SAR is a remote sensing technique capable to deliver actual data at any time and under bad weather conditions. Before
launch of TerraSAR-X, RADARSAT-2, or COSMO-SkyMed, the rather coarse resolution of operational SAR satellite
systems allowed an analysis of spaceborne SAR data in case of disaster management only for medium scale products.
The new generation of spaceborne SAR satellites permits a more detailed analysis at the object level even for urban
areas, which was before restricted to airborne SAR sensors. Change detection in SAR images is an important field of
research. In general, the appearance of objects in SAR images strongly depends on the viewing angle and look direction.
This makes a comparison of images on a pixel level difficult. The changeover from pixel- to object level leads to the
possibility, to look for object-features that are more stable concerning different imaging constellations. Bridges are keyelements
of man made infrastructure. In this paper the appearance of bridges in SAR data is analyzed and features are
derived that are exploitable for change detection. Here the focus is on analysis at the object level to derive features that
are either stable concerning the imaging constellations or that can be predicted based on a given imaging constellation.
Thereby, the usage of different sensors will be possible to achieve the goal of real time information. The investigations
are supported by simulations, which allow the creation of SAR images for different imaging constellations, bridge
materials, and even for situations with destroyed bridges.
State-of-the-art SAR sensors suggest utilizing InSAR-Data for the analysis of dense urban areas. The appearance of
buildings in SAR or InSAR data is characterized by the effects of the inherent oblique scene illumination, such as
layover, occlusion by radar shadow and multipath signal propagation. Therefore, especially in dense built-up areas
reconstruction quality can be improved by a combined analysis of multi-aspect data.
The presented approach focuses on reconstruction of buildings in residential districts supported by knowledge based
analysis considering the mentioned SAR-specific effects. The algorithm of building extraction starts with the
segmentation of primitives, such as lines and edges, followed by the assembly of building hypotheses based on typical
building features like linearity and right-angularity. The subsequent post-processing of building hypotheses contains the
analysis of InSAR phases to improve footprint or to detect roof-type of buildings. The results are presented by using
optical data and a high resolution LIDAR surface model as ground truth data.
Operational SAR satellite systems such as ENVISAT-ASAR and RADARSAT-1 deliver image data of a rather coarse
resolution, which allows the recognition or feature extraction only for large man-made objects. State of the art airborne
SAR sensors on the other hand provide spatial resolution in the order well below a half meter. In such data many features
of urban objects can be identified and used for recognition. Core elements of man-made infrastructure are bridges. In
case of bridges over water, the oblique side looking imaging geometry of SAR sensors may lead to special signature in a
SAR image depending on the aspect. In this paper, the appearance of bridges over water in SAR data is discussed.
Geometric constraints concerning the changing of this signature are investigated using simulation techniques based on an
adapted ray tracing. Furthermore, an approach is presented to detect bridges over water and to derive object features
from spaceborne and airborne SAR images in the context of disaster management. RADARSAT-1 data with a spatial
resolution of about 9 m as well as high-resolution airborne SAR data of geometric sampling distance better than 40 cm
The enhancement and improvement of classifiers for SAR signatures are a permanent challenge. The focus of this paper is the development of an integrated decision-and-reject method suitable for a kernel-machine-based target classification framework for SAR scenarios. The proposed processing chain consists of a screening process identifying ROIs with target cues, a pre-processing, and a high-performance classifier. A feasible screening method has to provide a maximum of detections namely object hypotheses while the false alarm rate is of lower interest. Therefore the quality of the following classification step significantly depends on the capability of reducing the false alarms. In complex scenarios standard approaches may classify clutter objects incorrectly as targets. To overcome this problem a novel classification scheme was developed. Class discriminating information is computed in a pre-classification step by a family of two-class kernel machines. Thus, a feature vector for an additional classification stage is provided. A comparative assessment was done using a SAR data set provided by QinetiQ. First results are given in terms of ROC curves.
The improved ground resolution of state-of-the-art synthetic aperture radar (SAR) sensors suggests utilizing this technique for analysis of urban areas. However, building reconstruction from SAR or InSAR data suffers from consequences of the inherent oblique scene illumination, such as foreshortening, layover, occlusion by radar shadow and multipath signal propagation. Especially in built-up areas, building reconstruction is often hardly possible based on single SAR or InSAR data sets alone. An approach is presented to improve the reconstruction quality combining multiaspect InSAR data.
Building object primitives are extracted independently for two directions from the magnitude and phase information of the interferometric data. After projection of these initial primitive objects from slant range into the world coordinate system they are fused. This set of primitive objects is used to generate building hypotheses. SAR illumination effects are discussed using real and simulated data. The simulation results have been compared with real imagery. Deviations between simulations and real data were the base for further investigations. The approach is demonstrated for two InSAR data sets of a building group in an urban environment, which have been taken from orthogonal viewing directions with spatial resolution of about 30 cm.
Change detection plays an important role in different military areas as strategic reconnaissance, verification of armament and disarmament control and damage assessment. It is the process of identifying differences in the state of an object or phenomenon by observing it at different times. The availability of spaceborne reconnaissance systems with high spatial resolution, multi spectral capabilities, and short revisit times offer new perspectives for change detection. Before performing any kind of change detection it is necessary to separate changes of interest from changes caused by differences in data acquisition parameters. In these cases it is necessary to perform a pre-processing to correct the data or to normalize it. Image registration and, corresponding to this task, the ortho-rectification of the image data is a further prerequisite for change detection. If feasible, a 1-to-1 geometric correspondence should be aspired for. Change detection on an iconic level with a succeeding interpretation of the changes by the observer is often proposed; nevertheless an automatic knowledge-based analysis delivering the interpretation of the changes on a semantic level should be the aim of the future. We present first results of change detection on a structural level concerning urban areas. After pre-processing, the images are segmented in areas of interest and structural analysis is applied to these regions to extract descriptions of urban infrastructure like buildings, roads and tanks of refineries. These descriptions are matched to detect changes and similarities.
The focus of this paper is the classification of military vehicles in multi-polarimetric high-resolution spotlight SAR images in an ATR framework. Kernel machines as robust classification methods are the basis of our approach. A novel kernel machine the Relevance Vector Machine with integrated Generator (RVMG) controlling the trade-off between classification quality and computational effort is used. It combines the high classification quality of the Support Vector Machine by margin maximization and the low effort of the Relevance Vector Machine caused by the special statistical approach. Moreover multi-class classification capability is given by an efficient decision heuristic, an adaptive feature extraction based on Fourier coefficients allows the module to do real time execution, and a parameterized reject criterion is proposed in this paper.
Investigations with a nine class data set from QinetiQ deal with fully polarimetric SAR data. The objective is to assess polarimetric features in combination with several kernel machines. Tests approve the high potential of RVMG. Moreover it is shown that polarimetric features can improve the classification quality for hard targets. Among these the simple energy based features prove more favorable than complex ones. Especially the two coplanar polarizations embody the essential information, but a better generalizability is caused by using all four channels.
An important property of a classifier used in the ATR framework is the capability to reject objects not belonging to any of the trained classes. Therefore the QinetiQ data are divided into four training classes and five classes of confusion objects. The classification module with reject criterion is controlled by the reject parameter and the kernel parameter. Both parameters are varied to determine ROC curves related to different polarimetric features.
Ground surveillance and target recognition by radar has become increasingly important over the years. Modern digitally controlled radar systems have the ability to operate quasi simultaneously in two or more different modes, e.g. after detection of moving targets by MTI these target hypotheses are recorded by a high-resolution spotlight SAR. To classify the SAR signatures different techniques have been investigated. The objective of our work was to support the decision process in choosing the best combination of methods for the problem of ground target classification in high-resolution SAR images. The criteria of optimizing the classification are correctness (low false alarm rate (FAR)), robustness, and computational effort. The investigations have been carried out using the MSTAR public target dataset. In the paper we describe the examination of new classifier approaches like support vector machine (SVM) and relevance vector machine (RVM) in combination with superresolution methods like the CLEAN algorithm. For this purpose we have developed an experimental software system. Its processing chain consists of the following modules: preprocessing, feature extraction, and classification. The tests with the SVM have shown that without preprocessing too many support vectors (up to 50 %) are used. Therefore the RVM has been chosen to overcome this disadvantage. The preprocessing methods have been used to reduce the noise and to restore / extract the significant SAR signature. The result of our investigations is an assessment of the different methods and several method combinations. Based on these results the investigation will be extended by more realistic new datasets with a resolution as high as or higher than the MSTAR data.
The extensive knowledge of scenes is necessary for change detection or mission planning. Unfortunately, InSAR images have zones with full or partial loss of information, e.g. in shadow areas. We have developed an algorithm for registration and fusion of multi-aspect InSAR images to replace shadow zones in the image of the first aspect by corresponding data taken from another aspect. The investigations were carried out with using of X band images obtained for two aspects. The two aspect images were recorded from two contrary flight courses. The images were delivered as intensity and unwrapped phase of forest, rural and urban terrain. The registration is based on the consecutive application of three separate matching algorithms: matching straight lines, matching contour sequences, and matching based on the Fourier-Mellin transform. The different matching algorithms are optimized for working with the SAR images having different contents. The main peculiarities of the registration algorithm are the methods of structural matching brightness contours calculated by standard edge filters. The shadow borders are excluded from the gradient field by a heuristic algorithm. The contours are extracted by a watershed- or a maximum gradient tracking algorithm. The matching is very robust. It is rather computationally expensive, thus a hierarchical structural matching is used to decrease the computational complexity. The multi-stage contour matching provides a fast and reliable registration and fusion.
Future conflicts will probably lead the armed forces also into regions, for which topographical data are missing as well as out of date. Here IMINT must support the planning and transaction of military operations through improved target recognition in combination with topographical information.
High-resolution LIDAR data, multi-spectral image data and GIS with orthorectified elevation data, combined with 3Dimage maps with high geometrical and spatial precision integrated in a network (Smart Sensor Web, SSW), open new additional possibilities of the reconnaissance.
To improve reconnaissance, we investigate the 3D-modeling of built up areas including texturing and visualization for the observer. In a future joint-sensor system the information of several sensors should be used in common and should also be combined with non-imaging knowledge (Rapid Terrain Visualization, RTV). By this, the technology is a key technology for military applications in urban warfare and in the battle against terrorism.
The improved quality of InSAR data suggests to utilize such data for analysis of urban areas. But, the phase information from which the height data is calculated, is often severely disturbed, depending on the signal to noise ratio. As a consequence, irregular height jumps occur even inside flat objects. In this paper we refer to investigations to stabilize and improve the InSAR height data. After preprocessing, a segmentation is carried out in the intensity and the height data. Inside the extracted segments the height data is smoothed, using the related intensity or coherence values as weights. For every segment the weighted average height is calculated. Preliminary hypotheses for buildings are identified a by significant height over surrounding ground. In a post-processing step, the intermediate results are analyzed and corrected due to a possible over- and under-segmentation. Adjacent objects with similar heights are merged and objects including shadow areas are split. The shadow areas are detected by structural image analysis in a production net environment exploiting collateral information, like sensor position and depression angle. The derived 3D information may be used for visualization or map update tasks. A test site including the airport of Frankfurt (Main) was chosen. For the visualization purpose, a 3D view of the smoothed height data is shown. The results are compared to a map and differences are depicted and discussed.
The SAR processing is optimized for motionless scenes. Moving objects cause artifacts like blurring or azimuth displacement in case of parallel or radial velocity components respectively. With along-track interferometry (SAR-MTI) even very slow radial velocities can be measured by the phase differences. Unfortunately, the phase information is often severely disturbed, depending on a insufficient signal to noise ratio. In this paper we refer to investigations to stabilize and improve the SAR-MTI velocity data. Reliability is enhanced by a combined exploitation of phase and intensity. After speckle filtering a binary mask is generated from the intensity data to fade out regions with insufficient signal to noise ratio, like regions with low backscattering coefficient. In a next step for every point in the intensity image the radial velocity is calculated by the phase difference of two channels. This image of velocities is masked with the binary mask derived from the intensity image. A region growing process is initiated in the velocity image to identify connected regions in the image with similar velocity. By this process we get first hints for moving objects. The approach was applied to images with slow moving cargo ships inside and nearby locks. The cargo ships are segmented and described by a simple model. Only cargo ships with a minimum velocity which match the longitudinal and transversal extension features of the model are concerned in further processing.
In image reconnaissance the analyst of remotely sensed imagery is confronted with large amounts of data. Especially the integration of multi-sensor data calls for support of the observer by automatic image processing algorithms. For this purpose we recently developed model based structural image analysis algorithms which deliver successful results. However, varying scenarios, different applications and changing image material often require a tuning of the algorithms. Therefore, we suggest techniques to support and automate the adaptation of the image processing to changing requirements. Our approach uses techniques form data mining to discover relationships between image properties and optimal parameter vectors. This paper addresses two points: a supervised tuning approach and suggestions for unsupervised tuning. For the supervised tuning a representative image database was set up, and a corresponding ground truth was interactively defined. The results of the structural image analysis for a set of parameters can be compared to the ground truth. For the example images the parameters were optimized using an evolutionary optimization loop. For the unsupervised tuning the data form the supervised optimization is analyzed. We present promising results form manual clustering and propose a clustering approach based on decision trees, and hierarchical and evolutionary cluster algorithms with different distance measures.
Geocoding based merely on navigation data and sensor model is often not possible or precise enough. In these cases an improvement of the preregistration through image-based approaches is a solution. Due to the large amount of data in remote sensing automatic geocoding methods are necessary. For geocoding purposes appropriate tie points, which are present in image and map, have to be detected and matched. The tie points are base of the transformation function. Assigning the tie points is combinatorial problem depending on the number of tie points. This number can be reduced using structural tie points like corners or crossings of prominent extended targets (e.g. harbors, airfields). Additionally the reliability of the tie points is improved. Our approach extracts structural tie points independently in the image and in the vector map by a model-based image analysis. The vector map is provided by a GIS using ATKIS data base. The model parameters are extracted from maps or collateral information of the scenario. The two sets of tie points are automatically matched with a Geometric Hashing algorithm. The algorithm was successfully applied to VIS, IR and SAR data.
The operation of high resolution sensors can enhance the reconnaissance capability, but as a consequence it is accompanied by an increasing amount of data. Therefore, it is necessary to relieve the data down link and the image analyst, e.g. by a screening process. This task filters out regions of interest (ROI) and recognizes hypotheses of potential objects. In this paper, we describe the support of image analysis by thematic vector data from a GIS. First, the image is automatically or interactively geocoded based on matched objects in the image and map. Three different ways on the use of GIS depending on its information content are presented. In the first approach, the desired ROIs are selected by a thematic query to the GIS and then extracted from the image. Inside these ROIs a subsequent detail analysis can be performed. In the second case in which objects are not integrated in the vector data or interesting objects are not maintained, a structural image analysis approach is used to detect extended objects like, for example, airfields. The GIS is then supplemented by the result of structural image analysis. In the third case, thematic vector data is used to extract training areas for classification and parameter settings for image segmentation.
This contribution presents a comprehensive framework for algorithm evaluation. When we speak of evaluation, we have in mind that first the performance of an algorithm is measured and then the measured performance is assessed with regard to a given application. The performance assessment is done by applying an assessment function that uses desired values for the performance measures and weighting factors giving the importance of each measure, thus considering the application- specific requirements. The algorithm evaluation's goal is to verify the specification of an algorithm. This specification is mainly given by the definition of the input data and the expected output data, both of which are determined by the application. Prior to the evaluation process the algorithm specification has to be laid down by analyzing the application in order to deduce its requirements as well as by defining the application relevant data sets. To organize this sequence of preparatory steps and to formalize the accomplishment of the evaluation we have developed a 3-phase approach, consisting of the definition phase, the tuning phase, and the evaluation phase. An extensive software toolbox has been developed to support the evaluation process.
Screening and thematic data exploitation are important tasks in the reconnaissance cycle. To reduce the work load of the image analysts, an automatization of these tasks is required. For an automatic image exploitation structural image analysis algorithms were developed for multisensor data. The approach is based on the primitive objects line which are generated by appropriate edge detectors. For preprocessing in SAR images, i.e. the change from the iconic to the symbolic level, not only speckle noise prejudices automatic segmentation but also multiple scattering of hard targets in the scene. Several different approaches to filter speckle noise and to detect edges prior to structural image analysis have been investigated. The structural analysis of complex scenes uses the blackboard-based production system (BPI) as framework. The primitive objects are built up step by step applying the productions. For an IMINT report the image data has to be referenced to a cartographic grid. In an automated reconnaissance cycle an automatic image-to-map registration is required. The necessary control points must be detected in the image data and a correspondence between map and image control points must be found. The transformation parameters can then be calculated. Additionally, when using map information, expectation areas can be defined and processing needs can be reduced efficiently.
It is known that the performance of recognition by humans is improved by time integration. We have investigated this phenomenon in image sequences. The object hypotheses are detected in the image sequence utilizing object regions and motion in analogy to human perception. A multiple thresholding segmentation and a change detection process by wavelet transformation are used for detection. The detected segments are tracked over time. Our classification approach assumes, that the objects are recognized as a connected entity. Therefore a structural object description derived from the image sequence was developed. This description contains the geometric relations and shape features of the individual object parts as well as the motion behavior. Normalized difference measures of the structural descriptions have been derived for the classification. The differences are determined and combined by a fuzzy approach. The results have shown, that the classification can be improved and stabilized by the object description derived from the image sequence.