capabilities of object recognition approaches, i.e. how well a given system can be transferred to new and previously unseen data. In this paper, the proposed recognition model is trained and tested on a unique dataset of 25 high-resolution TerraSAR-X images (X-band), acquired over four different airports in Staring Spotlight mode. We show how classification performance changes for different application scenarios which require different training and evaluation setups.

Random ferns—as a less known example of ensemble learning—have been successfully applied in many computer vision applications ranging from keypoint matching to object detection. This article extends the random fern framework to the semantic segmentation of polarimetric synthetic aperture radar images. By using internal projections that are defined over the space of Hermitian matrices, the proposed classifier can be directly applied to the polarimetric covariance matrices without the need to explicitly compute predefined image features. Furthermore, two distinct optimization strategies are proposed: the first based on preselection and grouping of internal binary features before the creation of the classifier and the second based on iteratively improving the properties of a given random fern. Both strategies are able to boost the performance by filtering features that are either redundant or have a low information content and by grouping correlated features to best fulfill the independence assumptions made by the random fern classifier. Experiments show that results can be achieved, which is similar to a more complex random forest model and competitive to a deep learning baseline.

Annotated datasets have become one of the most crucial preconditions for the development and evaluation of machine learning-based methods designed for the automated interpretation of remote sensing data. In this paper, we review the historic development of such datasets, discuss their features based on a few selected examples, and address open issues for future developments.

Many benchmark datasets in remote sensing still consist of only a single image that is used to train and evaluate machine learning algorithms. Random sampling - while being the virtual standard in the field to divide the available data into train and test set - is the worst choice in this scenario as it leads to a large degree of spatial correlation between these two sets which are assumed to be independent. Already pointed out in several earlier works, this paper reconfirms this fact and compares several alternatives. Two of the evaluated techniques allow to draw a large amount of unbiased test samples and lead to significantly less biased error estimates. The urgent recommendation is using multi-image datasets or - if only a single image is available - refraining to use random sampling and instead rely on either of these or similar methods.

With the increasing importance of monitoring urban areas, the question arises which sensors are best suited to solve the corresponding challenges. This letter proposes novel node tests within the random forest (RF) framework, which allows them to apply them to optical RGB images, hyperspectral images, and light detection and ranging (LiDAR) data, either individually or in combination. This does not only allow to derive accurate classification results for many relevant urban classes without preprocessing or feature extraction but also provides insights into which sensor offers the most meaningful data to solve the given classification task. The achieved results on a public benchmark data set are superior to results obtained by deep learning approaches despite being based on only a fraction of training samples.

The estimation of soil permittivity under fully covered grassland is a challenging task that can be approached by either model-based polarimetric decomposition techniques or data-driven machine-learning (ML) methods. In this study, we test the benefits and limitations of those techniques when individually or jointly applied to estimate the permittivity of the top soil (lower than 5-cm depth) from the L-band full-polarimetric SAR data. Training (needed for the ML approach) and reference data for accuracy assessment are based on the soil-permittivity measurements from an in situ sensor network. The applied polarimetric decomposition approaches are unable to estimate high soil-permittivity ranges (permittivity higher than 25-30) under the full-cover grassland leading to an underestimation compared with the in situ values. Purely data-driven ML techniques, here a case-adapted Random Forest (RF) architecture, applied directly to the SAR data achieve similar results as the decomposition approaches, where estimation quality mostly depends on the quality of the training set. The combination of both techniques works best and is able to represent high soil-permittivity ranges. The joint estimation decreases the mean absolute error drastically compared with applying any of the two approaches alone (i.e., from 4.88 and 4.99 of an informed physical model and ML on SAR only to 3.35).

Stacked Random Forests (SRFs) sequentially apply multiple Random Forests (RFs) where each instance uses the estimate of the predecessor as additional input to further refine the prediction. They have been shown to improve the performance for semantic segmentation of Polarimetric Synthetic Aperture Radar (PolSAR) images. Both, RFs and SRFs, not only provide an estimate of the class label of a query sample, but instead make a probabilistic prediction, i.e. provide the full class posterior. The probabilistic predictions of RFs are known to be usually well calibrated (i.e. the predictions match the expected probability distributions of each class). This paper answers the question whether stacking leads to overfitting on the training data or decreases the calibration quality of RFs. Results indicate that neither is the case. Instead, classification accuracy steadily increases and then saturates quickly after only a few stacking levels. The predicted probabilities are generally well calibrated where calibration quality also increases slightly for higher stacking levels.

This paper presents first results for an automatic interpretation of SAR images of sea ice acquired in the Davis Strait off the coast of Baffin Island in 2019. While the study provides multi-frequency and interferometric data collected by the DLR F-SAR airborne SAR sensor, we focus on the analysis of C-band as one of the most commonly used frequencies for sea ice monitoring. We apply an iterative version of polarimetric k-Means which allows to work on the local variance-covariance matrices directly. The obtained clusters show very distinct polarimetric as well as topological properties, which indicates that they are closely related to different sea ice types.

Due to the increasing amount of remotely sensed data, methods for its automatic interpretation become more and more important. Corresponding supervised learning techniques, however, strongly depend on the availability of training data, i.e. data where measurements and labels are provided simultaneously. The creation of reference data for large data sets is very challenging and approaches addressing this task often introduce a significant amount of label noise. While other works focused on the influence of label noise on the training process, this paper studies the impact on the evaluation and shows that the corresponding effects are even more adverse.

The deployment of numerous air- and space-borne remote sensing sensors as well as new data policies led to a tremendous increase of available data. While methods such as neural networks are trained by online or batch processing, i.e. keeping only parts of the data in the memory, other methods such as Random Forests require offline processing, i.e. keeping all data in the memory of the computer. The latter are therefore often trained on a small subset of a larger data set that is hoped to be representative instead of exploiting the information contained in all samples. This paper shows that Random Forests can be trained by batch processing too making their application to large data sets feasible without further constraints. The benefits of this training scheme are illustrated for the use case of land-use classification from PolSAR imagery.

Random Forests (RFs) are a powerful machine learning technique used for various applications including classification, regression, clustering, and manifold learning. The interpretation of a given Random Forest usually relies on statistical values, such as the distribution of path length, leaf impurity, leaf size, etc. All those measures focus on specific aspects and are incapable to provide a holistic understanding of the RF. In this paper, we propose a two-dimensional, easy-to-grasp visualization technique that follows a botanical approach and illustrates several key parameters necessary to understand why a given RF performs in a certain way. The method allows customized mappings of RF characteristics to visual properties and provides the possibility to interactively analyze the forest structure. This allows to determine trees that perform extraordinarily well or bad, to analyze the reasons for their performance, and thus to gain insights into how to change parameter setting to increase performance or efficiency.

While previous work primarily focused on using Tomographic Synthetic Aperture Radar (TomoSAR) data to analyze the 3D structure of the imaged scene, we study its potential for the generation of semantic land cover maps in a supervised framework. We extract different features from the covariance matrices of a tomographic image stack as well as from the tomograms computed by tomographic focusing. To assess the impact of our approach, we compare our results to classification maps obtained from a fully polarimetric image. We show that it is possible to outperform classification results from polarimetric data by carefully designing hand-crafted features which can be extracted either from multi-baseline single polarization covariance matrices or from tomograms obtained after tomographic focusing. Our experiments show a significant gain in the classification accuracy, especially on challenging classes such as heterogeneous city and road.

Random Forests have been shown to able to be applied directly to polarimetric synthetic aperture radar (PolSAR) data instead of to extracted hand-crafted features by adapting the internal node tests. This paper investigates different polarimetric distance measures and their potential to be used by Random Forests for the classification of PolSAR images. The experiments show that using distance measures tailored towards the statistics of PolSAR data outperforms the usage of individual hand-crafted polarimetric features and their combination. However, the differences between accuracies obtained by different suitable distance measures are insignificant allowing to take other aspects into consideration such as computational efficiency.

The automatic generation of semantic maps from remotely sensed imagery by supervised classifiers has seen much effort in the last decades. The major focus has been on the improvement of the interplay between feature operators and classifiers, while experimental design and test data generation has been mostly neglected. This paper shows that sampling strategies applied to partition the available reference data into train and test sets have a large influence on the quality and reliability of the estimated generalization error. It illustrates and discusses problems of common choices for sampling schemes, i.e. the violation of the independence assumption and the illusion of the availability of global knowledge in the training data. Furthermore, a novel sampling strategy is proposed which circumvents these problems and achieves a less biased estimate of the classification error.

With increasing availability of different sensors for earth observation, data fusion gained more and more importance. While previous publications focussed on new sensor combinations, new fusion techniques, or new applications, this work investigates at which stage of the image analysis pipeline the fusion process is most beneficial. The fusion of an optical and a SAR image for the task of land cover classification serves as an example. The experimental results indicate, that although the fusion of complementary data is generally advantageous, it is most helpful at later stages of the classification process.

One of the major problems of keypoint-based alignment of SAR and optical images is that keypoint operators react to very different object structures in both image types. This leads to a small mutual overlap in the corresponding sets of keypoints. This paper proposes to cast the task of keypoint detection as a classification problem. A machine-learning based classifier is trained to predict whether a SAR image pixel corresponds to a keypoint in the optical image or not. Experimental results indicate, that the mutual overlap of keypoints can be doubled by the proposed approach.

Random Forests and their many variations developed to one of the most successful instruments to automatically analyse image data. One of the most crucial parts is the definition and selection of node tests within the individual trees, which among other things allow for trade-offs between accuracy and computational load. This paper discusses several different approaches to test creation and compares them based on their classification performance on polarimetric synthetic aperture radar data. The experiments show that selecting the best out of multiple randomly generated node tests leads to the highest accuracy with the smallest computational effort.

Segmentation of Synthetic Aperture Radar (SAR) images is often only understood as the partitioning of the image into rather small regions which are homogeneous with respect to scattering processes. This paper proposes an adaption of the graph-cut image segmentation framework to the unique characteristics of polarimetric SAR images by using a Wishart-distribution based distance measure for local segmentation cues and simple, real-valued features derived from the complex-valued coherency matrix. The proposed method is evaluated on different polarimetric SAR images, for different objects of interest, and with a wide range of parameters. The results show that the proposed framework is able to derive accurate object/non-object segmentations. Best results are obtained for forest areas by usage of a log-transform of the polarimetric intensities.

Building detection from Synthetic Aperture Radar (SAR) images states a particular important as well as difficult problem. The high-resolution which is necessary to distinguish single buildings as well as the geometric and di-electric properties of dense urban areas cause most assumptions to fail, that are commonly made in SAR data analysis. This paper proposes the usage of Random Forests for building detection from high-resolution Polarimetric Synthetic Aperture Radar (PolSAR) imagery. Random Forests can handle high-dimensional input and therefore a large set of different features, they are known to lead to good classification performance in terms of robustness and accuracy, and are nevertheless seldomly applied to analysis of PolSAR images in general and building detection in particular. This paper presents first results of Random Forests when applied to a building detection task and shows their successful applicability.

Supervised learning algorithms are important methods to automatically interpret image data in general as well as PolSAR data in particular. However, they suffer from the need of a training set, which has to contain manually labelled data. Un-supervised methods do not demand this kind of data, but cannot be directly used to assign user-defined class labels to image regions. This paper proposes a semi-supervised method to overcome both shortcomings. The data is analysed by an un-supervised clustering algorithm under the usage of all available information. Simultaneously each pixel is classified by a supervised method using the information available at the current phase of clustering.

This paper presents a distributed approach for time- domain focusing, which significantly enhances the overall efficiency by distributing the computational load across a (potentially large) number of networked computers. The system described includes the so-called master, responsible for pre-processing, the distribution of fragments of raw-data data and the collection of processed image fragments. Fragments of raw-data are passed to so-called slaves, any number of which can be connected to the master, which are responsible for the focusing itself. Master and slave actively communicate over the network to organise the entire process in a scalable manner. In this way, time-domain processing can be accelerated by a factor that is virtually linear in the number of participating slaves. This paper summarises the current status of software development, realised in a platform-independent way using the IDL and Java languages. Additionally, some preliminary evaluations of performance, scalability and the required network infrastructure are given. Some examples of SAR data, acquired by the airborne sensor E-SAR of DLR, and processed with the system described are shown.