The paper discusses the impact of sense and assessment models for videometry. Full automation of calibration and orientation procedures appears to be as necessary for enlarging the field of applications as the use of explicit geometric and semantic scene knowledge. The focus on achieving highest possible accuracy needs to be embedded into a broader context of scene analysis. Examples demonstrate the feasibility of tools from Computer Vision for image metrology.
The paper deals with our research on developing videogrammetry based 3D modeling procedure for indoor facilities. The modeling is based on recorded video sequences, and the analysis procedure is reverse if compared to the traditional one. Here a functional 3D model is built first and the measurements for exact geometry are done thereafter: `Modeling first--then measuring'. The final output will be a photorealistic model.
The key to navigate in a 3D environment or designing autonomous vehicles that can successfully maneuver and manipulate objects in their environment is the ability to create, maintain, and use effectively a 3D digital model that accurately represents its physical counterpart. Virtual exploring of real places and environments, either for leisure, engineering design, training and simulation, or tasks in remote or hazardous environments, is more effective and useful if geometrical relationships and dimensions in the virtual model are accurate. A system which can rapidly, reliably, remotely and accurately perform measurements in the 3D space for the mapping of indoor environments is needed for many applications. In this paper we present a mobile mapping system that is designed to generate a geometrically precise 3D model of an unknown indoor environment. The same general design concept can be used for environments ranging from simple office hallways to long winding underground mine tunnels. Surfaces and features can be accurately mapped from images acquired by a unique configuration of different types of optical imaging sensor and dead reckoning positioning device. This configuration guarantees that all the information required to create the 3D model of the environment is included in the collected data. Sufficient overlaps between 2D intensity images, in combination with information from 3D range images, insure that the complete environment can be accurately reconstructed when all the data is simultaneously processed. The system, the data collection and processing procedure, test results, the modeling and display at our virtual environment facility are described.
The subject of this paper is the research that aims at efficiency improvement of acquisition of 3D building models from digital images for Computer Aided Architectural Design (CAAD). The results do not only apply to CAAD, but to all applications where polyhedral objects are involved. The research is concentrated on the integration of a priori geometric object information in the modeling process. Parallelism and perpendicularity are examples of the a priori information to be used. This information leads to geometric constraints in the mathematical model. This model can be formulated using condition equations with observations only. The advantage is that the adjustment does not include object parameters and the geometric constraints can be incorporated in the model sequentially. As with the use of observation equations statistical testing can be applied to verify the constraints. For the initial values of orientation parameters of the images we use a direct solution based on a priori object information as well. For this method only two sets of (coplanar) parallel lines in object space are required. The paper concentrates on the mathematical model with image lines as the main type of observations. Advantages as well as disadvantages of a mathematical model with only condition equations are discussed. The parametrization of the object model plays a major role in this discussion.
Image metamorphosis is a popular technique for creating a smooth transition between two images. In this paper, we present a new object-oriented method for range image metamorphosis. A metamorphosed range image X(t) is represented by a linear Minkowski combination (l-t)A + tB of two given range images A and B based on the theory of Mathematical Morphology. In consideration of the different features of individual objects in range images, the presented method is extended to three different kinds of models, i.e. volume metamorphosis between convex objects, volume metamorphosis between concave objects, and volume metamorphosis between convex-decomposed concave objects. Finally, some practical examples for range image metamorphosis are given to demonstrate the principle and efficiency of the suggested approach.
The increasing need for describing 3D spatial data in urban areas has led to the development of 3D city models or more widely 3D GIS. Recently, the airborne remote sensing industry has supplied a new way for fast acquisition of 3D spatial data with high densities and increased accuracy at lower cost. At mean time, exiting maps and 2D spatial databases have also supplied a well-organized infrastructure to be used for 3D data extension. In this paper, a integrated method for large scale 3D-city modeling based on multi-viewed laser range images and existing maps is presented, in which we want to integrate different kinds of data from multi-sources, optimizing the different algorithms, and to find a suitable way for operationally generating 3D spatial databases and reducing the related cost and time. Finally, an example is given to demonstrate the operational applicability of the suggested approach.
Real-time measurement using multi-camera 3D measuring system requires three major components to operate at high speed: image data processing; correspondence; and least squares estimation. This paper is based upon a system developed at City University which uses high speed solutions for the first and last elements, and describes recent work to provide a high speed solution to the correspondence problem. Correspondence has traditionally been solved in photogrammetry by using human stereo fusion of two views of an object providing an immediate solution. Computer vision researchers and photogrammetrists have applied image processing techniques and computers to the same configuration and have developed numerous matching algorithms with considerable success. Where research is still required, and the published work is not so plentiful, is in the area of multi-camera correspondence. The most commonly used methods utilize the epipolar geometry to establish the correspondences. While this method is adequate for some simple situations, extensions to more than just a few cameras are required which are reliable and efficient. In this paper the early stages of research into reliable and efficient multi-camera correspondence method for high speed measurement tasks are reported.
The demand for high accuracy, for example in 3D-measurements or in quality control, requires high precision calibration of CCD-cameras. In this paper we present a test-field-based high-precision calibration procedure with the focus on the techniques which enable an autonomous calibration: automatic control of the imaging process with geometric tests, algorithms for automatic point detection and matching, automatic determination of approximate values for the orientation parameters and automatic model selection of the interior orientation parameters. Finally, the parameters are estimated in an iterative least squares adjustment.
Portable digital image cameras such as the Kodak DCS420 and 460 are widely used for metric applications of close-range photogrammetry. The use of these cameras is typically in one of two forms, either as a mobile single camera for offline industrial measurement applications, or in a multi-camera system for online operation in a `work-cell' environment. In common with any metric system, the calibration of the cameras is of paramount importance to maintain the fidelity of the collinearity solution, and therefore the maximum possible accuracy. Any variation in the calibration during image capture for an offline or an online system will inevitably have a deleterious effect on the measurement accuracy. This paper describes a series of experiments concerning the calibration stability of DCS420 and 460 cameras in offline and online configurations. The calibration strategy is described and the effect of a minor modification to the CCD sensor mounting is presented. The variations in RMS image residuals and calibration parameters are quantified in terms of magnitude and significance for the different modes of operation.
Today's developments in industrial robots focus on aims like gain of flexibility, improvement of the interaction between robots and reduction of down-times. A very important method to achieve these goals are off-line programming techniques. In contrast to conventional teach-in-robot programming techniques, where sequences of actions are defined step-by- step via remote control on the real object, off-line programming techniques design complete robot (inter-)action programs in a CAD/CAM environment. This poses high requirements to the geometric accuracy of a robot. While the repeatability of robot poses in the teach-in mode is often better than 0.1 mm, the absolute pose accuracy potential of industrial robots is usually much worse due to tolerances, eccentricities, elasticities, play, wear-out, load, temperature and insufficient knowledge of model parameters for the transformation from poses into robot axis angles. This fact necessitates robot calibration techniques, including the formulation of a robot model describing kinematics and dynamics of the robot, and a measurement technique to provide reference data. Digital photogrammetry as an accurate, economic technique with realtime potential offers itself for this purpose. The paper analyzes the requirements posed to a measurement technique by industrial robot calibration tasks. After an overview on measurement techniques used for robot calibration purposes in the past, a photogrammetric robot calibration system based on off-the- shelf lowcost hardware components will be shown and results of pilot studies will be discussed. Besides aspects of accuracy, reliability and self-calibration in a fully automatic dynamic photogrammetric system, realtime capabilities are discussed. In the pilot studies, standard deviations of 0.05 - 0.25 mm in the three coordinate directions could be achieved over a robot work range of 1.7 X 1.5 X 1.0 m3. The realtime capabilities of the technique allow to go beyond kinematic robot calibration and perform dynamic robot calibration as well as photogrammetric on-line control of a robot in action.
This paper describes the parameters and method used to calibrate an anamorphic lens specially designed for a 3D triangulation based laser range sensor. It expands the more `conventional' spherical lens calibration technique that has been developed in photogrammetry to include the extra distortions introduced by the strong astigmatism and by the different optical principal planes of the lens of the anamorphic design. Experimental results using a prototype of the lens for the Biris range sensor are presented.
The problem of calibrating a vision system is extremely important for practical applications such as 3D reconstruction and pose estimation of 3D objects. In this paper we present a method for optimally estimating the internal and external camera parameters from point or line correspondences. First we extend the linear method of Faugeras & Toscani (1986) to line correspondences, second, we develop for both point and line correspondences a new method based on the minimization of an error function. We show how to minimize this error function in order to compute the unknown parameters (i.e., the internal and external camera parameters). This minimization leads to a non-linear optimization problem. We introduce an elegant way to automatically establish 2D to 3D feature correspondences using projective geometry. We perform a stability analysis both for our method and for the linear method. In the light of this comparison, the non-linear method seems to be the most robust one with respect to noise. Finally, we present some experiments to completely calibrate a binocular stereo rig with different calibration patterns. This stereo rig forms the main vision sensor of the Janus humanoid robotics system we are currently developing.
In recent years, considerable developments have occurred in the field of digital photogrammetry. These have been due mainly to increases in computing power, the refinement of feature- and area-based image matching algorithms and the reduction in cost of equipment capable of producing near real-time images in digital format. Despite the progress, a major limitation to the widespread application of digital photogrammetry is the resolution of the digital images being used. Much time and effort has been expended trying to improve resolution through hardware solutions consisting of imaging sensors with increased numbers of pixels. An alternative solution is offered in this paper. An algorithm which uses a combination of photogrammetric techniques and a rigorous mathematical solution to enhance the resolution of digital images is described.
High resolution digital stillvideo cameras have found wide interest in digital close range photogrammetry in the last five years. They can be considered fully autonomous digital image acquisition systems without the requirement of permanent connection to an external power supply and a host computer for camera control and data storage, thus allowing for convenient data acquisition in many applications of digital photogrammetry. The accuracy potential of stillvideo cameras has been extensively discussed. While large format CCD sensors themselves can be considered very accurate measurement devices, lenses, camera bodies and sensor mounts of stillvideo cameras are not compression techniques in image storage, which may also affect the accuracy potential. This presentation shows recent experiences from accuracy tests with a number of large format stillvideo cameras, including a modified Kodak DCS200, a Kodak DCS460, a Nikon E2 and a Polaroid PDC-2000. The tests of the cameras include absolute and relative measurements and were performed using strong photogrammetric networks and good external reference. The results of the tests indicate that very high accuracies can be achieved with large blocks of stillvideo imagery especially in deformation measurements. In absolute measurements, however, the accuracy potential of the large format CCD sensors is partly ruined by a lack of stability of the cameras.
Digital cameras providing a 4k X 4k high resolution CCD area array sensor have been developed.; The acquired digital image data can be used for high precision photogrammetric object reconstruction, e.g. in industrial applications such as quality control and assurance. Technical specifications of some imaging systems are given in the paper. The metric quality of cameras based on the 4k X 4k Dicomed BigShot digital camera is determined from test measurements.
In this paper geometric and radiometric investigations performed with two Leica/Helava DSW200 scanners will be presented. Good quality test patterns and accurate processing methods for their performance evaluation have been employed. The geometric tests include global and local geometric errors, misregistration between color channels, geometric repeatability, and determination of the geometric resolution. Efforts were made to separate the contribution of various error sources (especially mechanical positioning, electronic errors and lens distortion) on the total error. The radiometric tests include investigation on noise, linearity, dynamic range, spatial, temporal and spectral variation of noise, and artifacts. After a brief description of the scanner, details on the above investigation, used analysis methods and achieved results will be presented. The most important errors, especially for organizations employing the scanner in production, are nonrepeatable geometric errors that are due to mechanical positioning inaccuracies and vibrations. This causes whole CCD image tiles (4 million pixels) to be systematically shifted with respect to their neighboring tiles, with obvious negative influence on any subsequent measuring task.
Depth-from-Defocus using the Spatial-Domain Convolution/Deconvolution Transform Method (STM) is a useful technique for 3D vision. STM involves simple local operations in the spatial domain on only two images recorded with different camera parameters (e.g. by changing lens position or changing aperture diameter). In this paper we provide a theoretical treatment of the noise sensitivity analysis of STM and verify the theoretical results with experiments. This fills an important gap in the current research literature wherein the noise sensitivity analysis of STM is limited to experimental observations. Given the image and noise characteristics, here we derive an expression for the Root Mean Square (RMS) error in lens position for focusing an object. This RMS error is useful in estimating the uncertainty in depth obtained by STM. We present the results of computer simulation experiments for different noise levels. The experiments validate the theoretical results.
For many years, photogrammetry has been in use at TRW. During that time, needs have arisen for highly repetitive measurements. In an effort to satisfy these needs in a timely manner, a specialized Robotic Video Photogrammetry System (RVPS) was developed by TRW in conjunction with outside vendors. The primary application for the RVPS has strict accuracy requirements that demand significantly more images than the previously used film-based system. The time involved in taking these images was prohibitive but by automating the data acquisition process, video techniques became a practical alternative to the more traditional film- based approach. In fact, by applying video techniques, measurement productivity was enhanced significantly. Analysis involved was also brought `on-board' to the RVPS, allowing shop floor acquisition and delivery of results. The RVPS has also been applied in other tasks and was found to make a critical improvement in productivity, allowing many more tests to be run in a shorter time cycle. This paper will discuss the creation of the system and TRW's experiences with the RVPS. Highlighted will be the lessons learned during these efforts and significant attributes of the process not common to the standard application of photogrammetry for industrial measurement. As productivity and ease of use continue to drive the application of photogrammetry in today's manufacturing climate, TRW expects several systems, with technological improvements applied, to be in use in the near future.
This paper presents a concept for the recognition and localization of objects which relies on multi-sensor fusion and active exploration. Today, researchers in photogrammetry generally agree that the use of complementary sensors, e.g. ranging and imaging cameras, is important for simplifying interpretation and related tasks. But no notice has been taken so far of the role of active exploration. Our work is part of a research program where five institutes of Stuttgart University co-operate to develop an experimental measuring system for flexible inspection and gauging. The system will be capable of determining automatically the shape, form and class attributes of an industrial object. It then solves in a self-acting manner the measuring task associated with that object. The paper, which focuses on the object recognition concept, briefly describes the experimental measuring system and the sensors employed. A number of subsequent processing steps of the whole procedure are illustrated through use of initial experimental results.
In this paper a general geometric description of the optical methods for 3D coordinate measurement is presented. Similar to holographic interferometry this new approach is based on the concept of measuring sensitivity. As a special case the derived basic relation is applied to the fringe projection technique using a physical model of this measurement method. Moreover a geometric 3D model that contributes to a dramatic reduction of systematic distortions of measured 3D coordinates is presented. On the one hand this model is sufficiently general but on the other hand still easy to handle. It permits an explicit and direct determination of 3D coordinates from primary measuring data as well as a calibration of the measuring set-up using linear identification methods mainly. The described 3D model can be applied also with advantage to multiview registration tasks.
This paper presents a method for motion detection from image sequences with a moving camera. The most important problem is how to cancel the motion of backgrounds caused by the motion of camera. Existing methods can be divided into two groups. First group is using information about the motion of camera. Second group is using only image sequences. First group can be used only in restricted environment. Second group is free from environmental condition. But calculated accuracy of this method is not so high, and complex preprocessing is necessary for the detection of moving objects. In this paper, we proposed a method which can detect a moving object by using precise geometric transformation for background. Experimental results showed the effectiveness of this method.
There are three essential requirements for real-time 3D measurement using targeted points: fast 2D image processing; a fast solution to the correspondence problem; and fast computation of 3D coordinates. This paper brings together research work to produce such solutions and considers other work which has appeared during the project duration.
Videometric measurements in wind tunnels can be very challenging due to the limited optical access, model dynamics, optical path variability during testing, large range of temperature and pressure, hostile environment, and the requirements for high productivity and large amounts of data on a daily basis. Other complications for wind tunnel testing include the model support mechanism and stringent surface finish requirements for the models in order to maintain aerodynamic fidelity. For these reasons nontraditional photogrammetric techniques and procedures sometimes must be employed. In this paper several such applications are discussed for wind tunnels which include test conditions with Mach numbers from low speed to hypersonic, pressures from less than an atmosphere to nearly seven atmospheres, and temperatures from cryogenic to above room temperature. Several of the wind tunnel facilities are continuous flow while one is a short duration blow-down facility. Videometric techniques and calibration procedures developed to measure angle of attack, the change in wing twist and bending induced by aerodynamic load, and the effects of varying model injection rates are described. Some advantages and disadvantages of these techniques are given and comparisons are made with non-optical and more traditional video photogrammetric techniques.
This study examines a personal computer system for automatic measurement of the traffic flow, vehicle traveling velocity, and traveling tracks, as well as identification of types of vehicle, on the basis of the image of road traffic that is videotaped by a video camera mounted on an airship type balloon. Since observation is made almost right above the observation point in this traffic flow monitor system, the influence of occlusion due to the vehicle shape is small, and simplification of measurement processing is possible. In that case, the shaking of the airship type balloon causes shaking of the video image. In order to eliminate such an influence, a method to automatically pursue ground control points in order to find the movement of the background so as to cancel the apparent movement was adopted. In the meantime, a vehicle detection/extraction function is necessary for automation of the traffic flow monitor system. As a method to detect and extract a vehicle by means of comparison between the movement of the background and the inclination of the time-space images was adopted in this study. For verification of the measuring accuracy of a series of sequential image analysis methods, experiments were conducted with the help of test vehicles, and data obtained by the sequential image analysis method were compared with actually measured values. As a result, it was found that automatic measurement of the traffic flow, vehicle traveling velocity, and traveling tracks, as well as identification of types of vehicles, would be possible with practical accuracy.
This paper addresses an application installed in quality control where a 100% inspection of geometry (3D) and surface of cuboid (parallelepiped) and ring-shaped magnets is done using a system of 2 CCD matrix cameras, one of which is equipped with on-board processing components and a transmitted-light sensor with microcontroller based data processing for the measurement of the height of the objects. The geometry and surface properties are measured with a diffuse indirect IR-LED flash and a telecentric lens to avoid perspective distortions due to different heights of the measured objects. The surface inspection looks for broken pieces, surface faults and cracks. The second CCD camera uses the same illumination and algorithm to inspect the surface of the other side of the objects after it has been turned around in a return conveyor belt. The main aspect of this paper is the theory of the geometry measurement including sub-pixel accuracy and the surface inspection using a telecentric lens and diffuse illumination. The geometry is measured with an accuracy of 1/3 pixels, independent of the background properties. Using a PentiumTM 100 processor the system is able to inspect up to 5 pieces per second.
Japanese island is covered mountains nearly 80%, then railways and highways are run through the mountainous area. Recently those mountains are running to wild, therefore railway or highway companies are being stand in need of maintenance strictly. Relative heights for railway slope are measured and video maps are produced using High Definition TeleVision system. The HDTV system is a system designed to collect highly definitioned and stabilized images with GPS data turned out by HDTV camera. The 237.2 km along the railway are photographed. And relative height of 530 points are measured. Then measurement accuracy was 0.9 m average. The HDTV system is good for such as long target, sporadic points measurement and rough accuracy request.
Sensor choice is critical in all object inspection imaging systems and the suitability of different types of sensor must be thoroughly assessed before a decision is made. Despite its wide use, a television type camera may not represent the best choice for certain object inspection applications. Specifically, when inspection of an object that has a degree of cylindrical symmetry is required, a line-scan camera is a viable alternative to the television type and, in general, any matrix camera, offering a number of unique advantages. By applying rotational motion to the object of interest and using a line-scan device, an 360 degree(s) view of the object is obtained. The cylindrical surface of the object is effectively unfolded into a planar 2D one, allowing for the efficient inspection of the entire surface of the object from a single, continuous image. To allow accurate object space co-ordinate measurement, a line-scan camera calibration technique has been developed, catering for both interior and exterior parameter calibration. The former accounts for the lens effective focal length, the pierce-pixel value and the timing of the line-scan camera, while the latter yields the relative position and orientation of the camera with respect to a reference object space co-ordinate system.
In the compilation of archival records for archeological artifacts, true ortho-graphic drawings of these artifacts have to be drawn by the archaeologists themselves or part- timers, taking a great deal of time, labor and skill. For saving the labor, the authors have developed ortho projection system using CCD camera. 3D measurement system using ortho projection system are described in this paper. Finally, it demonstrates wireframe model for jomon-pottery by using this system.
Actually, the rapid acquisition and measurement of 2- and 3- D contours and shapes represent an important assignment in modern production processes and quality control. Especially, the nontactile methods like optical triangulation and the digital image processing get more and more importance. With the methods of color image processing a new field of nontactile 3D object ranging arises. Using color coded illuminations it is possible to generate a 3D object description by processing only a single RGB-image. Two of these methods are: the color coded phase-shift method and the color triangulation. The color triangulation combines the methods of the color image processing with the distance measurement by triangulation. The color coded phase-shift method is based on the gray scale phase-shift method. Since both methods are able to generate a 3D object description by processing only a single RGB-image, it is possible to combine these two methods to realize a fast and precise active ranging system without ambiguities.
It is growing to measure the Plant World Cultural Herritages, in order to make the replicas and to check for the size and the shape. The photogrammetry for World Cultural Herritages is very difficult, because they could not be moved and touched. And also, there are in the deep mountains and forest, we must carry up heavy measurement equipments by foot. We developed the close range photogrammetry for the Plant World Cultural Herritage with CAD. It is very effective for the measurement and the planning through the CAD models. These CAD models show many 3D information about botanist and engineers to reconstruct the plant models.
It has been reported the origin of rice crop is going up 10,000 years ago, and also the origin of bronze product is going up 5,500 years ago in China. We researched two huge ancient archaeological sites by photo-interpretations using high-resolution satellite images and low-altitude aerial color pictures and infrared pictures taken by radio- controlled model helicopter. Low-altitude aerial photogrammetry is available to research about soil and plant marks in archaeological fields. And we also have used electromagnetic-wave research for underground archaeological remains.
In this paper a novel method of measuring coaxiality and perpendicularity using laser alignment, diffraction imaging and CCD image detecting is proposed. In this method, we use a laser diode fiber alignment beam as the datum reference of straightness, deviation of the center of the hole from this datum, coaxiality, can be detected by a CCD camera. By rotating a penta-prism that makes this alignment beam bend a right angle a scanning datum plane is generated which can be used to detect the perpendicularity of the axis to an end plane. In the paper the factors that affect the accuracy of the system are analyzed. The experiment result shows that the measurement accuracy of the system is 3 + 2.3 X 10+6.
The effectiveness of the real-time ortho projection and drawing system using a CCD camera and line lasers developed by the authors has been indicated. However, 3D visualization for archeological artifacts, such as wireframe model, surface model or texture model should be added to this system. There are some methods for performing a digital 3D modeling such as the range finder method. The most remarkable point of this system is its ability to get a real-time ortho projection image and a 3D model with one CCD camera.
A research on using linear array CCD detecting technique in measuring form and position errors of large scale dimension is described in this paper. The basic idea of this kind of measurement is that using a stable aligned laser beam incident on a 1D asymmetrical phase plate to form a black line as the reference of the beam, form and position datum lines and planes can be established through a scanner. A linear array CCD is used to measure deviation between the datum and the quantity be measured directly and absolutely. In the paper, we also discuss the reading error of the CCD probe caused by various factors in detail and introduce the CCD probe and its signal processing circuit. Some of experiment results and its application are shown.