Knowledge about the location of a defect is essential for damage assessment. In terms of a digitalised maintenance, inspection data is combined with position information. The presented approach regards the manual ultrasonic inspection, where the ultrasonic probe and the inspected component are both hand-held. By using markerless tracking technologies, it is possible to track the component without any markers. The ultrasonic probe is tracked by a more stable marker-based tracking technology. This results in a hybrid tracking system, which allows a referencing of the non-destructive testing (NDT) data directly to the local coordinate system of the 3D model that corresponds to the inspected component. Transferring this approach to other manual inspection technologies allows for a superimposition of recorded NDT data without any postprocessing or transformation. A better damage assessment is thus enabled. The inspection system, the inspection tool calibration and the camera registration process are described and analysed in detail. This work is focused on the analysis of the system accuracy, which is realised by using a reference body.

In industrial manufacturing processes, targets for infrared marker based tracking have to be robust and must integrate into the tools without disturbing the work flow. In this paper, we propose cylindrical markers attached directly to the tool. We show that, targets equipped with cylindrical markers can be tracked with about the same precision as targets equipped with spherical markers, if a correction for the reduced symmetry is applied. Additionally, the markers can be placed on different parts of a tool with a flexible connection, which is also considered in this paper.

Marker-based optical outside-in tracking is a mature and robust technology used by many AR, VR and motion capture applications. However, in small environments the tracking cameras are often difficult to install. An example scenario are ergonomic studies in car manufacturing, where the motion of a worker needs to be tracked in small spaces such as the trunk of a car. In this paper, we describe how to extend the tracking volume in small, cluttered environments using small and flexible wireless cameras in form of unmodified mobile phones that can quickly be installed. Since those small cameras are not synchronized with the main tracking cameras, we describe several modifications to the tracking algorithms, such as inter-frame interpolation, the replacement of the least-squares adjustment by a Kalman filter and the integration of rolling-shutter compensation. To support the quick setup of additional cameras while the tracking system is running, the system is extended by an on-line calibration technique that determines the extrinsic camera parameters without requiring a dedicated calibration step.

In most close-range photogrammetry applications, the cameras are modelled as imaging systems with perspective projection combined with the lens distortion correction as proposed by Brown in 1971. In the 1980s, the calibration of video cameras received considerable attention. This required compensation for further systematic effects caused by the digitization of the analogue image signal. Modelling the image process in that manner has become the widely-applied standard since then. To take advantage of the increased field of view of individual cameras, the use of wide angle as well as fisheye lenses became common in computer vision and close-range photogrammetry, again requiring appropriate modelling of the imaging process to ensure high accuracies.A.R.T. provides real-time tracking systems with infra-red cameras, which are in some cases equipped with short focal length lenses for the purpose of increased fields of view, resulting in larger trackable object volumes. Unfortunately the lens distortion of these cameras reaches magnitudes which can not be sufficiently modelled with the customary Brown model as - mainly at high excentricities such as image corners - the calculation of the correction is not applicable. Considerations to avoid modelling these lenses as fisheye projections led to an alternate and rather pragmatic approach, where the distortion model is extended by a fourth radial distortion coefficient. Due to numeric instabilities, a stepwise camera calibration is required to achieve convergence in the bundle adjustment process.This paper presents the modified lens distortion model, describes the stepwise calibration procedure and compares results in respect to the conventional approach. The results are also compared to the approach wherein the camera lens is modelled as a fisheye projection. The introduction of a fourth radial lens distortion parameter allows the correction of lens distortion effects over the full sensor area of wide angle lenses, which increases the usable field of view of that specific camera and therefore the size of the trackable observed object volume. The approaches with the extended lens distortion model and the fisheye projection were successfully implemented and tested, and are on target to become part of the A.R.T. product range.

In many augmented reality applications, in particular in the medical and industrial domains, knowledge about tracking errors is important. Most current approaches characterize tracking errors by 6×6 covariance matrices that describe the uncertainty of a 6DOF pose, where the center of rotational error lies in the origin of a target coordinate system. This origin is assumed to coincide with the geometric centroid of a tracking target. In this paper, we show that, in case of a multi-camera fiducial tracking system, the geometric centroid of a body does not necessarily coincide with the point of minimum error. The latter is not fixed to a particular location, but moves, depending on the individual observations. We describe how to compute this point of minimum error given a covariance matrix and verify the validity of the approach using Monte Carlo simulations on a number of scenarios. Looking at the movement of the point of minimum error, we find that it can be located surprisingly far away from its expected position. This is further validated by an experiment using a real camera system.

One of the instruments on board of the ALOS satellite, launched by the Japanese Aerospace Exploration Agency (JAXA) in 2006, is the Panchromatic Remote-sensing Instrument for Stereo Mapping (PRISM). PRISM has three cameras with different viewing directions (nadir, forward, backward), recording imagery with a ground resolution of 2.5 m. A main characteristic of raw ALOS PRISM imagery is that depending on the observation mode each scene is split into four or six separate strips, each related to an individual CCD chip. Basic imagery is delivered as one image data file per strip. We have developed a pushbroom sensor model that is capable of dealing with individual CCD chips sharing some orientation parameters and that can thus be applied to ALOS PRISM imagery. It has previously been shown that pixel-level results can be achieved for georeferencing of ALOS PRISM imagery using this sensor model and a moderate number of ground control points. However, the distribution of resulting residuals suggested that the parameters describing the relative alignment of the individual CCD chips provided by JAXA might not be perfect. Thus, the sensor model was expanded to be capable of self-calibration of these CCD alignment parameters. In this paper, the sensor model will be outlined and the new self-calibration technique described. The effectiveness of self-calibration will be assessed as well as the calibration process carried out by JAXA, in the latter case comparing a set of CCD alignment parameters calibrated in October 2006, and thus representing a very early stage of system calibration, to an updated parameter set obtained in July 2007. Three scenes (forward, backward, nadir) covering a test field in Melbourne (Australia), consisting of more then 100 points surveyed by kinematic GPS, were used for this assessment. Our results show that self-calibration changes the relative alignment of the CCD chips by up to two pixels. If the original calibration data are used, self-calibration can improve the accuracy of the results by 33% and from pixellevel to sub-pixel level. The updated calibration parameters provided by JAXA yield considerably better results than the original ones. In this case, self-calibration essentially helps to increase the height accuracy by about 20%.

A new generic pushbroom sensor model for high-resolution satellite images in which the orbit and attitudes are modelled by splines is presented. Direct observations for the satellite orbits and attitudes provided in imagery metadata files are used to determine the parameters of these splines. As such observations are contaminated by systematic errors, the new sensor model also incorporates error correction for the orbit and attitude angles. Camera model parameter definitions and file formats generally differ between satellite pushbroom scanners, and the system image- to object-space transformation models are not always compatible with a sensor model based on the familiar perspective model applied in photogrammetry. In order to make the new sensor model applicable to a large number of satellite imaging systems, these vendor-specific definitions are first mapped to the definitions of the new sensor model during data import. This is illustrated for QuickBird, SPOT 5 and ALOS PRISM imagery. The model has been extensively tested using imagery from these same three satellites, over test sites in Melbourne and Bhutan. The tests have shown that the new sensor model can produce georeferencing accuracy of 1 pixel and better when biases in orbit and attitude data are compensated.

The CARTOSAT 1 satellite, launched by the Indian Space Research Organisation (ISRO) in 2005, can provide panchromatic alongtrack stereo imagery with a ground resolution of 2.5 m. Along with the imagery, encrypted files with rational polynomial coefficients (RPCs) and meta-data are distributed by ISRO. The RPCs allow direct georeferencing within certain limits depending on the on-board systems for registering the orbit path and attitudes of the satellite. At the Cooperative Research Centre for Spatial Information at the University of Melbourne (Australia), the software package Barista for the processing of high-resolution satellite images is being developed. Barista offers three techniques for precise georeferencing of such image data, namely the 3D affine model, bias correction for RPCs, and a generic pushbroom sensor model. The 3D affine model can only be applied when ground control points (GCPs) are available. The RPC model can be improved beyond the limits of direct georeferencing by correcting for the biases contained in the original RPCs. This process requires at least one well-defined GCP per image. Whereas the meta-data for CARTOSAT 1 imagery do not contain all the information required for using the generic pushbroom sensor model for direct georeferencing, they provide initial values for such a sensor model to be determined if enough GCPs are available. In this paper, the authors compare the geopositioning accuracy achievable with CARTOSAT 1 imagery via the 3D affine, bias-corrected RPC and generic pushbroom sensor models. A stereo pair of images covering Hobart, Australia, was processed using Barista. In addition to the imagery, an object point array of altogether 69 3D GPS-surveyed points was utilised. They were distributed all over Hobart and covered about one quarter of the scene. In order to assess the georeferencing accuracy that can be achieved using CARTOSAT 1 images, bundle adjustment was carried out using all three sensor models and nine well-distributed GCPs. The absolute accuracy was then assessed via the remaining 60 points, which served as independent checkpoints. The georeferencing results obtained for CARTOSAT 1 in the Hobart test field are very encouraging. Whereas direct georeferencing using the RPCs provided by ISRO yielded sub-optimal results, the provision of a small number of GCPs is enough to boost the positioning accuracy to subpixel level in planimetry and to make it slightly better than 1 pixel in height, independent from the sensor model used.

* Modified version of the paper Application of a Generic Sensor Orientation Model to SPOT 5, Quickbird and ALOS Imagery presented at the 28 Asian Conference on Remote Sensing, Kuala Lumpur, November 2007. Abstract This paper will present the results of experimental applications of a new, generic sensor orientation model for pushbroom imaging sensors to ALOS PRISM imagery. Within the model, which has been incorporated into the BARISTA software system for metric information extraction from satellite imagery, the sensor orbit and attitudes are modelled by splines. In order to determine the parameters of the splines, direct vendor-provided observations for the satellite orbits and attitudes, available in the metadata files, are employed. These direct observations are usually contaminated by systematic errors and so a rigorous model is used to compensate perturbations in the orbit and attitude data. The new sensor model has been designed to be applicable to a large variety of sensors, with satellite-specific definitions being mapped to the definitions of the sensor model during data import. The experimental evaluations of ALOS PRISM imagery discussed, which cover test field applications in Australia and Bhutan, demonstrate that the model is capable of sub-pixel level geopositioning accuracy.

A new generic pushbroom sensor model for high-resolution satellite images is presented. The sensor orbit and attitudes are modelled by splines. In order to determine the parameters of the orbit and attitude splines, direct observations for the satellite orbits and attitudes that are provided by the data vendors in metadata files are considered. As these direct observations are usually contaminated by systematic errors, the pushbroom sensor model also requires a correction model for these systematic errors. Unfortunately, the definitions of file formats and model parameters provided by the vendors are usually different and sometimes not compatible with a sensor model based on a perspective transformation. Our new sensor model is designed to be applicable to a large variety of sensors. Vendor-specific definitions are mapped to the definitions of our sensor model during data import. A rigorous model is employed for compensating systematic errors in the orbit and attitude data. In this paper, we present the sensor model and describe the way the vendor-specific definitions are mapped to the definitions of the new sensor model for Quickbird, SPOT 5 and ALOS/PRISM. First results for the correction of systematic errors will be given for Quickbird and SPOT 5 for test sites in Melbourne and Bhutan. * Corresponding author.

The extraction of metric information from image data is highly relevant for the update of geospatial data bases. Photogrammetric methods offer tools to extract such information in a reliable and accurate manner from mostly multiple image setups using appropriate sensor models. These methods require a certain amount of expertise and often expensive software systems to process the image data. An alternative to extract metric information from a multi-image configuration is feature extraction from single, already georeferenced images. The accuracy potential from georeferencing may be sufficient for various applications. This paper describes the method of georeferencing from orthorectified and non-orthorectified high-resolution satellite imagery. During the orthorectification process, georeferencing information enabling the transformation from pixel coordinates to two-dimensional object space coordinates is generated. Mathematically this georeferencing information – often supplied in the commonly used Tiff World File (TFW) format – describes an affine transformation modelling a two-dimensional shift, two scale factors and two rotations. This additional information is often part of an orthorectified image product. Alternatively, the georeferencing information can also be generated with the use of ground control points (GCPs) and can be applied to both orthorectified and non-orthorectified images. For the generation of a set of the six affine transformation parameters, a minimum of three GCPs is necessary. If only two GCPs are available, a two-dimensional conformal transformation using four parameters can be applied for the image to object space transformation. The paper analyses the two methods of coarse georeferencing applied to orthorectified and non-orthorectified highresolution satellite imagery on the basis of three different data sets and different projection systems. The data processing was performed with the software package Barista. The paper quantifies the accuracy yielded with georeferencing, discusses its limitations and concludes with comments about its suitability for potential applications.

The demand for accurate and up-to-date spatial information is increasing and its availability is becoming more important for a variety of tasks. Today’s commercial high-resolution satellite imagery (HRSI) offers the potential to extract useful and accurate spatial information for a wide variety of mapping and GIS applications. The extraction of metric information from images is possible due to suitable sensor orientation models, which describe the relationship between two-dimensional image coordinates and threedimensional object points. With IKONOS and QuickBird imagery, camera replacement models such as rational polynomial coefficients (RPCs) or alternative models such as the affine projection model are used to describe the relationship between image space and object space. With the sensor orientation determined, accurate metric 3D information can be extracted from HRSI through multi-image processing as well as from single images via monoplotting. Monoplotting is a well-known photogrammetric technique for extracting 3D spatial information from single aerial imagery of terrain described by a digital elevation model (DEM). The method also offers potential for single-image analysis of high-resolution satellite imagery (HRSI). This paper describes the implementation and application of monoplotting functions in the photogrammetric software package Barista and investigates the prospects of single IKONOS and QuickBird images for 3D feature point collection and the generation of 3D building models. The experimental determination of the accuracy of monoplotting from IKONOS and QuickBird imagery is also reported.

Monoplotting is a well‐known photogrammetric technique for extracting 3D spatial information from single aerial imagery of terrain described by a digital elevation model. The method also offers potential for single‐image analysis of high‐resolution satellite imagery (HRSI). This paper investigates the prospects of single IKONOS and QuickBird images for 3D feature pointcollection and the generation of 3D building models. The implementation of monoplotting functions in the photogrammetric software package ‘Barista’ is described and an experimental determination of the accuracy of monoplotting from IKONOS and QuickBird imagery is reported.

Various applications such as meteorology, climatology or hydrology require information about the soil hydraulic properties over large areas. Microwave radiometry is a promising approach to gather this type of information. The microwave emission from soils is strongly affected by the roughness of the soil surface. This effect has therefore to be quantified to get a reasonable estimation of the hydraulic properties. In a cooperation of the Institute of Terrestrial Ecology, Soil Physics with the Institute of Geodesy and Photogrammetry digital surface models of soils were generated to study the influence of the surface roughness on the soil measurements. Accurate Digital Surface Models (DSM) can be derived by the application of photogrammetric measurement techniques and provide the spatial basis to extract roughness information. In this paper an approach to determine the roughness of the topsoil surface is presented.

We compare methods used to measure the water content near the soil surface. The primary objective of this project is to link remotely sensed surface water contents to the soil water regime, in particular to the regime of structured soils. We attempt to use the dynamics of spatially averaged surface water contents measured with microwave radiometry to predict preferential infiltration and drainage. Under field conditions the so-called macropore flow plays an important role in the infiltration and drainage behavior of a soil, as well as in the mass transfer of all kinds of solutes to larger soil depths. These rapid processes are only detectable during the first few hours after a rainfall event, when most of the larger pores are still water filled. The main focus of our project lies in an areal integration of such processes on a field scale. For this reason, we depend on areal data with a high temporal resolution that allow to characterize the soil water dynamics. In this study we report on a field experiment with two different ground-based radiometers (1.4 GHz and 11.4 GHz, respectively) centered at a 5 m × 10 m bare soil plot. The brightness temperature measured with passive microwave sensors contains information on surface water content that is already spatially averaged. Furthermore the water content was measured in-situ with time domain reflectometry probes (TDR) assembled at five depths. In the same depths we measured matric potential (pressure head of soil water) and soil temperature. These data were recorded every 30 min from May to July 2002. In addition, we determined the moisture profile over the top 15 cm using neutron radiography. Transmission radiographs of soil slabs vertically taken from the surface horizon allow for a high spatial resolution of the water distribution. In order to characterize the surface roughness of the soil on a mm-scale we used optical measurement techniques. We illustrate the implications of the results from this field campaign on the dynamics of surface water content.

3D-Particle Tracking Velocimetry (PTV) is one of the most flexible techniques for flow measurement, which allows the determination of three-dimensional velocity fields. Research activities in this field performed by the Institute of Geodesy and Photogrammetry at ETH Zurich for more than a decade have reached a status of an operational and reliable measurement method used in hydrodynamics and space applications. The method is based on the visualization of a flow with small, neutrally buoyant particles and recording of particle image sequences with 3-4 CCD cameras. In cooperation with the Institute of Hydromechanics and Water Resources Management at ETH Zurich further progress has been achieved in the improvement of the existing hardand software solutions. Regarding the algorithmic aspect of the method a new spatiotemporal matching method was developed, implemented and tested on different data sets. In former implementations the determination of the 3D particle positions was separated from the tracking of each particle, while the new method uses a combination of image and object space based information to establish spatio-temporal correspondences between particle positions of consecutive time steps. A system based on 4 CCD cameras is capable of tracking up to 1000 particles at video frequency (25Hz) with a relative accuracy of the velocity vectors of approximately 1:4000 of the field of view. The latest developments of the algorithmic aspects of 3D PTV are described and some examples of the successful application of the method are given in this paper. INTRODUCTION The 3D PTV is a technique for the determination of 3D velocity fields in flows. The existing 3D PTV solution developed at the Institute of Geodesy and Photogrammetry applying a object space based tracking algorithm should be improved in a way that the redundant information in image and object space is exploited more efficiently. The use of image and object space based information in combination with a prediction of the particle motion was thought to lead to enhanced results in the velocity field determination. The most important result to be expected from this work is a substantial increase of the tracking rate in 3D PTV. This is of importance mainly in the context of a Lagrangian analysis of particle trajectories, which can be considered the actual domain of the technique. Long trajectories are an absolute requisite for a Lagrangian flow analysis, as integral time and length scales can only be determined if long correlation lengths have been recorded. In addition, the number of simultaneous trajectories should be large enough to form a sufficient basis for a statistical analysis. Due to interruptions of particle trajectories caused by unsolved ambiguities the number of long trajectories decreases exponentially with the trajectory length. Very long trajectories over hundred and more time instances can so far only be determined if the probability of ambiguities is reduced by a low seeding density, thus concurrently reducing the spatial resolution of the system and the basis for a statistical analysis. A reduction of the trajectory interruptions due to unsolved ambiguities can multiply the yield of long trajectories and thus the usefulness of the results of 3D PTV, which further enlarges the application potential of the technique. Within the framework of a research project of the Swiss National Science Foundation a new spatio-temporal matching algorithm was developed and implemented. The technique has reached a status of an operational and reliable measurement tool used in hydrodynamics and space applications (Becker et al, 1995, Maas et al, 1997, Willneff and Maas, 2000). NOMENCLATURE XO,YO, ZO: 3D coordinates of projective center O Xi,Yi, Zi: 3D coordinates of object point Pi xi, yi: Image coordinates of Pi c: Principle distance of the camera xh, yh: Coordinates of the principle point ω, φ, κ: Rotation angles of the direction of the optical axis Drot: Rotation matrix as function of the rotation angles ω, φ, κ aij: Elements of 3x3 rotation matrix ti: Time step of image sequence