This PDF file contains the front matter associated with SPIE Proceedings Volume 6491, including the Title Page, Copyright information, Table of Contents, Introduction (if any), and the Conference Committee listing.

We propose a 3D surveillance system using multiple cameras surrounding the scene. Our application is concerned with identifying humans in the scene and then identifying their postures. Such information can help with automatic threat assessment of a scene. The cameras are fully calibrated and assumed to remain fixed in their positions. Object detection and interpretation are performed completely in 3D space. Using depth information, persons can easily be separated from the background and their posture identified by matching with 3D model templates.

Object tracking from multiple Pan Tilt Zoom (PTZ) cameras is an important task. This paper deals with the evaluation of the result of such a system. This performance evaluation is conducted by first considering the characterization of the PTZ parameters and then by the trajectories themselves. The camera parameters with be evaluated with the homography errors; the trajectories will be evaluated according to the location and miss-identification errors.

In any surveillance or remote monitoring system, the location and configuration/orientation of directed optical sensors is critical for efficient resource utilization as well as correct system behavior. In this paper we introduce a method to determine the set of configurations (pan-tilt-zoom) for a system of cameras in a given a 3D environment with well-characterized human activity. The mainstay of our method is a metric that measures the ground plane coverage by a sensor. Our goal is to seek configurations that provide maximal coverage of areas in a scene that have been deemed important or have intense human activity. Our method was deployed on a portion large urban campus covering a region with multiple buildings, sidewalks, trees and other occluding objects. We built a 3D CAD model of the environment and tested our method on a system of six pan-tilt-zoom cameras. We used activity information gathered from surveillance of the environment to determine the coverage requirements. The resulting sensor configurations satisfied the real-world security and surveillance concerns of the environment.

During the autumn of 2004, a team of 3D imaging scientists from the National Research Council of Canada (NRC) was invited to Paris to undertake the 3D scanning of Leonardo's most famous painting. The objective of this project was to scan the Mona Lisa, obverse and reverse, in order to provide high-resolution 3D image data of the complete painting to help in the study of the structure and technique used by Leonardo. Unlike any other painting scanned to date, the Mona Lisa presented a unique research and development challenge for 3D imaging. This paper describes this challenge and presents results of the modeling and analysis of the 3D and color data.

The calibration of a range camera greatly influences the whole 3D acquisition and modeling process, allowing to minimize the equipment inaccuracy. However, depending on the range camera "openness", we might have systems precalibrated only once by the industrial manufacturer or systems requiring a regular (and mandatory) end-user calibration before any scan session. Independently of the calibration approach, the metrological system characterization represents a point of paramount importance for making the user aware of the actual performances of his equipment. This permits the choice of appropriate resolution in 3D scan planning and allows to properly interpret the feedback indices during the alignment of several range maps trough Iterative Closest Point (ICP). Finally, in polygonal model editing, the modification of geometrical features is greatly helped by the awareness about the 3D capturing device performances. These remarks are effective for both triangulation based instruments, like Minolta Vivid 910, ShapeGrabber SG100 and SG1000 evaluated in this paper, or TOF based instruments. The proposed experimental method is based on post processing of the range data produced by acquiring the surface of a precise test object with a 3D laser scanner. In this procedure resolution, accuracy, and precision parameters are obtained sequentially, through the application of a set of simple geometric processing steps. Such operating easiness make this approach a possible candidate as a mandatory step in any 3D acquisition and modeling project.

The paper deals with the metrological characterization of 3D scanners, and proposes a procedure for their experimental verification in accord with the suggestions of the ISO GUM. The procedure is based on the application of a statistical method for the evaluation of the standard uncertainty to the results of a comparison with a Coordinate Measuring Machine (CMM). Finally the results of the experimental verification of a fringe pattern system are reported and discussed in detail.

When TOF laser scanners were introduced on the market, their performances were rather poor, having in general a measurement uncertainty in the range of centimeters. For this reason it was clear that their application was definitely limited to environment and architecture survey, where the large size of the involved objects makes acceptable the relative measurement error. But with the progressive improvement of technology, and the consequent increase in the measurement precision, the potential range of purposes have been widened. In this paper an application to museum objects have been considered. Studying the scanner performance when working at a low range, and using such results to properly interpret the acquired data, it was possible to survey a famous wooden model of S. Peter basilica in Rome, remodeling its shape with a 3D surface modeler. Resolution, precision and accuracy have been studied at distances ranging from 1 to 3 meters, in working conditions similar to those imposed by the museum constrains. The results were used to properly set-up some post processing steps instrumental characterization as a key step in the 3D modeling process, both for increasing the geometrical data reliability, and for processing them (e.g. in the smoothing phase) in a way compliant with their metrological characteristics.

This paper proposes a real-time phase-stamp range finder (PSRF) using a three-phase correlation image sensor (3PCIS). The system onsists of the same components as in the conventional light-stripe ange finder except the 3PCIS and a three-phase reference signal enerator. During a single scan of the sheet beam, the 3PCIS "stamps" the phase of the reference signals at the moment the light stripe is imaged on each pixel. The three-dimensional shape of the object is obtained in real time by converting the detected phase to the angle of the projected sheet beam. Compared to time-stamp VLSI range finders, the proposed PSRF is not affected by spatial nonuniformity in surface reflectance of the object and background illumination. It can be realized with a simpler architecture than the modulation-type PSRF previously proposed by the authors. Experimental results on a developed PSRF system are demonstrated.

This paper summarizes the causes of uncertainty in 3D data measurement, some basic theory of 3D imaging that explains the origin of some of these causes, and, describes the properties and performances of a 100 square meter facility to perform research in traceable 3D imaging metrology. Built in 2006, the laboratory space allows accurate measurements of 3D data from devices operating at standoff distances from a few centimeters up to 10 meters. A laminar flow of 20°C air at 50% humidity level is maintained within ±0.1°C. The total volume of air in the lab is changed twice a minute (18000 cfm). This characteristic combined with the air filtering design allows air cleanliness to be exceptionally good (Class 100).

Range Imaging (RIM) is a new suitable choice for measurement and modeling in many different applications. RIM is a fusion of two different technologies. According to the terminology, it integrates distance measurement as well as imaging aspects. The distance measurement principle is dominated by the time-of-flight principle while the imaging array (e.g. CMOS sensor) enables each pixel to store also the distance towards the corresponding object point. Due to the technology's relatively new appearance on the market, with a few different realizations, the knowledge of its capabilities is very low. In this paper we present our investigations on the range imaging camera SwissRanger^TM (realized by the Swiss Center for Electronics and Microtechnology, CSEM). Different calibration procedures are performed, including a photogrammetric camera calibration and a distance system calibration with respect to the reflectivity and the distance itself. Furthermore we report about measurement applications in the field of surveillance and biometrics. In particular, range imaging data of moving people are analyzed, to identify humans, detect their movements and recover 3D trajectories.

Solid-state full-field range imaging technology, capable of determining the distance to objects in a scene simultaneously for every pixel in an image, has recently achieved sub-millimeter distance measurement precision. With this level of precision, it is becoming practical to use this technology for high precision three-dimensional metrology applications. Compared to photogrammetry, range imaging has the advantages of requiring only one viewing angle, a relatively short measurement time, and simplistic fast data processing. In this paper we fist review the range imaging technology, then describe an experiment comparing both photogrammetric and range imaging measurements of a calibration block with attached retro-reflective targets. The results show that the range imaging approach exhibits errors of approximately 0.5 mm in-plane and almost 5 mm out-of-plane; however, these errors appear to be mostly systematic. We then proceed to examine the physical nature and characteristics of the image ranging technology and discuss the possible causes of these systematic errors. Also discussed is the potential for further system characterization and calibration to compensate for the range determination and other errors, which could possibly lead to three-dimensional measurement precision approaching that of photogrammetry.

The modelling of complex objects and sites involve the acquisition of a large number of texture maps and range images; each one of them representing a particular viewing angle. These views must be combined and registered in order to create an accurate model of the original. The complexity of the resultant models, and consequently the number of views required, has increased tremendously over the past decade. Furthermore major projects involve multinational and multilingual teams, each with different underlying methodologies. In such conditions, it is difficult to make sense of the annotation and to determine which views should be registered together. We propose a new approach in which similar views are found by content-based indexing and retrieval of textures (2D) and range images (3D). The views are described automatically, according to their appearance and their 3D geometry. A search engine allows retrieving similar and related views for registration.

A system reconstructing arbitrary shapes from images in real time and with enough accuracy would be paramount for a huge number of applications. The difficulty lies in the trade off between accuracy and computation time. Furthermore, given the image resolution and our real time needs, only a small number of cameras can be connected to a standard computer. The system needs a cluster and a strategy to share information. We introduce a framework for real time voxel based reconstruction from images on a cluster. From our point of view, the volumetric framework has five major advantages: an equivalent tree representation, an adaptable voxel description, an embedded multi-resolution capability, an easy fusion of shared information and an easy exploitation of inter-frame redundancy; and three minor disadvantage, its lack of precision with respect to method working at point level, its lack of global constraints on the reconstruction and the need of strongly calibrated cameras. Our goal is to illustrate the advantages and disadvantages of the framework in a practical example: the computation of the distributed volumetric inferred visual hull. The advantages and disadvantages are first discussed in general terms and then illustrated in the case of our concrete example.

We propose a support system of range data acquisition by a laser rangefinder for wide area outdoor modeling in order to reduce un-observed portions of generated model. The system presents the operator a recommendation degree map which illustrates recommendation position of acquisition of range data in the objective area. The operator decides a next acquisition position in consideration of movement distance of sensor system and recommendation degrees of the map. The recommendation degree is computed by measurement density as the index. The recommendation degree at a position is given by the difference between measurement density acquired by rangefinder and measurement density estimated by the system in reachable area of the laser beams. The reachable area of the laser beams is estimated by using a 3D model generated from the acquired range data. The system computes the measurement density by the reachable area of the laser beams. The recommendation degrees in the objective area are computed by the model generated from range data whenever a range data is acquired. Moreover, the system judges whether overlapping portions of the range data can be acquired for the registration by ICP algorithm from a work area which the sensor system can enter.

Craniofacial anthropometry (the measurement and analysis of head and face dimensions) has been used to assess and describe abnormal craniofacial variation (dysmorphology) and the facial phenotype in many medical syndromes. Traditionally, anthropometry measurements have been collected by the direct application of calipers and tape measures to the subject's head and face, and can suffer from inaccuracies due to restless subjects, erroneous landmark identification, clinician variability, and other forms of human error. Three-dimensional imaging technologies promise a more effective alternative that separates the acquisition and measurement phases to reduce these variabilities while also enabling novel measurements and longitudinal analysis of subjects. Indiana University (IU) is part of an international consortium of researchers studying fetal alcohol spectrum disorders (FASD). Fetal alcohol exposure results in predictable craniofacial dysmorphologies, and anthropometry has been proven to be an effective diagnosis tool for the condition. IU is leading a project to study the use of 3D surface scanning to acquire anthropometry data in order to more accurately diagnose FASD, especially in its milder forms. This paper describes our experiences in selecting, verifying, supporting, and coordinating a set of 3D scanning systems for use in collecting facial scans and anthropometric data from around the world.

In vivo quantitative studies of cardiac function in mouse models provide information about cardiac pathophysiology in more detail than can be obtained in humans. Quantitative measurements of left ventricular (LV) volume at multiple contractile phases are particularly important. However, the mouse heart's small size and rapid motion present challenges for precise measurement in live animals. Researchers at Duke University's Center for In Vivo Microscopy (CIVM) have developed noninvasive time-gated microcomputed tomography (micro-CT) techniques providing the temporal and spatial resolutions required for in vivo characterization of cardiac structure and function. This paper describes analysis of the resulting reconstructions to produce volume measurements and corresponding models of heart motion. We believe these are the most precise noninvasive estimates of in vivo LV volume currently available. Our technique uses binary mixture models to directly recover volume estimates from reconstructed datasets. Unlike methods using segmentation followed by voxel counting, this approach provides statistical error estimates and maintains good precision at high noise levels. This is essential for long term multiple session experiments that must simultaneously minimize contrast agent and x-ray doses. The analysis tools are built into the Pittsburgh Supercomputing Center's Volume Browser (PSC-VB) that provides networked multi-site data sharing and collaboration including analysis and visualization functions.

Highspeed cameras form a central component of photogrammetric 3D measurement systems in applications requiring a high temporal resolution in fields such as fast object tracking or high frequency deformation measurement. Current highspeed cameras offer data rates in the order of 1 Gigabyte per second or beyond, delivering for instance 1000 frames per second at 1000×1000 pixels. Most applications using highspeed cameras for visualization or measurement purposes are still 2D. Highspeed camera stereo systems are rare because of the costs of cameras and the lack of synchronizibility of some camera types. The paper gives a short overview on single camera 3D measurement system concepts. It presents two concepts of setting up photogrammetric 3D measurement systems based on a single highspeed camera in detail, one based on a single camera with a stereo mirror system and one on based on a camera-projector combination: A stereo mirror system in front of a camera generates multiple views on one sensor. These views can be considered virtual cameras of a multi-camera photogrammetric measurement system. The paper presents a flexible quadruple-mirror system to be used in 3D motion analysis applications and discusses geometric modeling and system calibration. Photogrammetric surface measurement systems can be configured by a camera and a projector device projecting strip or dot patterns onto a surface. Using the projector as an active element in the system design, dense and accurate 3D surface representations may be generated using coded light approach or phase shift techniques. In the geometric modeling and calibration, the projector can be considered an inverse camera with infinite frame rate. The paper presents results of an accuracy test obtaining a precision potential in the order of ¹/₅₀ pixel from a standard consumergrade beamer. Besides the system design and the geometric modeling of the two concepts, the paper shows results from pilot studies and practical applications.

The use of multi camera systems to determine 3D object coordinates is a common approach in close range photogrammetry. The paper discusses system configurations based on a single camera and mirror systems to generate a virtual multi camera system. The two main advantages of mirror-based multiple-view vision systems can be seen in the lack of synchronisation requirements and the costs of only one camera. The latter applies especially if high speed cameras are required to capture the dynamics of an application. These advantages have to be weighted against the disadvantages of a reduced active image format per view and the restrictions to the imaging geometry. In the paper we present two different systems used for motion analysis applications: A fixed two mirror system is used to analyse pedestrian protection testing in the context of vehicle safety development, and a flexible four mirror system is used to capture 3D velocity fields of particles visualising gas flows in a wind tunnel. A geometric model for multi-mirror stereoscopic systems has been developed, representing the mirror system by multiple virtual cameras in order to warrant compatibility with existing photogrammetric stereo data processing software solutions. The accuracies achieved in practical tests are almost comparable to those obtained with multi camera systems.

We have combined the basic human ability to recognize other individuals with functional anatomical and biomechanical knowledge, in order to analyze the gait of perpetrators as recorded on surveillance video. The perpetrators are then compared with similar analyses of suspects. At present we give a statement to the police as to whether the perpetrator has a characteristic gait pattern compared to normal gait, and if a suspect has a comparable gait pattern. We have found agreements such as: limping, varus instability in the knee at heel strike, larger lateral flexion of the spinal column to one side than the other, inverted ankle during stance, pronounced sagittal head-movements, and marked head-shoulder posture. Based on these characteristic features, we state whether suspect and perpetrator could have the same identity but it is not possible to positively identify the perpetrator. Nevertheless, we have been involved in several cases where the court has found that this type of gait analysis, especially combined with photogrammetry, was a valuable tool. The primary requisites are surveillance cameras recording with sufficient frequency, ideally about 15 Hz, which are positioned in frontal and preferably also in profile view.

This paper presents an overview of 3D body scanning technologies with applications to the fashion and apparel industry. Complete systems for the digitization of the human body exist since more than fifteen years. One of the main users of this technology with application in the textile field was the military industry. In fact, body scanning technology is being successfully employed since many years in military bases for a fast selection of the correct size of uniforms for the entire staff. Complete solutions were especially developed for this field of application. Many different research projects were issued for the exploitation of the same technology in the commercial field. Experiments were performed and start-up projects are to time running in different parts of the world by installing full body scanning systems in various locations such as shopping malls, boutiques or dedicated scanning centers. Everything is actually ready to be exploited and all the required hardware, software and solutions are available: full body scanning systems, software for the automatic and reliable extraction of body measurements, e-kiosk and web solutions for the presentation of garments, high-end and low-end virtual-try-on systems. However, complete solutions in this area have still not yet found the expected commercial success. Today, with the on-going large cost reduction given by the appearance of new competitors, methods for digitization of the human body becomes more interesting for the fashion and apparel industry. Therefore, a large expansion of these technologies is expected in the near future. To date, different methods are used commercially for the measurement of the human body. These can be divided into three major distinguished groups: laser-scanning, projection of light patterns, combination modeling and image processing. The different solutions have strengths and weaknesses that profile their suitability for specific applications. This paper gives an overview of their differences and characteristics and expresses clues for the selection of the adequate method. A special interest is given to practical examples of the commercial exploitation of human body digitization with applications to the fashion and apparel industry.

Membrane distillation (MD) [1] is a relatively new process that is being investigated world-wide as a low cost, energy saving alternative to conventional separation processes such as distillation and reverse osmosis (RO). This process offers some advantages compared to other more popular separation processes, such as working at room conditions (pressure and temperature); low-grade, waste and/or alternative energy sources such as solar and geothermal energy may be used; a very high level of rejection with inorganic solutions; small equipment can be employed, etc. The driving force in MD processes is the vapor pressure difference across the membrane. A temperature difference is imposed across the membrane, which results in a vapor pressure difference. The principal problem in this kind of system is the accurate measurement of the recipient volume change, especially at very low flows. A cathetometer, with up to 0,05 mm resolution, is the instrument used to take these measurements, but the necessary human intervention makes this instrument not suitable for automated systems. In order to overcome this lack, a high resolution system is proposed, that makes automatic measurements of the volume of both recipients, cold and hot, at a rate of up to 10 times per second.

Though the commercial landscape of industrial metrology has changed significantly in recent years, the strengths and advantages of digital photogrammetry, or vision metrology (VM), ensure its role as a highly competitive, three-dimensional coordinate measurement tool. Used predominantly for inspection and reverse engineering, today it is increasingly utilized for building within the manufacturing process. This paper presents an overview of innovation and the current status of VM in industry and engineering. A brief description of the state-of-the-art is followed by summaries of five unique applications that illustrate the basic types of available VM systems. In concluding remarks, selected emerging trends are discussed.

In this contribution we describe the modern 2D and 3D technologies for the documentation, analysis and restoration of paintings. RGB color imaging, IR and UV fluorescence sensors, together with highly precise active sensors are among the most widely technologies in this field. The devices provide information on the painting's materials, on the employed technique and on the conservation's state of the art work. However, all information must be correctly registered to be able to draw safe conclusions and perform the most adequate conservation interventions. We also present a complete example where multispectral visible images, IR reflectography, UV fluorescence and 3D data are acquired and then combined, showing how the integration gives a new and significant improvement in the analysis of painting.

The generation of 3D models of objects has become an important research point in many fields of application like industrial inspection, robotics, navigation and body scanning. Recently the techniques for generating photo-textured 3D digital models have interested also the field of Cultural Heritage, due to their capability to combine high precision metrical information with a qualitative and photographic description of the objects. In fact this kind of product is a fundamental support for documentation, studying and restoration of works of art, until a production of replicas by fast prototyping techniques. Close-range photogrammetric techniques are nowadays more and more frequently used for the generation of precise 3D models. With the advent of automated procedures and fully digital products in the 1990s, it has become easier to use and cheaper, and nowadays a wide range of commercial software is available to calibrate, orient and reconstruct objects from images. This paper presents the complete process for the derivation of a photorealistic 3D model of an important basalt stela (about 70 x 60 x 25 cm) discovered in the archaeological site of Tilmen Höyük, in Turkey, dating back to 2nd mill. BC. We will report the modeling performed using passive and active sensors and the comparison of the achieved results.

This paper shows the applicability of non-contact 3D imaging technology to the dimensional monitoring of wooden artworks. The results of a study on a wooden test artifact submitted to sudden environmental parameters changes (temperature and humidity) are presented. It was possible to verify that the range maps generated by a 3D camera based on optical triangulation have the necessary resolution to show dimensional variations in the order of a few tenths of a millimeter. 3D models generated before and after the parameter variation were processed with a specific software in order to highlight the amount of deformation through a color coded image. The results demonstrate that such technique represents a unique instrument to capture and track the deformations of wooden artifacts before they become permanent. This study has been conducted in collaboration with the laboratories for restoration of the "Opificio delle Pietre Dure" at Florence, Italy.

The digital documentation of monuments and architectures is an important field of application of the 3D modeling where both visual quality and precise 3D measurement are important. This paper proposes an integrated approach based upon the combination of different 3D modeling techniques for the virtual reconstruction of complex architectures like those found in medieval castles. The need of combining multiple techniques, like terrestrial laser scanning, photogrammetry and digital surveying comes from the complexity of some structures and by the lack of a single technique capable of giving satisfactory results in all measuring conditions. This paper will address modeling issues related to the automation of photogrammetric methods and to the fusion of 3D models acquired with different techniques, at different point densities and measurement accuracies. The test bench is a medieval castle placed in Trentino A.A., a tiny region in Northern Italy.

3D documentation and visualization of Cultural Heritage objects is an expanding application area. The selection of the right technology for these kinds of applications is very important and strictly related to the project requirements, budget and user's experience. Active sensors, i.e. triangulation based laser scanners and structured light systems are used for many kinds of 3D object reconstruction tasks and in particular for 3D documentation of cultural heritage objects. This study presents some experiences in the results of two case studies in which a close-range structured light system is used for the 3D digitization. The paper includes all necessary steps of the 3D object modeling pipeline from data acquisition to 3D visualization.

This paper describes techniques for analyzing 3D volume data by using extended Laws' convolution kernels. Laws' kernels are well known for 2D texture analysis, and have been used for various pattern recognition applications. Although typical Laws' convolution kernels are represented in 2D masks, we have extended the kernels to form 3D masks. The three dimensional extension of the masks allows a pattern recognition system to handle 3D volume data, whereas a traditional approach can handle only 2D image data. Also our approach can be extended for use with various lengths of kernels to generate multiple resolutions of masks. In our experiments, mask resolutions of 3 × 3 × 3, 5 × 5 × 5, 7 × 7 × 7, and 9 × 9 × 9 were tested.

Versatile three-dimensional (3D) image data as in the form of I(X,Y,Z) defined in the camera coordinate system are acquired from sets of color image and depth image, which are simultaneously captured at a video frame rate by a 3D camera named Axi-Vision camera. The X and Y coordinates are calculated from the image coordinates (x, y) on the image plane of CCD based on the perspective projection theory. The Z coordinate can be obtained for each pixel of the color image directly from the depth image without any complex processing. Thus, the 3D image data of objects, I(X,Y,Z), is obtained from the relation between the camera coordinate and the image coordinate systems. A stereoscopic video image display is demonstrated using the 3D image data using an integral photography (IP) system, which combines a 4- inch VGA liquid crystal color display panel and a pinhole array. Aliasing, an issue from mismatch of image qualities between the 3D image data and the IP system, is effectively suppressed by pre-processing the color Axi-Vision image data using a two-dimensional lowpass filter, which is designed based on analyses of the maximum spatial frequency of the 3D image transformed appropriate for the LCD size and the Nyquist spatial frequency of the IP system.

An ultra-high speed camera of 1Mfps was applied to visualize the crack propagation. Change of stress field around the propagating crack tip was captured as a change of the fringe pattern by means of the photo-elastic imaging technique. Newly developed video trigger system is employed to detect the occurrence of the crack propagation as a trigger in the experiment. The trigger successfully perceived the initiation of the crack propagation stably. Also its response time was fast enough even for the image capturing with 1Mfps. As a result, it is revealed that the elastic wave, propagating in the continuous body, has a significant effect on the velocity and kinking behavior of the propagating crack.

Since the 1990s photogrammetry has been used in forensic medicine to help identifying perpetrators from crime scenes covered with surveillance video. Some case studies [1,2] have shown promising results measuring both heights and segment lengths. Photogrammetry is used with high precision when measuring clearly defined points but less is known about the reproducibility of hidden body-points. In this study we quantified the inter- and intra-observer variability of bodily measures with the software PhotoModeler Pro in low resolution images. The study showed that the body height and shoulder height were reproducible within ± 1 cm and ± 2 cm respectively. The measurement error were markedly higher for all measurements between points hidden by clothes and measurements of segment lengths were only usable in the intra-observer situation when flexion in the joints were present. In the inter-observer study the measurement error was so high that few measurements could be used beside shoulder and body height. Measurements are suggested to be done with the perpetrator and suspect in the same pose.

A novel approach is presented for the location of discrete targets images with subpixel accuracy. It can be adopted as first stage of any method that uses planar and no-planar grids with circular, square or cross shaped targets in order to calibrate close range cameras. The approach has been performed on several calibration sessions each of them followed by the three-dimensional reconstruction of a reference object that was previously measured through a coordinate measuring machine. The performance of the proposed approach has been expressed in terms of deviation between the reference coordinates and the corresponding ones provided by triangulation. The results obtained have also been compared with those achieved through ShapeCapture^TM, a software widely used in photogrammetry for camera calibration and 3D model reconstruction from images.

This paper outlines a special microscope under development, named "Ultra-high-speed bionanoscope" for ultra-highspeed imaging in biological applications, and preliminary design of the image sensor, which is the key component in the system. The ultra-high-speed bionanoscope consists of two major subsystems: a video camera operating at more than 10 Mfps with ultra-high-sensitivity and the special microscope to minimize loss of light for seriously reduced illumination light energy due to the ultra-high-speed imaging. The ultra-high-frame rate is achieved by introducing a special structure of a CCD imager, the ISIS, In-situ Storage Image Sensor, invented by Etoh and Mutoh. The ISIS has an array of pixels each of which equips with a slanted linear CCD storage area for more than 100 image signals for reproduction of smoothly moving images. The ultra-high-sensitivity of the sensor of less than 10 photons is achieved by introducing three existing technologies, backside-illumination, cooling, and the CCM, Charge Carrier Multiplication invented by Hynecek.

Parameter optimization is an important aspect for achieving highest accurate results. This paper investigates different parameters of the one dimensional Log-Gabor filter approach to produce optimal recognition rates. ROC curves are adopted to measure the recognition accuracy. More than 20 million matches have been performed in our experiment. Our experimental results show that optimal wavelength is linear change with the normalized length in angular direction in the polar coordinates. This paper also finds the optimal phase number for pattern encoding, and the optimal radial length.