Low intensity, contributions of background light, reabsorption and amplification can distort the signal and prevent fast and accurate determination of the measured quantity in the fluorescent lifetime based sensors. Methods such as Prony's least-squares method or lock-in phase sensitive techniques are usually employed. They are all prone to errors, as it was shown by several authors. They can be improved, but at the expense of heavy computation. An alternative way to accurately determine decay rate from measurement is a simple digital signal processing technique presented in the article. It is based on modified digital lock-in technique, enabling on-line monitoring. Effects of DC offset, linear and exponential contributions on the proper decay rate determination are discussed in the article and effectively eliminated by the technique. Optimal process parameters are evaluated and experimental confirmation with the basic scheme given. A digital processing was implemented with software, but its simplicity enables simple digital hardware implementation.

Material processing by laser still offers a wide topic for investigations since interaction between light and material is very sensitive to inner and outside influences. Big efforts have to be taken to control the quality of laser cut edges, up to 6 mm. Using obtained optical signals from the cutting process efficient control highly depends on sophisticated mathematical methods. We present first contexts between two sensor signals of different wavelengths by the use of different mathematical algorithms. Important is the well done combination of these signals to achieve controlling rules both for the cutting speed and the laser power considering computation time for its evaluation.

In a previous paper we showed, that it is possible to increase the information of the measurement vector by extending the available number of measurements in the measurement vector y(k) with additional measurements. Theoretical considerations proved, that the additional measurements do not contain any new information, but information in another for, i.e. in another coordinate system. This new coordinate system is generated by a mapping of the original components of the measurement vector into the additional components. In this paper we are going to occupy ourselves with the restrictions and the preliminaries of such additional measurements. Therefore we will look at information in the possible mappings and we will give some hints how to create the optimal extension of the measurement vector.

In treating 2D arrays we may use either the cartesian coordinate system or the cylindrical coordinate system to describe pictures, fields, etc. Both descriptions must contain the same amount of information about the device, that is to be examined. As a consequence the joint distribution density of real and imaginary part of a vector and the joint distribution density of absolute value and phase of the same vector have to contain the same information. Approximating the phase density by a gaussian density we take the variance of the phase density to generate the gaussian density. In worst case this variance becomes (Delta) ²/12 (where (Delta) is the width of the phase interval (Delta) equals 2(pi) ). However, that would suggest that the phase data would contain information through the distribution density of the phase is an uniform density, where we cannot favor any value of the phase, i.e. we do not have any information about which value the phase estimate should have. In the following paper we are going to find a solution to this problem by examining the difference between information and inverse variance in the phase distribution.

During the last few years the Center for Sensorsystems (ZESS) develops an image guided 3D measuring system consisting of a 3D laserradar scanner and a CCD camera. The field of application are all kinds of industrial inspection and surveillance tasks where it is necessary to detect, measure and recognize 3D objects in distances up to 10 m with high flexibility. One main problem, which has to be solved during the development phase is the calibration of the single sensors and the multisensor arrangement for such a large working volume. This paper presents a flexible calibration procedure for the 3D multisensor system which makes it possible to determinate separately the outer and inner parameters of the camera and the laserradar scanner. Using the outer parameters which describe the position and orientation of the sensors, it is feasible to integrate the single sensors into a multisensor system. The calibration procedure for the image guided 3D measuring system is based on the method of bundle adjustment. For this we use a 3D arrangement of 3D calibration points which could be detected optimally both by the CCD camera and the laser scanner. It is necessary to known the position of all calibration points. We set store by examining the dependence between the accuracy of the calibration and the expense for providing the calibration normal and the accuracy of detecting the 3D points during the calibration. Beside the use of real sensor data, it is possible to work with data from a system model, which considers the sensors as well as the calibration normal and possible measuring errors. Because of the opportunity to simulate the complete calibration procedure it is practicable to find out an ideal arrangement of calibration points and sensor views as well as estimate the accuracy of the calibration and the steps which lead to a worthwhile improvement of accuracy. This paper gives a survey over the multisensor concept and outlines the mathematical models which are used to describe the single sensors and the complete multisensor system. Further we explain the calibration procedure and the relevant influences in detail.

This paper addresses an installed application in quality control where a 100% inspection of geometry (3D) and surface of cuboid (parallelpiped) 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, mounted in a ring around the object 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 due to spalling/chipping and for 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. All components are triggered by separate light barriers and perform their tasks independently. The integration of the results of each measurement is done by an SPC which also controls the actors that handle the three different classes of objects (good, bad, rework). These actors are valves and the objects are separated by pressurized air. The main concern of this paper is the system aspect, how the measurement results are evaluated and combined to achieve a correct classification of the objects which are inspected by three independent sensors and arrive at unpredictable time intervals.

One way to reduce the increasing waste streams of used polymers is an efficient material recycling. This requires a technology for the separation of polymer mixtures into different material fractions. For this purpose the principal suitability of laser-induced breakdown spectroscopy was investigated. Plasma emission spectra of LDPE, HDPE, PP, PET, PVC, and PS were studied. Basic investigations were performed in order to assess the influence of different measurement parameters and to optimize the analytical performance. More than 140 spectra lines are identified, which can be related to C, H, O, N, C₂, CN and CH from the bulk material and the atmosphere and to Al, Ca, Cu, Fe, Mg, Sn, Ti and Zn from additives of the polymer. Estimated detection limits of down to 2 ppm are achieved for metallic additives. Different artificial neural networks were tested for the evaluation of the spectra. PET and PVC can be identified unambiguously detecting the characteristic elements oxygen and chlorine. For plastics, which differ in their contents of inorganic additives, the line emission of additives can be used as `fingerprints' of the plastics. In this way identification accuracies of 87% to 100% for PE, PP, PET and PVC are achieved.

A two-channel synchronous receiver circuit for optical instrumentation applications has been designed and implemented. Each receiver channel comprises a.o. transimpedance preamplifier, voltage amplifiers, programmable feedback networks, and a synchronous detector. The function of the channel is to extract the slowly varying information carrying signal from a modulated carrier which is accompanied by relatively high levels of noise. As a whole, the channel can be characterized as a narrow band filter around the frequency of interest. Medical applications include arterial oxygen saturation (SaO₂) measurement and dental pulp vitality measurement. In both cases, two optical signals with different frequencies are received by a single photodiode. The measured performance of the optical receiver shows its suitability for the above mentioned applications. Therefore the circuit will be used in a small sized, battery-operated sensor prototype to test the sensing method in a clinical environment. Other applications include the signal processing of optical position-sensitive detectors. A summary of measured receiver channel performance: input reduced noise current spectral density between 0.20 and 0.30 pA/(root)Hz at all relevant frequencies, total programmable channel transimpedance between 7 M(Omega) and 500 M(Omega) , lower -3 dB frequency of at least 50 Hz, upper -3 dB frequency of 40 kHz, maximum voltage swing at the demodulator output of 2.4 V.

Implementations and test results of one single-axis and 2- axis CMOS PSDs as well as a BiCMOS integrated receiver channel are presented. The single-axis PSD has a conventional LEP structure. It uses the well-substrate junction as a photodetector and pinched well as current dividing layer. An interelectrode resistance, NEP and position sensing accuracy of 152 k(Omega) , 1.6 pW/(root)Hz (850 nm) and 0.1%_rms are achieved with the detector measuring 5 X 0.2 mm². The first 2-axis PSD is a tetralateral LEP but instead of having strip-like continuous edge-electrodes it has electrodes composed of 100 discrete contacts, each of which is connected to the output current line using MOS switches. Linear position response is provided by disconnecting one opposite pair of the electrodes from preamplifier inputs and measuring one dimension at a time. An interelectrode resistance, NEP and accuracy of 4.5 k(Omega) , 10 pW/(root)Hz and 0.07%_rms were achieved with this PSD using the alternate measurement mode. The second 2-axis PSD has an operating principle similar to a basic LEP but is composed of an array of phototransistors and polysilicon resistors. The NEP and position sensing nonlinearity of the sensor were 0.5 pW/(root)Hz and 0.04%_rms, respectively. The third 2-axis PSD has the same construction as the second one but the phototransistors and polysilicon resistors were replaced with well-substrate photodiodes and pmos transistors, respectively. By driving one or more adjacent transistors in off-state the LEP mode can be changed to a segmented PSD mode providing the means to combine the high lateral sensitivity of the segmented mode with the large and linear measurement fields of the LEP. The BiCMOS receiver channel is composed of a transimpedance preamplifier, voltage amplifiers, gain control and offset cancellation blocks and a synchronous detector. The transimpedance at four different gains for a signal frequency ranging typically from 5 kHz to 10 kHz were 7 M(Omega) , 33 M(Omega) , 143 M(Omega) and 488 M(Omega) . The measured noise current density was lower than 0.3 pA(root)Hz, and the area and power consumption were 2.9 X 0.45 mm² and 37 mW, respectively. As the achieved results fulfill the demands set for a typical signal conditioning channel of a PSD sensor system, the properties appear to be suitable for integrated PSD systems.

The implementation and test results of a 16 by 16 binary CMOS photodetector array for position sensing applications are presented. Unlike conventional position sensitive detectors (PSDs), it can be used for simultaneous multiple spot detection with high accuracy. To decrease the signal processing overhead characteristic of this kind of area arrays, binary detection and random access readout is used. The fill factor and pitch of the array are 30% and 50 micrometers , respectively. Spatial quantization limited position sensing accuracy of 4.3 micrometers is achieved using a uniform spot of 280 micrometers in diameter, total signal power of about 8 nW and pulse width of 8 ms. At lower signal levels, the accuracy is decreased by the spatial noise caused by the nonuniformity of the threshold currents. The array outperforms a typical large-area (approximately 100 mm²) lateral effect PSD (LEP) when a measurement bandwidth smaller than 10 kHz is used.

The displacement sensing accuracy of the reflected beam method in turbulent environment is studied experimentally. The method includes an illuminated corner cube retroreflector (CCR) and a lateral effect photodiode (LEP) by which the image position of the CCR is detected. The results show turbulence (C_n approximately 5*10^-7 m^-1/3) limited measurement resolution to vary from 0.6 to 9 mm (standard deviation) in the distance range from 50 to 300 m. The result is 1.4 to 2.1 times worse than that measured for the direct beam sensor in the same atmospheric conditions. The type of CCR used is found to have no systematic effect on the achievable resolution but the method is found very sensitive to receiver defocusing. An approximately tenfold increase in standard deviation is observed when the receiver is defocused by 2%. The possibility to average out part of the angular fluctuations by using multiple CCRs is verified. By using two CCRs instead of one, when positioned 1.3 m apart from each other at the distance of 150 m, a resolution better by 22% is achieved. The averaging in case of the LEP, however, is not effective because the noncorrelated intensity fluctuations of the reflections from the different CCRs cancel the averaging effect.

In this paper the high frequency behavior of integrated pn- photodiodes is discussed and measurement results of two different types of photodiodes, one implemented in a standard 1.2 micrometers BiCMOS process and the other in a 0.8 micrometers CMOS process are presented. The rise times and responsivities of the photodiodes are under 5 ns and 0.26 A/W in the CMOS process and about 30 ns and 0.23 A/W in the BiCMOS process, respectively. Furthermore, the suitability of the technique for 3D vision has been investigated by designing an array of photodetectors and measuring the isolation between detector blocks.

An adaptive, selfdiffractive image sensor with space resolution greater than 11,000 lines/mm is presented. Practically unlimited reversibility in continuous hologram registration regime without pretreatment (chemical, temperature, electrical field etc.) allows creation of a new portable interferometers with high I_signal/I_noise ratio. The advantages of proposed systems for diffusely reflected objects are discussed.

The performances of a CCD have been evaluated in a very wide spectral region, which comprises the near IR, the visible, the near and far UV, EUV and soft x-ray spectral regions. The CCD detector is a back illuminated one, 512 X 512 format, 24 X 24 micrometers ² pixel. The measurements performed consist mainly on the determination of the quantum efficiency and of the uniformity of response in the 1 - 1100 nm spectral region. Three different experimental setup have been used for the various spectral range: a Czerny-Turner monochromator for the 1100 - 250 nm region, a Johnson-Onaka monochromator for the 250 - 30 nm region, and a grazing incidence monochromator for the 30 - 1 nm region. The tested CCD exhibits high values of quantum efficiency and a good uniformity of response in the analyzed spectral range.

A method of color image recording by photothermoplastic media consists in the recording of optical data in a mode of imaging storing of a total array of color-divided components with its later development.

An all electronic double-pulsed TV holography system is described. Two separate image-plane holograms are recorded using a CCD camera. A small angular offset between the reference and object beams introduces carrier fringes in the image plane. Thus, the Fourier spectrum of the image contains distinct parts that can be filtered out and inverse transformed to yield the phase information. Experiments showing propagating transient bending waves in an aluminum plate generated by a focused laser pulse are performed. Phase maps of good quality are unwrapped and presented as 3D plots of the out-of-plane displacement. The system has also been used to evaluate the 3D distribution of transient acoustic fields in air generated by an electrical discharge. Several projections of the acoustic field is recorded and the 3D pressure field is reconstructed by computerized tomography.

Random access active pixel CMOS image sensors generally suffer from non-uniformity in their pixel outputs. This document describes a simple mixed analogue-digital integrated circuit for fixed-pattern-noise compensation. The method has been applied to the range of sensors developed by IMEC, and improves their operation beyond mere static noise suppression.

Talbot effect and interferometrical fringes phenomena combined with computer-generated moire are employed in order to enhance the detection of the displacements of one surface. The high sensitivity achieved, easily assembly and the possibility to change automatically the reference pattern tourn competitive such methods as compared to other but complex. The range of application for each case is estimated.

In this paper, the new method on evaluating and measuring roughness of optical surface by Fourier spatial frequency analysis is presented. Authors have gotten the equipment, in which the electron photomicrographs of optical surface is scanned and analyzed by CCD camera--microcomputer system. The new method have both good virtual and lateral resolution, fast scanning speed and better measuring accuracy.

A simple, single-element, afocal, refractive optical device with two aspheric surfaces has been designed and fabricated for transformation the Gaussian intensity profile of a He-Ne laser into a collimated beam of uniform profile. The working principle, the method of design, the method of fabrication are presented. Optical and geometrical properties of the fabricated sample have been tested. Device parameters and simulated behavior are compared with test results in detail.

In this paper a theoretical and an experimental study of a sensor which uses a self-mixing interference effect inside a microchip laser are reported. The so-called FMCW approach has been implemented by using a intracavity electrooptic modulator which provides a total 10 GHz frequency-shift of the optical wave emitted by a Nd³⁺:YAG-LiTaO3 microchip laser. We demonstrated that the resonant behavior of the microchip laser provides a substantial amplification of the return wave and that such a phenomenon can be used to perform a highly sensitive detection for absolute distance (from 1 to 20 m +/- 0.1%) and velocity (from 0.1 to 25 m.s^-1 +/- 0.1%) measurements of a remote scattering target.

Triangulation is a method often used to register absolute coordinates of an object. The basic idea of this method is to measure the 3D coordinates by structuring the illumination using spectral or spatial light encoding. This technology has been established in the field of quantitative measuring free-styled surfaces, including the generating of CAD-description of the test-object (reverse engineering). For fringe-projection a periodic grating with constant frequency is mainly used. Advantage of this kind of grating is that it is easy to produce. But sometimes the measurement fails since shadows, high density of the fringes, edges, overmodulation, reflection and blurred regions of the object refuse exact evaluation. This paper describes a method using fringe-projection-technology which is more robust. This aim is accomplished by monitoring a programmable, high- resolution, high-contrast spatial light-modulator using information from the result of projection. Consequently the main component of this process is a programmable light modulator based on LCD. First step is the detection of problematical regions on the object caused by shadows, overmodulation, reflection, edges of the object and high density of fringes. This analysis is based on skeletonizing the projected fringes in the image. After detecting the faulty areas on the object the pattern, frequency and intensity of the projector is modified for the detected area only until the best result is guaranteed. With these modifications to the conventional fringe-projection- technology a fast, precise and complete shape measurement is achieved.

The paper presents a phase unwrapping algorithm based on a linearized extended Kalman filter. The Kalman filter exploits a so called `Basic-Slope Model', enabling the filter to incorporate additional information about the mean local phase variation. This phase variation will be estimated directly from the sample frequency spectrum of the interferogram by a local slope estimator. Additionally, the local slope estimator calculates the error variance of the estimated phase variation by using Fisher's information. The Kalman filter then fuses the information directly obtained from real and imaginary part of the interferogram with information from the local slope estimator. The paper outlines the principle operation of the phase unwrapping algorithm and explains the cooperation of the Extended Kalman filter with the local slope estimator. At last the efficiency of this phase unwrapping algorithm will be shown by simulations and real InSAR images.

One of several methods to demodulate noisy, angular modulated signals is using Kalman Filtering techniques, offering a combination of two advantages: demodulation and filtering in one step. Different Kalman Filter models have been examined to achieve good filtering and error-reduced demodulation results without consuming too much time. The models usually use at least two inputs for filtering--the quadrature components and supplementary inputs. Using the quadrature components, a more effective exploitation of the given phase information is achieved. The principle of data fusion is increased furthermore by adding other relevant input signals. The filtering results become more precise but unstable at the same time. Smoother filters are expected to increase the stability of the Kalman Filter. Filtering from two directions to determine one estimated phase value, leads to a better detection of outliers and keep the control parameters stable. Simulations and Online-Filtering is done to work out the best fitting model for this problem.

A fringe projector based on a low-cost spatial light modulator has been used to measure the shapes of discontinuous objects. By changing the fringe phase and the fringe pitch, a sequence of wrapped phase maps can be acquired at different sensitivities. This sequence can then be converted to surface profile by the recently-proposed method of temporal unwrapping rather than by a conventional spatial unwrapping approach. The main advantages are that the method is simple and robust, and that objects with surface discontinuities are profiled as easily and accurately as smooth ones. Modifications to the basic temporal unwrapping method are proposed which provide significant improvements in reliability, accuracy and computation time.

A method to reduce the sensitivity of phase-shifting interferometry to external vibrations without sacrificing lateral resolution is described. The returning interferogram is amplitude split to form two series of interferograms; a fast sampled, low spatial resolution data set and a slow sampled, high spatial resolution data set. The fast sampled set is used to calculate the true phase increment between intensity measurements in the high spatial resolution data set and a generalized phase extraction algorithm then uses these phase increments when calculating the topographical phases for this set. The measured topography thereby benefits from the best qualities of both data sets, providing increased vibration immunity without sacrificing high spatial resolution.

Triangulation systems that are based on an autosynchronized scanning principle to provide accurate and fast acquisition of 3D shapes are able to scan large fields. It is done generally by a coordinate measuring machine (CMM) carrying a small-volume 3D camera. However the acquisition speed is limited by the CMM movement and also by the image fusion time required to get the complete 3D shape. This paper describes some practical consideration for large volume 3D inspections with emphasis on telecentric scanning. We present the analytical and the optical design of a large telecentric scanner using a large reflective surface. Some results of the laboratory prototype will be presented. We also discuss applications and the viability of this new approach.

A number of sensors exist which measure 3D range information using both active (projected light patterns or laser scanning) or passive (shape from shading or stereo camera) methods. For many applications, use of active systems is impractical, and the range of depth achievable with a stereoscopic camera system is limited to the baseline separation of the cameras. Presented here is a passive 3D range sensor based on a pupil-division and signal correlation scheme. The method for the extraction of range information compares signals from the upper and lower entrance pupil halves of the objective. The phase of the signals obtained from the correlation of light from both pupil halves with an oscillating optical grating are compared, and the result provides the range information. Using a microlens and sensor array, this system has been expanded to analyze 49 channels in parallel in the field of view. Preliminary studies indicate that the system is capable of measuring ranges from 2 to 10 meters with a precision of less than one percent. The measurement speed is limited by the mechanical oscillation frequency of the grating and is currently in the range of 1 - 3 frames/second. A single channel experimental system has been completed and a 49 channel prototype will be completed during the summer of 1997. Presented here are details regarding the development of this sensor. Suitable theoretical background to understand the system is provided. Results of the prototype system and applications of sensors that extract range data from a 2D field of view and a 1D strip will also be presented. Planned future developments to miniaturize and improve speed performance will also be outlined.

This paper presents a CCD line scanning method, which allows the parallel measurement of up to six length measurement channels with a scanning rate of higher than 1000 Hz per channel. Mathematical methods--interpolation between the measured points--allow the combination of the results to find approximately the curvature of a moving surface or the misalignment of the sensor respectively. The sensor could be fitted in a compact case and works in a measurement range of 15 mm with a resolution better than 10 micrometers . A subpixel resolution of 1 micrometers can be achieved by computation. The requirements to high-speed operating sensors with high accuracy for length measurement are rising constantly. The instrumentation may readily be integrated into the production process (in-process measurement), e.g. for the ends of quality assurance and to make an automatic compensation of deviations possible. A very important fact is that the advantage of contactless measurement is avoiding the wear and tear of the probe tip, the material, respectively, on the other hand it reduces the error of measurement results from a thin film of cooling oil and chips. High pressure jets make sure that the workpiece surface in the measurement areas are blown free. Variations of the ambient temperature doesn't effect the results since the instrumentation is in a housing at stable temperature. Inductive and capacitive measurement methods have problems with curved surfaces. These methods are only used with expanded methods of mathematical compensation on these circumstances. Therefore the optical sensors are very interesting for achieving the requirements of the in-process measurements.

3D-measurement systems based on active triangulation using projection of structured light and recording by CCD-cameras are gaining more and more importance in industrial application. Their advantages compared to conventional coordinate measurement machines are the non-contact measurement and the fast acquisition of dense point clouds. This paper presents an alternative method to the well-known combination of Graycode and phase-shift projection. It will be shown how to use the heterodyne principle for overcoming the unwrapping problem of phase functions and for estimating the parameters for a reliable unwrapping procedure. The redundant phase information is used to increase the accuracy of the photogrammetrical 3D-coordinate determination. Finally, the heterodyne principle is used to realize on-line measurement based on color-coded fringe projection.

This paper presents first hardware implementation and investigations of a new electro-optic modulator (EOM), called the Photonic Mixer Device (PMD). The PMD is a semiconductor device combining the characteristics of fast optical sensing and modulation. Arranged to a PMD-matrix it looks like a CCD-matrix but additionally provides the depth information of each image pixel using an appropriately modulated scene illumination. Besides this feature of 3D- imaging by means of time-related correlation the PMD-chip will enable on-chip 2D-image processing by means of spatial correlation. The principle of operation of the PMD, possible technologies for realization, facilities and applications will be described. This new device offers high potential for optical sensory systems due to an amazingly simple and powerful procedure of electro-optical mixing and correlation. Both CCD and CMOS are appropriate technologies. The latter will be treated in detail as a single-element Photonic Mixer Device. Here we propose several architectures--including two quadrant (2Q)-PMDs--with readout and preprocessing circuits for both, the phase resp, time-of-flight values and the pixel intensities. Arranging PMD-pixels to a PMD-line or to a PMD-matrix will provide a new generation of flexible and powerful solid-state 3D- cameras based on time-of-flight. According to the EOM- principle, the proposed 2D-mixer requires no additional optical devices, no broadband electronic amplifiers and mixers. This solid-state PMD-array offers even more unique facilities, e.g., performing high speed spatial light modulation up to the GHz range or optical CDMA- communication. The modulation characteristic of the PMD is an important aspect of this mixer. It is associated with optimizing the layout design for the PMD, which will be presented and discussed in this paper, too. PMD test chips have been realized in CMOS technology. Some simulation and first test results of the chip are also presented. The simulation results prove the operation principle of the PMD and provide us with parameters for an optimized layout design. First experimental results verified the expected operation principle of the test PMD.

For advanced performances related to 3D metrology in production, security and traffic control, the Institut fur Nachrichtenverarbeitung is currently engaged in the investigation of a new 3D-camera system based on rf- modulation interferometry, other than triangulation. The new 3D-camera system proposed in this paper uses the techniques of optical rf-modulation of the active illumination source and the correlation of these signals reflected from a target within a large aperture. This method yields all phase- correlation functions, relating to each voxel of the 3D- scene in parallel. Hence, the system delivers a fast parallel measurement and evaluation process by completely using a 2D transmitting and receiving path. To fulfill this concept, we have set up a binocular system using a modified Pockels-cell arrangement and a modulated laser transmitter. Since the electro-optical modulator plays a key important role in realizing the whole system, we would discuss the modified arrangement of stacked birefringent KD*P crystals that incorporates the properties of large aperture and high frequency modulation in the region of some tenth of megahertz. Additionally, an optical reference path is introduced to compensate for the fluctuation of the light source, thus enhancing the capability of the system. The experimental results of this reformed system will be presented in this paper, e.g., the standard deviation of (lambda) /2500 being obtained.

The measurement of 3D object shapes for the purpose of digitization of CAD-models and for the complete manufacturing control of components are important tasks of modern industrial inspection. The proposed 3D measurement system using structured-light illumination has the ability to avoid illumination-caused difficulties, like shadowing and excessive light intensities by light reflection and diffraction at the surface of the object, while measuring technical surfaces. For this purpose, the object under test is successively illuminated with a periodic grating structure from at least three different directions, using a telecentric projection system. At least three linearly independent phase-measurement values are measured by gray- code techniques to calculate the 3D coordinates of the object points. The experimental setup allows the determination of phase-measurement values with illuminations from up to 16 different directions. This is connected with a simultaneous variation of the intensity of the projected grating structures. Thus, areas of shadows are `shifted' across the object surface to spots where they have no influence on the result of the measurement, and also specular effects can be suppressed. Furthermore, in order to obtain the entire surface, the object to be digitized must be covered by many overlapping views taken from different directions. To view the entire surface, the object is moved into various measuring positions, using a second rotation axis. These views are merged within an object-centered coordinate system and are automatically rearranged into a uniform grid. For this purpose, a calibration procedure has been developed to measure absolute coordinates within a defined object coordinate system, so that the combination of the particular images is simple, because all measurements are performed within the same system of object coordinates. The power of this concept has been experimentally demonstrated, for example, by measuring the complete 3D shape and/or the roughness of specularly reflecting technical surfaces of different volumina.

A prototype of a liquid level gauge based on pulsed time-of- flight method was developed for measuring liquid level accurately at distances from zero to 30 meters. The system consists of an optomechanical sensor head and electronics unit which are connected to each other by means of two optical fibers. The developed level gauge utilizes mirror- like reflection of the liquid surface and by proper design of the optics the received signal is constant in the whole measurement range. The losses in the optical path of the level gauge are relatively small, thus allowing the use of a 1 mW CW semiconductor diode laser as a light source. The pulsing frequency of the laser is 1.5 MHz and the time interval between start and stop pulses is digitized by means of a simple and inherently linear synchronous digital counter. The correct orientation of the measurement head is achieved by using sensitive bearings and by allowing gravity to keep the correct orientation. An attenuation system based on eddy currents effectively attenuates the oscillation of the sensor head. Measurements carried out in laboratory show that the accuracy of some millimeters can be achieved with the measurement time of some seconds.

With this paper we would like to present an intelligent low- cost-camera. Intelligent means that a microcontroller does all the controlling and provides several in- and outputs. The camera is a stand-alone system. The basic element of the camera is a linear sensor that consists of a photodiode array (PDA). In comparison with standard CCD-chips this type of sensor is a low cost component and its operation is very simple. Furthermore this paper shows the mechanical, electrical and electro-optical differences between CCD- and PDA-sensors. So the reader will be able to choose the right sensor for a particular task. Two cases of industrial applications are listed at the end of this paper.

The paper presents an innovative instrumentation for in- process measurements of the actual diameters of workpieces with rotational symmetry. This instrumentation may readily be integrated into turning or grinding machine, e.g. for the ends of quality assurance and to make an automatic compensation of deviations of diameters possible. It consists of two optical Fresnel diffraction sensors mounted diametrically to the nominal axis of the workpieces under process such allowing a fast measurement of their actual diameters with errors less than +/- 5 micrometers at scanning rates of more than 1 kHz. An automatic follow-up system, based on grated glass scales has been constructed to achieve the requirements of a resolution of 1 micrometers up to diameters of 300 mm. Special high precision guide rails and bearings have been constructed to minimize the effects of imperfect guiding to the over-all accuracy. Variations of the ambient temperature doesn't effect the results since the complete instrumentation is in a housing at stable temperature. High pressure jets make sure that the workpiece surface in the measurement areas are blown free from oil, cooling fluids, and chips.

A novel image processing method is presented that allows to obtain images with maximal contrast by means of fusing a series of images which were acquired under different illumination constellations. For this purpose, the notion of illumination space is introduced, and strategies for sampling this space are discussed. It is shown that the signal of interest contained in a physical texture often would be lost if standard image acquisition methods were used. In contrast to this, the presented approach shows a robust and reproducible way to obtain high-contrast images containing the relevant information for subsequent processing steps.

In typical optical sensor setups for CO₂ laser material processing scraper mirrors or beam selectors (ZnSe mirrors) are used to allow sensor positions which have a coaxial path for both the laser beam and the monitored visible or near- infrared signal. Calculations and experiments have shown a vital influence of the focusing element for the laser beam on the measured signal. Critical topics are the distance between focusing element and monitored process, as well as the dependence of focal length on wavelength.

In the center for sensor system at the polytechnic university of Siegen an image processing system for the measurement of torsions of special section tubes has been developed. Especially considering the tube geometry and the areas of the profiles used for the determination of torsions, an angular resolution below 0.25 degrees results. The measuring system is applied within the scope of quality control.

Due to the changes on the international market that derive :from the introduction of product liability and the augmenting international competition in relation with an increasing awareness for quality, quality inspection of industrial manufacturing processes gains more and more importance. The growing automation of manufacturing processes as well as the introduction of flexible manufacturing systems also require the automation of the quality inspection. There are various reasons for that: • The manual-visual control is liable to high slip rates of undetected faults. • Human beings judge subjectively. Thus, the reproducibility of the results is low. • At high clock-pulse rates, human beings are unable to keep up with the production clock-pulse. • Visual control is highly monotonous work. • The documentation of the inspection results requires great effort. Today, image processing systems are being used for manifold tasks in industrial manufacturing processes. They can be used to identify workpieces as well as to control the flow of material. In automatic handling processes image processing systems control the position of the objects. When used for quality inspection they are employed for measuring, surface control and the control of assembling processes.

In this paper we would like to present a new idea for measuring high precision tubes with an accuracy of +/- 10 micrometers . The most interesting values of the tubes are wall thickness, diameter, eccentricity and so on. The cycle of measuring is not more than 5 seconds inclusive the complete tube handling. The preferred sensor concept is the triangulation technique with a matched sensor head.

Within the scope of an extended monitoring of the production a 100%-control of the workpiece is absolutely important. Especially the measurement of inside contours in small drillings and hollows is a problem in various industrial areas. Dimensional reasons are already enough to exclude most of the tried and tested methods for the inside measurement in the workpiece control.

Introduction of a compact sensor system to detect abnormalities on high graded, polished surfaces in production process. Usable for TQM in line of coating quality of lenses, glass plates, wafers or other high quality products. Optimized for non destructive, high speed scanning (2.5 m/s) of transparent materials with a low reflection rate and a resolution down to some micrometers 's. Reachable even in a noisy industrial environment. Available in a 19' rack with profibus data-link.

This paper deals with frequency modulation radar ranging principles applied to an incoherent multi channel fiber optical laser radar (ladar) system. The amplitude of the power output of a semiconductor laser beam is linearly frequency modulated (chirp). The laser beam is guided into an optical fiber and then splitted in as many channels as needed. The light leaving the fibers illuminates the objects under surveillance. The backscattered light from the objects is coupled back into the fibers, guided to a single photo detector diode and converted to an electrical signal. Mixing this signal with a reference yields a new signal that includes explicitly the distance information in it's frequency components. The basic analysis techniques of this system are derived from estimation theory. Frequency as well as phase of the mixed signal are used to extract distance information. It is shown that the additional information in the phase of the signal lowers the variance of distance measurements.

The principle of structured light 3D vision sensing based on spatial encoding image is presented. The spatial encoded images is produced by the means of laser scanning and modulating according to time-sequences. Is realized its key technology, i.e. synchro control among the camera, laser diode modulation and scanning polyhedron. The spatial image encoder is developed.

In a conventional bistatic LIDAR system, a laser pulse is generated and is reflected by the target. The round trip time determines the position of this target. In a greatly scattered medium, like water, the returned target signal is hidden by the random backscatter due to the existence of the suspended particules within the ocean mass. The improvement of the target detection sensitivity is possible in this backscatter clutter by modulating an optical pulse laser. We propose a theoretical study of the behavior of a modulated pulse in the underwater medium. We show that the randomly distributed backscatter does not maintain coherency of the modulation envelope whereas the target signal keeps this information. Due to this, the detection of the modulation envelope tends to suppress the incoherent backscatter clutter thereby improving the target contrast. We present a bistatic scheme detection in underwater medium, simulate it and compare results with a laboratory experiment. Simulation of the laser pulse propagation takes into account the single scattering approximation. Numerical integrals are calculated with an original technical based on the division into cubes of the definition domain. Detection of the modulated envelope is based on a fast frequency filter. The optical source used in the experiment is a 1500 nm laser diode controlled by a modulated current (500 Mhz). The ocean mass is simulated with a plastic fiber. A short pulse (10 ns) is transmitted into fiber. The returning pulse is collected by photodiode and analyzed with signal processing method. Experimental results show the efficiency of the modulated signal by a gain of the SNR ratio.