An earlier paper in this conference series, described the hardware and software for a Remotely Operated Prototyping Environment (ROPE). This enables a vision engineer, working a long way from the site of a proposed installation, to experiment with an image-acquisition sub-system that is already in situ. Using ROPE, a vision engineer is able to control the lighting, optics, camera, and object presentation, without ever leaving his own office. During the last 24 months, progress has been made, in enhancing ROPE’s facilities, most notably its control functions, user interface and the display of video image sequences. A real-time person-to-person communication system, intended to allow the vision engineer to perform a virtual tour of a remote factory, is also included. Recent developments in the associated image processing software (CIP) are the subject of a separate paper at this conference.

Following many years of development, interactive vision systems have become an effective tool for problem analysis and algorithm selection for industrial inspection, process monitoring and machine / robot control. Until recently such systems were inevitably based upon a stand-alone computer. However, more recent work has highlighted the need for a system that can be operated via the Internet. Such a system, called CIP, has been developed by the authors and was reported in a recent paper at this conference. This article describes enhancements of that system, including: the ability to write scripts, extensions to the user interface and command repertoire, as well as the ability to control the program remotely. Very recently, an interface to a shareware version of Prolog, written in Java, was completed and provides the ability to write intelligent programs for analysing visual scenes. CIP forms an integral part of a larger suite of programs that is being developed by the authors and their colleagues, for the remote acquisition and processing of images from a factory production environment. The latter is described in an accompanying article in this conference.

This paper presents an updated version of a general method for carrying out binary template matching, which is useful for image analysis in general and automated visual inspection and quality control in particular. In a series of 23 papers, image processing implementations based on the SKIPSM (Separated- Kernel Image Processing using finite-State Machines) paradigm have been shown to be faster or much faster than conventional implementations. One of the earliest of these papers, published in 1994, was devoted to binary template matching of various types. As with all the papers of that era, the theory was presented in general form but the specific applications used LUTs (lookup tables) and pipelined hardware. The results were impressive - templates 35x35 or even larger could be executed in the same time as the identical hardware, programmed conventionally, could execute a 3x3 template. This paper develops and extends the same basic approach to provide fast and highly efficient binary template matching on ordinary desktop computers. This implementation does not use LUTs, because computers with pipelined instruction streams and vector data structures perform relatively slowly when using LUTs.

Shot peening is a technique used to increase the flexural fatigue strength of machine parts which are heavily loaded by alternate bending. The impacts of the projectiles induce a compressive strain tangential to the surface which increases its endurance limit. To achieve a defined surface coverage with projectile impacts, this process has to be calibrated by measuring the surface coverage as a function of time. Up to now, this is done visually by inspecting test surfaces with a microscope. Following, the surfaces are compared with a catalog of reference patterns showing different coverage factors. This paper presents a model- based technique enabling an automated inspection of shot peened surfaces. For this purpose, an image series is recorded by varying the surface illumination systematically, and subsequently the series is fused to a symbolic result describing the areas showing shot impacts. Thanks to the simultaneous analysis of signal intensities in illumination space, no consideration of neighboring pixels is necessary to classify each single surface point. However, to assure a consistent result, additional constraints are introduced. Thus, a robust and precise detection of the interesting areas is attained.

We describe an application of color-based pattern matching, where a real-time vision system needs to detect and exactly localize textile patterns woven into carpet flooring material. These patterns are distributed on a large web in a periodic fashion. The task to be solved is recognition of these patterns by matching them with stored prototypes, computing the exact location and use this information to guide a cutting machine to produce perfect replica of desired tiles. The pattern matching part is challenging because of the presence of distortion, scaling, and rotation of the 2D patterns, and rather high demands on the localization accuracy. Also, the task needs to be solved under real-time constraints. We describe the building blocks used in our system. These are color-based segmentation of the patterns to achieve 2D representation in a graph-like manner, followed by graph-based matching. This block solved the graph-isomorphism problem in real-time tolerating distortions, additions, deletions, rotation, translation, and scale variations between the trained and tested versions of the patterns. We demonstrate the concept showing example images and matching results.

This research investigates the techniques using the image subtraction to find the discrepancy between the healthy and illness MRI images. The technique developed in this research moves the healthy MRI image to overlap with the illness MRI image. Then, the healthy MRI image and the illness MRI image are aligned to the same orientation. After the healthy MRI image overlapped with the illness MRI image, the illness MRI image is subtracted from the healthy MRI image. If there is discrepancy in the illness MRI image, after the image subtraction, the discrepancy will remain in the subtracted result. From beginning to end the inspection is done by the machine automatically. There is no further human effort involved. The technique developed in this research can very accurately find the discrepancy of the healthy and illness images. This paper explains the method using the second moment to find the orientations of the MRI images. By the orientations of the MRI images, the healthy MRI image and the illness MRI image can be aligned to the same orientation. The detailed process of image rotation is addressed in this paper.

Recognition of license plate images is the topic of this paper. The major issue in this problem is the accurate extraction of the license plate character string from varying backgrounds using processing techniques that are reasonably fast. The images are also characterized by non-uniform illumination. Recognition of the string is relatively straightforward, if the extraction process has been correctly designed. The authors present three different approaches to extraction and their study revealed that the combination of gray-scale morphology with a log gray-scale transform provided accurate extraction of the string. Recognition studies with 700 images captured during day and night periods indicated an overall acceptance rate of nearly 90% with .7% confusion.

Canadian Coast Guard has the mandate to maintain the navigability of the St.-Lawrence seaway. It must prevent ice jam formation. Radar, sonar sensors and cameras are used to verify ice movement and keep a record of pertinent data. The cameras are placed along the seaway at strategic locations. Images are processed and saved for future reference. The Ice Images Processing Interface (IIPI) is an integral part of Ices Integrated System (IIS). This software processes images to extract the ice speed, concentration, roughness, and rate of flow. Ice concentration is computed from image segmentation using color models and a priori information. Speed is obtained from a region-matching algorithm. Both concentration and speed calculations are complex, since they require a calibration step involving on-site measurements. Color texture features provide ice roughness estimation. Rate of flow uses ice thickness, which is estimated from sonar sensors on the river floor. Our paper will present how we modeled and designed the IIPI, the issues involved and its future. For more reliable results, we suggest that meteorological data be provided, change in camera orientation be changed, sun reflections be anticipated, and more a priori information, such as radar images available at some sites, be included.

A microscopic 3-D shape measurement system based on a new high-speed phase shifting technique is developed and experimented. By taking advantage of the color channel switching characteristic of a commercial Digital Micromirror Device (DMD) based video projector, a potential 3-D shape measurement speed of up to 100 frames/sec can be achieved. A computer generated color fringe pattern is projected onto an object surface. The red, green and blue channels of this pattern are programmed to be sinusoidal fringe patterns with 120 degrees in phase difference. When the color filter of the projector is removed, three grayscale fringe patterns are actually projected in sequence with a cycle time of approximately 10 ms. Through a stereomicroscope, the fringe patterns generated by the DMD is projected onto a small surface area of the object and then captured by a CCD camera via the same microscope objective. The 3-D microscopic shape of the object surface is reconstructed by using a phase wrapping and unwrapping algorithm. Experimental results demonstrated the feasibility of this technique for high-speed surface profile measurement.

A problem common to automated assembly in manufacturing or in automated docking of spacecraft is angular and lateral alignment of components. A hybrid video system utilizing both conventional imaging and 3-D video moiré has been developed to automatically align a test target with three translational and two rotational degrees of freedom. Alignment was demonstrated via computer controlled translation and rotation stages. The video moiré system is operated in an error map mode, in which a structurally illuminated reference surface is used to chromakey the image of an identical structurally illuminated alignment target. The output is a moiré image generated by the misalignment of the two surfaces; further processing indicates the degree and direction of the misalignment. The translation alignment error signals (x, z, zoom) are generated by conventional image processing operating on filtered or non-structurally illuminated images. In one design, translational and rotational alignments are handled by separate processors controlled by a single host, and might use synthesized images as reference: the off center and smaller image of a laterally misaligned and distant target can be zoomed and centered with respect to a synthetic image. In another design, the synthetic image is passed to the angular alignment program while translational alignment takes place, so that coarse angular alignment can proceed essentially in parallel with the lateral and range alignment. Thus both sequential and parallel processing can be utilized to solve the alignment problem.

During the design and manufacturing processes of specular surfaces, waviness and shape defects may occur, reducing the quality of the surface regarding its visual appearance and/or its technical performance. Typically, these defects are only a few micrometers deep and a few centimeters wide. Beside the plain defect recognition, the reconstruction of the surface topography is of great interest, because it allows to characterize defects quantitatively. To reconstruct specular surfaces, a reflection technique based on structured-lighting is used. The specular object is considered as a part of the optical system. This technique is sensitive to changes of the slope of the surface. A series of patterns produced by an illumination system and reflected at the specular surface is observed by a camera. The distortions of the patterns in the acquired images contain information on the shape of the surface. This information is recovered through a model of the imaging function of the optical system. In contrast to previous approaches, with the proposed method it is possible to reconstruct the topography of the specular surface without an iterative approximation. This is achieved by applying a smoothness constraint to the surface data and directly calculating the surface topography from the imaging function.

We demonstrate that a grazing incidence configuration can be used with a surface relief diffraction grating having a shallow frequency chirp in one dimension and straight rules in the other to produce a long stand-off profilometer with significant anamorphic magnification.

Authors have worked out a nonstationary signal analysis method on an example of research of laser lines. This method disclosed relationship between signal approximation coefficients and geometry signal characterizations (for instance, energy center, moment of inertia). Examples, which is demonstrating an application of this method for exact coordinate determination problem in laser line at displacement compensation in laser imaging are present.

The systems presented in this paper are based on phase-shift measurement. The first solution developed here is the heterodyne technique associated with direct counting. An improvement of the system consists in using digital PLL to reduce the phase noise. A second technique based on under-sampling technique, applied to digital synchronous detection, is described. Its main advantage is a global simplification of the electronic system, leading to a quite simple development of a twofold modulation system. This new technique is also very interesting to move towards a kind of smart range finder able to adapt different parameters to the different steps of the measurement.

Accordion fringe interferometry (AFI) is an active-triangulation, surface-profiling technique based on projecting interference fringes onto the surface ofan object from a source at one location and viewing with a camera at another location. Interference fringes have the advantage of infinite depth of field, which allows large or complex objects to be illuminated without the need for source focus. Conventional AFI produces the interference fringes by using a laser to form two mutually coherent source points. Coherent illumination, however, produces speckle in the image that limits range resolution. A technique that significantly improves the range resolution ofAFI is described and demonstrated experimentally. This technique uses broadband illumination to eliminate speckle effects, while maintaining the advantages of interference-fringe projection.

Traditionally, one of the driving forces behind the development of industry robots was the goal to replace tedious or hazardous manual work such as welding or varnishing. However, as robot prices are falling constantly, the use of robots becomes an economically sound solution in more and more applications – including areas which have nothing in common with the traditional “harsh environment” scenario. With the wider application scope, there also come more and tighter requirements, one of them certainly being accuracy. The accuracy of a robot can be asserted by measuring the position and orientation of its end effector. For this purpose there exist a number of techniques, one of them being photogrammetric measurement. Robot manufacturers have used photogrammetry since several years in order to carry out a factory calibration. However, there is a growing need for systems which are able to calibrate and constantly monitor a robot on the factory floor. This paper describes the possibilities which are offered by recent developments in sensors, image processing and close range photogrammetry. It then shows first results towards an on-line photogrammetric robot tracking system which have been obtained by a research group at the University of Stuttgart, Germany.

Object pose estimation is an important application in three-dimensional (3-D) object recognition. A 3-D geometrical feature descriptor, Angle Distance Map (ADM), is proposed to describe the surface feature around a local point in range images. A triangular mesh model of the 3-D object is used for reducing the computational complexity. The principal component analysis (PCA) method is applied on the ADM descriptions of surface mesh vertices. The projected vectors of an ADM description in subeigenspace are then used to match the feature points and determine the object pose. The matched feature points are refined by a proposed outlier elimination approach. The pose information is computed according to the final 3-D feature points. The proposed approach is tested in an application for flexible robot assembly. The experimental results show that accurate pose estimation can be obtained using the ADM descriptions on 3-D object surfaces.

In this paper, a new on-line measurement and accuracy analysis method for part configuration and surface is presented by combining computer vision and neural networks. Different from conventional contact measurement, it is non-contact measurement method, and it can operate on-line. In this method, the 3D configuration and surface of part are reconstructed from stereo image pair taken by computer vision system. The architecture for parallel implementation of part measurement system is developed using neural networks. Several relevant approaches including system calibration, stereo matching, and 3D reconstruction are constructed using neural networks. Instead of conventional system calibration method that needs complicated iteration calculation process, the new system calibration approach is presented using BP neural network. The 3D coordinates of part surface are obtained from 2D points on images by several BP neural networks. Based on the above architecture and the approaches, the part measurement and accuracy analysis system for intelligent manufacturing is developed by making fall use of the advantages of neural networks. The experiments and application research for this system is also presented in this paper. It is proved through the actual application that the method presented in this paper can meet the needs of on-line measurement for parts in intelligent manufacturing. It has important value especially for on-line measurement of parts that have complicated surface.

We present a digital vision system for acquiring the complete 3D model of an object from multiple views. The system uses image focus analysis to obtain a rough 3D shape of each view of an object and also the corresponding focused image or texture map. The rough 3D shape is used in a rotational stereo algorithm to obtain a more accurate measurement of 3D shape. The rotational stereo involves rotating the object by a small angle to obtain stereo images. It offers some important advantages compared to conventional stereo. A single camera is used instead of two, the stereo matching is easier as the field of view remains the same for the camera (but the object is rotated), and camera calibration is easier since a single stationary camera is used. The 3D shape and the corresponding texture map are measured for 4 views of the object at 90 degree angular intervals. These partial shapes and texture maps are integrated to obtain a complete 360 degree model of the object. The theory and algorithms underlying rotational- stereo and integration of partial 3D models are presented. The system can acquire the 3D model (which includes the 3D shape and the corresponding image texture) of a simple object within a 300mm x 300mm x 300mm volume placed about 600 mm from the camera. The complete model is displayed using a 3D graphics rendering software (Apple’s QuickDraw 3D Viewer). Both computational algorithms and experimental results on several objects are presented.

The usage of chrome or highly polished precision tooling (reference) spheres is common in the calibration and operational characterization of measurement systems such as a Coordinate Measurement Machine (CMM). The usage of a three-dimensional, (3D) laser triangulation, non-contact measurement system on CMMs and other scanning systems pose several obstacles. The highly specular mirror finish on the tooling sphere provides an accurate mechanical entity that has adverse results with laser sensors. The development of tooling spheres with a diffuse surface would benefit laser based measurement systems. The surface roughness and reflectivity properties have an effect on the laser measurements' accuracy. Efforts to develop spheres and establish meaningful measurements of spheres with modified surface finishes are investigated.

With the improving capabilities and acceptance of laser triangulation based systems, the natural extension of these systems to 3D measurement tools is gradually coming about. As with any metrology tool, calibration, including the ability to verify calibration in the field is becoming an important issue to potential users. There is currently no standard method of doing these calibrations. The methods used in Coordinate Measurement Machines, such as looking at polished spheres, were developed to account for some of the limitations of probing capabilities of the time such as touch probe lobing errors and machine axis squareness. Optical probes do not work the same as mechanical touch probes, and as such have new problems as well as capabilities. The ability to establish a calibration over an entire field at one time, using tens to hundreds of thousands of points rather than tens to a hundred points is both viable and necessary to effective 3D system calibration. This paper will discuss the basic calibration problems particular to triangulation based 3D systems that need to be addressed. The issues to be addressed include both optical aberrations such as field curvature and distortion as well as the effects of light distribution and surface reflectivity. Finally, we will describe and contrast a number of potential methods for performing effective full-field calibrations and discuss the pros and cons of each approach. Experimental results will be presented to illustrate these comparison points.

A machine vision application for the fully automatic straightening of steel bars is presented. The bars with lengths of up to 6000 mm are quite bent on exit of the rolling mill and need to be straightened prior to delivery to a customer. The shape of the steel bar is extracted and measured by two video resolution cameras which are calibrated in position and viewing angle relative to a coordinate system located in the center of the roller table. Its contour is tracked and located with a dynamic programming method utilizing several constraints to make the algorithm as robust as possible. 3D camera calibration allows the transformation of image coordinates to real-world coordinates. After smoothing and spline fitting the curvature of the bar is computed. A deformation model of the effect of force applied to the steel allows the system to generate press commands which state where and with what specific pressure the bar has to be processed. The model can be used to predict the straightening of the bar over some consecutive pressing events helping to optimize the operation. The process of measurement and pressing is repeated until the straightness of the bar reaches a predefined limit.

Volume measurement is an important process for various industries such as food processing, fruit and vegetable grading, etc. Value or price is often determined by the size of product. In seafood industry, for example, oyster meat is separated into four grades before being packaged. Large size grade means higher selling price than small size. More consistent packaging size is also an indication of high quality. Product size can be measured optically with machine vision technology for on-line inspection and grading systems. Most optical grading techniques use a two-dimensional area projection or the weight of the product to estimate the actual product volume. These methods are subject to measurement inaccuracy because of the missing thickness information. An algorithm combines laser triangulation technique with two-dimensional measurement to reconstruct a three-dimensional surface for volume measurement is introduced in this paper. The result of this technique shows a significant accuracy improvement from the area-projection method.

Progress in imaging sensors and computers create the background for numerous 3D imaging application for wide variety of manufacturing activity. Many demands for automated precise measurements are in wood branch of industry. One of them is the accurate volume definition for cut trees carried on the truck. The key point for volume estimation is determination of the front area of the cut tree package. To eliminate slow and inaccurate manual measurements being now in practice the experimental system for automated non-contact wood measurements is developed. The system includes two non-metric CCD video cameras, PC as central processing unit, frame grabbers and original software for image processing and 3D measurements. The proposed method of measurement is based on capturing the stereo pair of front of trees package and performing the image orthotranformation into the front plane. This technique allows to process transformed image for circle shapes recognition and calculating their area. The metric characteristics of the system are provided by special camera calibration procedure. The paper presents the developed method of 3D measurements, describes the hardware used for image acquisition and the software realized the developed algorithms, gives the productivity and precision characteristics of the system.

The paper presents a new method of measuring 3D small angles. In the method, a collimated parallel ray is projected on a double-face-reflector, a CCD detects the position changes of the returned light points and by them, the paper sets up the measurement model of 3D small angles. It describes the characteristics of this method and gives its resolving method. The simulation tests verify their correctness. Finally, the paper does some analyses of the effect of some system parameters on the measurement..

Hardware and software advancement in computer technology enable the development of faster and better application of laser inspection technology in the field of pipe surface corrosion inspection as well as weld profile inspection. Several software programs for inspecting surface defects on pipelines have been developed by EWI within the past few years.

This paper describes a new methodology verifying the accuracy of photogrammetric networks by known distances between imaged points on scale bars. The analysis presents three issues of photogrammetric accuracy (point, distance, and scale accuracy) together with the issue of scale-bar length accuracy. The latter, in turn, involves the uncertainty in reference length, measured temperature, coefficient of thermal expansion, and centering of targets. The methodology is based on a one-time determination of photogrammetric point accuracy using statistics generated by discrepancies between points in a photogrammetric network and their counterparts determined very reliably by other means. It focuses on geometrically strong networks, where the coordinate standard error is approximately the same for all three coordinates of a typical point. This methodology replaces much lengthier and more expensive numerical procedures, such as a Monte-Carlo technique, by a straightforward statistical analysis. Under stipulated conditions, it can assist users of digital photogrammetry in verifying the performance of their system(s) after each survey.

The analysis of surface properties represents one of the most challenging and rapidly developing applications for machine vision today. Numerous manufacturing and processing tasks involve the control of surface attributes, such as three-dimensional shape, surface topographic texture, and two-dimensional coloured patterns.

A look at the challenges surrounding the modification of a large field of view system (6" - 18") for small field of view (~0.5") inspection in the electronics industry. Camera-based 3D systems suffer serious losses in depth of field when attempting to look at small fields of view. Off axis lens placement and careful application of the Scheimpflug Rule can significantly improve any camera-based inspection system, but these improvements may require modifications in your camera calibration procedures. A manufacturing inspection application and the improvements brought about by these techniques are discussed along with the pitfalls that were encountered along the way

The technical characteristics, advantages and applications of an automated optoelectronic measuring systems for 3-dimensional inspection of complex objects (gas-turbine engine (GTE) elements, biomedical objects, etc) designed by "Optel" company, State Aviation Technical University of Ufa, are presented in this paper. The measuring apparatus can be applied in research and in industry. Its main advantages are non-contact non-destructive accurate inspection and rapid 3D measurements.

Our purpose is, in medium term, to detect in air images, characteristic shapes and objects such as airports, industrial plants, planes, tanks, trucks, … with great accuracy and low rate of mistakes. However, we also want to value whether the link between neural networks and multi-agents systems is relevant and effective. If it appears to be really effective, we hope to use this kind of technology in other fields. That would be an easy and convenient way to depict and to use the agents' knowledge which is distributed and fragmented. After a first phase of preliminary tests to know if agents are able to give relevant information to a neural network, we verify that only a few agents running on an image are enough to inform the network and let it generalize the agents' distributed and fragmented knowledge. In a second phase, we developed a distributed architecture allowing several multi-agents systems running at the same time on different computers with different images. All those agents send information to a "multi neural networks system" whose job is to identify the shapes detected by the agents. The name we gave to our project is Jarod.