This paper describes a black and white photographic chart called the 'Scanner Analyzer Target' developed as a practical aid toward analyzing scanned image quality produced by binary digital document scanners used as input devices in document image processing systems. The target is used by factory-line personnel as an 'adjustment-set' tool, by the quality control department as an image quality judgement tool, and by customers as an incoming inspection tool. The scanner uses a rotary drum paper transport to move the document in front of a fixed scan line. The document is illuminated by a pair of fluorescent lamps and reflected through a reduction type camera to a CCD image sensor. Scanning speeds range from 7 in./s to 12 in./s and scanning resolutions range from 200 to 300 dpi. The scanner control panel allows image quality adjustments by using 9 levels of darkness and 10 levels of sensitivity. The darkness setting controls the level of optical reflectance at which the scanner shifts its output from white to black. At low darkness setting a scanned image appears lighter than the original and at high darkness setting the resulting image is dark. The sensitivity setting controls the amount of adaptive thresholding included in the binary decision. Low sensitivity setting is adequate for high contrast originals; high sensitivity setting allows the scanner to detect low contrast information contained in different background shades (or colors). At maximum sensitivity, the scanner is extremely sensitive to high frequency (text) and tends to drop out the background information. The purpose of the target is to give the user the ability to measure a number of scanned image parameters: text legibility, resolution, focus, darkness range, shading distortion, sensitivity range, photograph dithering, vertical and horizontal pixel count, document or CCD skew, vertical resolution linearity, aspect ratio, video noise and page registration.

As the desk-top publishing market has been rapidly expanding, high-quality image output technologies have been increasingly required; especially as to restoration for images with gradation like photographs. To meet this requirement, the authors developed a new high quality image recorder and tried some experiments of simulation image recording. In the high quality image recorder, the laser scanning optical system is used, including an anamorphic optical system. The laser spot diameter in the deflection plane is small as 30 micrometers in the image plane, and the scan length is large as 300 mm. A polygonal mirror is used as the beam deflector and the scanning optical system corrects the tilt error of the polygon facets. By using this image recorder, the authors obtained simulation image samples, which is based on the bi- level pixel of 1200 dpi density. The simulation image samples mean halftone images with some screen rulings. The simulation procedure and the result of the examination are reported.

Image quality of scanned documents is examined in the context of a data entry department reading data from the images with intelligent character recognition (ICR) equipment. First, the equipment and its use are reviewed. An economic model of the cost benefit of the ICR equipment is developed, and the cost benefit of a specific application is examined. To complete the analysis, a test deck of documents is used to measure the specific parameters that relate to image quality and the performance of the equipment. These results are analyzed to show the impact of image quality on operating costs. A novel image processing algorithm which dramatically improves the quality of one bit per pixel images is shown to work with very low contrast documents. It also works on documents with noisy and/or dark backgrounds. A comparison of the enhanced images to those produced by off the shelf scanners is then made with several test documents. A test is made of the ICR system with the new algorithm in place to produce high quality images. The impact of the use of image enhancement is then shown to increase significantly the cost benefit of using ICR equipment. This test substantiates the relationship of dollars to image quality of scanned documents.

The image modulation and noise characteristics of monochrome cathode ray tube (CRT) displays are described using a simple physical model. The model includes digital-to-analog converter, the video amplifier, electron gun, phosphor and faceplate. Expressions are derived for both the signal modulation transfer function (MTF) and noise power spectrum in the terms of the subsystem characteristics.

The literature presents many examples of visual distress (Gunnarsson and Soderberg, 1983) and (Ostberg, 1975) which indicate that many VDT operators suffer from ocular discomfort. Task design, room lighting, VDT screen glare, improper refraction correction, and the physical workstation environment are often given as the reasons for observed visual distress without delving into the specific mechanisms or combinations thereof of visually induced fatigue. There has been little noticeable improvement made up to today even though many more VDTs are in use.

The adoption of document image processing is resulting in the CRT replacing paper as the primary medium for written information in the office environment. The anticipated rapid adoption of this technology will mean that more workers will be using CRTs for larger portions of the workday. The U.S. Department of Commerce estimates that the percentage of workers using CRTs will increase from 15% in 1989 to 50% in 1997. Increased CRT usage has brought with it the problem of increased eyestrain resulting in sub-optimal productivity and worker discomfort. It is well documented, but not widely known, that reading from commonly used 60 Hz 24-line CRTs is 30% slower and is the cause of increased eye fatigue when compared to reading from hard-copy paper. The culprit is poor CRT image quality; and it dramatically affects users who spend three or more hours a day using CRT screens. In order for document image processing to be accepted by users, computer displays must be much closer to the display quality of paper, which commonly used displays on PCs today do not approach. The minimum requirements for displays used for document image processing are the ability to display an entire 81/2 X 11 page and have the information on the page be easily readable without manipulation of the image by zooming or scrolling. In other words, the image must be as easy to read as paper. Displays that meet these criteria are called 'paper-like displays.'

Digital images of business and financial documents are replacing paper and microfilm in numerous applications. System bandwidth limitations usually necessitate data-compression of such images. Another way to decrease the bandwidth requirement for each image is to reduce the spatial resolution of the image, thereby reducing the total number of pixels which must be stored or transmitted. In this paper, the authors investigate the combination of spatial resolution reduction with the evolving ISO/JPEG grayscale image-compression standard. Both resolution reduction and ISO/JPEG data-compression cause distortion of the decompressed image with respect to the original. The system designer has some flexibility in allocating this distortion to each of the processes. The authors have conducted extensive testing to determine the best such allocation. Results of these tests indicate that legibility depends principally on compressed size. In particular, legibility is directly proportional to compressed size. The effects of spatial resolution on legibility are secondary. One such effect is a decrease in legibility with increasing spatial resolution as packet size is held constant at comparatively small values. Taken together, these results argue that the best legibility for document images compressed with the ISO/JPEG draft standard is obtained at the lowest adequate spatial resolution.

This is a presentation of commercially available galvanometric and resonant oscillatory scanners. It is intended to assist in the selection of component, single and multi-axis scanning arrangements. The features of galvanometric scanners are reviewed and figures of merit offered to rate electromagnetic drivers. Position transducers are examined and three capacitive sensors compared. Resonant scanners are analyzed with a presentation of the milestones to be resolved by commercial product manufacturers. Multi-axis scanning arrangements are examined and the distortions of four configurations reviewed. The last section has 11 tables grouping comparable commercial products. A list of manufacturers is offered.

Optical scanner specifications are subject to a wealth of nonstandardization and scanner terminology to considerable misuse. The need for a standardization of terms and definitions is readily apparent to anyone involved in specifying or furnishing this technology. There is considerable debate as to the correct method of test or to the appropriate terminology for a given phenomenon. This paper proposes standard definitions and highlights test methods that appear either most common in industry or most correct. This is an attempt to stimulate discussion on standardization of terms and tests, the first step to a definitive standard. A glossary of terms is provided with brief definitions and justification, where required. Several of these terms, which require more extensive treatment to fully understand their intricacies and significance, are reviewed in a discussion section that follows. The included examination of test methods brings out subtleties in parameter definitions and the clear need for, as well as the difficulty in, standardization.

The proper design of a control system for use in high demand scanning applications requires the understanding of how the needs of the application dictate the configuration of the system. Once the needs of the application have been defined, those needs can be translated into the required system performance parameters, and subsequently a review to the limits of the performance can commence. This paper discusses the needs of two common applications, high accuracy stepping systems and high speed vector systems. Then, mechanical models of a galvanometer are examined and common limits to performance for each application is discussed. Finally, some approaches to servo amplifier designs which extend performance are reviewed.

A two-axis, high-bandwidth, small-aperture steering mirror called the High Bandwidth Steering Mirror (HBSM) has been designed, fabricated, and tested. The mirror/mechanism prototype functions within a servo loop either scanning a field of view or tracking a radiation source. The design focused on elements making up the beam-steering mechanism: mirror, restoring flexures, actuators, position sensors, and encompassing housing, and the part each component plays in making a mechanical system suitable for high-bandwidth operation. Inclusion of a novel flexural support allows one-degree peak-to-peak angular stroke (shaft space) at low frequencies and a small-signal closed-loop bandwidth of up to 10 kHz without the usual mechanical resonance-induced loop instabilities. This increased bandwidth allows substantial rejection of a disturbance spectrum in the 10-1000 Hz range and execution of fast, complex scan patterns. Pointing accuracies of 0.2 micronrad have been achieved in the laboratory. Details of the mechanical design and fabrication issues as well as the control-loop implementation are discussed. Test data are presented along with reports of the mirror's performance in use as an extended sensor.

Attention is given to a family of single-mirror two-axis beam steering systems (TABS) developed for fine pointing control of an optical beam. The requirement is to control the line of sight of the optical system on board such a platform with sufficient bandwidth to overcome the perturbations and thus maintain the laser communication link between two satellites. The armature is supported on its center of gravity on a gimble. A balanced design is achieved with the mirror and tachometer magnets on one side of the gimbal and the torque magnets on the other. A negative magnetic spring constant which compensates for the angular restraining effects of the flexure bearings without any radial stiffness consequences is incorporated.

Linear high-speed opto-mechanical scanning at TV and sub-fractional TV line rates is commonly achieved today with multi-faceted rotating polygons. An alternate near-linear scanning means is now available: a linear resonant scanner. The linear resonant scanner is a dual mode torsionally resonant system designed to simultaneously resonate at a fundamental and a third harmonic of the fundamental. An electronic circuit with a precision position sensor is used to control the scan amplitude and phase lock the two natural resonant frequencies. Design criteria and performance parameters are discussed for a 2 kHz linear resonant scanner with a 0.5 radian near linear scan angle, 88% bidirectional scan efficiency, and 18 mm pupil.

Many mechanical systems use high accuracy position feedback in conjunction with servo electronics and mechanical actuators to obtain precision positioning of mechanical components. The accuracy of the position detector limits the system performance. Compensations for errors which have a repeatable nature are possible. Random errors must be made small compared to the expected system resolution. There are many well known technologies which can be used to detect the position of a moving element. This article examines those rotary applications where the servo-positioner must be capable of high accelerations. In such systems the moving parts of the sensor must have a low inertia compared to the total moving inertia. A low inertia sensor will allow for efficient use of the torque produced by the actuator. A sensor with a high inertia can cause the mechanical structure to have resonances which will limit the stable bandwidth of the servo-system. In these situations capacitive position sensing emerges as a practical sensing technique. Transducers with inertia ranging from .005 to 1.0 gm-cm2, and having useful angles of +/-45 degrees motion, can be made for systems with up to 20 kHz of bandwidth. The resolution will be lowest for the low inertia versions, typically around 100 (mu) radian. Capacitive sensors with an inertia of 1.0 gm-cm2 and an accuracy 0.5 (mu) radian are possible.

The authors describe an interferometric angle encoder for a high-speed galvanometer having the following characteristics: (1) zero added mechanical moment of inertia, (2) microradian resolution, (3) laser diode light source, and (4) small package. The design uses the rear face of the existing galvanometer scan mirror to scan a laser diode beam across a cylindrical scale (a grating width on the order of 1000 lines/inch) whose axis coincides with the galvo axis of rotation. The +/- 1 diffracted orders are collected after being again reflected by the galvo mirror and are made to interfere. The phase of the resulting interference fringe pattern is measured with a resolution of approximately equals 1/400-fringe (or 1/800-line spacing on the grating) to achieve an angular resolution of 1.3 microradian (0.2 arcsec). The measurement update frequency is nominally 4 MHz, corresponding to a maximum angular velocity for full on-the- fly resolution of 5 rad/sec (fringes can be accurately tracked at angular velocities of greater than 1000 rad/sec). The design concept is described in detail and test data are presented.

An interferometer consisting of a plane-parallel refractive plate rotatable with respect to a fixed flat mirror measures angle by the passage of fringes. It is an amplitude division two- beam interferometer like the one discovered by David Brewster in 1815. The folded Brewster interferometer is insensitive to errors of manufacture and assembly. It responds to movements in just one degree of freedom while being insensitive to movements in the other five. The dependence of retardation upon input angle is analyzed as a function of the design parameters. There is a refractive index which minimizes the moment of inertia of the moving element. Two ways to deploy sources and photosensors for angle measurement are discussed. A prototype folded Brewster interferometer built for a galvanometric scanner had a refractive plate measuring 3mm thick X 6 mm wide X 19 mm long. At (lambda) equals 633 nm one fringe corresponded to 128 (mu) radians, and an angular range of +/- .15 radian could be covered.

Vector scanning of laser beams is critical to the success of stereolithography, laser marking, CAD-output-to-microfilm and many other industrial applications. The design of a vector scan lens system can assume many different configurations. This paper will discuss pre-objective, and post-objective alternatives for two and three axis scanning. Various parameters required for system specification are reviewed and the basic configuration of the two axis scan head is presented. The pre-objective scan f-theta lens, its scan distortion, telecentric options, and single origin scan systems are described. Fixed focus and dynamic focus post-objective systems are presented discussing the relative merits of each approach and noting the limitations of each design solution. Data presented describe limiting field sizes for fixed focus systems, and limiting f-numbers for thin lens systems. Thin lens design equations for three post-objective configurations are presented with suggested optimization techniques for overall system layout. Examples of pre-objective and post-objective scan lens systems as well as pre- configured systems are presented.

General Scanning has developed its second generation of two-axis, continuous tone raster film recorder for medical imaging. This print engine utilizes a new scanning geometry, as well as new scanning and exposure control components that provide better image quality at higher speed and lower complexity and cost. A prototype has been developed and evaluated, meeting the demanding specifications of the medical market.

This paper investigates one of the possible designs of an optical-electronic system for identification of optical measuring signal parameters as a two-dimensional interference picture with spatial selection.

The need for extremely precise positioning of pre-objective galvonometer based laser pointing systems necessitates calibration to correct for various system distortions. Such distortion errors fall into two categories: those inherent to the geometry of the scanning system layout, including galvonometer mounting configuration and lens specification; and those related to the behavior of the galvonometer as a precise positioning mechanism. This paper examines system error sources in pre-objective scanning systems and proposes several techniques to correct them.

In the past few years, a new technology for producing fast turnaround prototype models utilizing state-of-the-art technology has emerged. This technology, known as rapid modeling or desktop manufacturing, has capabilities that were previously unavailable for prototypes. In essence, rapid modeling breaks down a CAD solid or surface model into many descriptive cross-sections, which are then re-created by machine and adhered to one another to form an almost infinite set of shapes as a real three-dimensional part. This process employs leading edge technology in scanning, computer hardware/software, and lasers.

For several years, the Naval Ocean Systems Center (NOSC) has been working on the development of laser-based display systems with the goal of upgrading the image quality and ruggedness of shipboard displays. In this paper the authors report work on the major task of developing a full-color laser-addressed liquid crystal light valve (LCLV) projection system.

A scene generation capability is under development at the Arnold Engineering Development Center (AEDC) for visible and lR focal plane array (FPA) testing that uses two-axis acousto-optic (AO) deflectors. A multifrequency RF input creates a rake of output beams which is step-scanned across the face of the imaging device. Each component of the rake is modulated independently, and the beams are blanked during the shift from one row of pixels to the next. The expected maximum frame rate is 20 kHz, with operation synchronized to the operation of the FPA. A modular concept is being investigated to address large (51 2 x 512 pixel) FPAs. This AEDC Direct Write Scene Generation (DW5G) technology is being applied toward development of a transportable and a fixed-site Scene Generation Test Capability (SGTC).

There are some limiting factors in a transmission film copy mostly caused by mechanically moving parts. The subject of this paper is a new design of the electro-optic modulator for temporal scanning of light with both spectral and intensity modulations. A cube-formed module with a pair of electro-optic phase retarders processes not only the amplitude of an electromagnetic wave but also its polarization state. Thus an electro-optic modulator operating at Mc temporal sampling of (a) color range (blue, green, red), and (b) that enables intensity modulation of each color in 73 levels used in a transmission film copy is to be developed. A great number of characteristics must be fulfilled by electro-optic crystals for this particular application - namely, (a) broad spectral transmittance over the whole visible range, (b) reasonably high value of electro-optic constant, (c) short relaxation time, (d) low tangent loss, and (e) low deformation coefficients. Analysis of some materials for electro-optic modulators is stated in a sequence to the final concept with crystals performing double Pockels effect (DPE) and reaching the criteria mentioned above, particularly crystals of 23 and 43m groups of symmetry.

A one-crystal 2-D scanner using the circular birefringence of tellurium dioxide was designed, built, and characterized. A scan angle of approximately 4 to 5 degrees was obtained along each axis for an acoustic frequency range of about 30 to 80 MHz. The intensities of the higher order beams were in the range of 10-2 to 10-4 relative to the first order intensity. The intensity of the second order beam was studied as a function of transducer length and frequency. The authors found a rough agreement based on (1) the diffraction spread of the acoustic beam, and (2) the slope of (theta) i vs frequency, d(theta) i/df, which is a measure of the tolerance of the input beam angle phase matching conditions to small angular deviations.

Optical beam distortion generated by an acousto-optic sound cell can be characterized in terms of a linear system methodology. Comparison of various first order diffraction profiles numerically obtained for a Gaussian input field demonstrates an inherent frequency-dependent asymmetry.

The mercurous halides, Hg2CI2, Hg2Br2, and Hg2l2 are materials with highly anisotropic and unusual properties. They are transparent from the blue through the far infrared and they have high refractive indices and large birefringence. Most important, they support shear acoustic wave propagation along the crystallographic <1 1 0> directions whose velocity is extraordinarily low, slower even than the speed of sound in air. Because of this slow velocity Bragg cells fabricated from the mercurous halides, offer very long time delay signal processing in a compact package. What is more, the other properties combine to produce a very high acousto optic figure of merit which in turn implies very high efficiency at low drive power. We have fabricated cells from the mercurous halides with delays up to 180 micro seconds which require only a few hundred milliwatts drive1. Key to exploiting the mercurous halides for high performance Bragg cells is the availability of single crystals with the correct orientation and with low optical and acoustic loss. Considerable effort on crystal growth technology has produced remarkable i mprovem ent in these properties2.3. Crystals have been grown which permit fabrication of Bragg cells approaching 200jis delay time with reasonable optical quality. In this application, the extremely low acoustic velocity allows construction of a long delay device in less than 1/2 the volume required by other materials. Thus, the shorter acoustic path leads to a small optical aperture which, in turn, translates to a compact and practical signal processing element. Thus, development of Bragg cells of the mercurous halides will lead to compact, higher performance acousto-optic processors .

This paper describes an optical correlator for use in sonar ranging. The sonar system must process low probability of intercept signals while consuming minimum power. An optical correlator is favored because of the high efficiency of optical components and the low duty cycle, afforded by the high speed of the optical processing. The correlator is of the time integrating class because this approach permits the processing of extremely large time- bandwidth-product signals ($OM106).

Cw mode-locking of a Nd:Glass Laser at 0.5 GHz repetition rate has been achieved. The average power is 20 mW and the pulse width is 9.2 ps.

Single and multiple line imaging systems are in general use in the field ofphoton imaging. CRTs, LED arrays, and lasers have formed the base technologies from which specific designs have evolved. In certain applications, such as the exposure of high quality color films, extremely tight tolerances are placed upon intensity uniformity, addressing accuracy, energy density, wave length, and resolution of the source. These demands must often be achieved within a configuration that is low priced, reliable, rugged, and small. Several years ago, the author and his associates were asked to consider the possibility of configuring an electron beam activated phosphor screen device, in which the phosphor plane and the electron beam were paraxial ( parallel axes). It was reasoned that such a structure would offer a substantial advantage in size, performance and cost that inhibited many applications for electron beam systems. The result of this activity is the Paraxial Electron Imaging System (see fig. 1).

The design of rotary mirror scanners is a complex process involving many trade-offs. This paper explores these issues emphasizing how the design can be optimized for a particular application. Examples range from thermal imagers to prepress equipment, and the trade-offs covered include scan accuracy vs cost, efficiency vs performance. The topics covered include polygons, motors, bearings and housings. The configuration of these parts and their interactions are considered. Particular emphasis is placed on reducing the power consumption of the scanner through careful consideration of windage - the drag on the polygon due to its movement through the surrounding gas. The main conclusion is that by careful design the effect of trade-offs can be minimized; this allows the design of low cost high accuracy scanners.

The study investigates the relationship between the change in the surface of the mirror and jitter when one of the causes of jitter is due to deformation of the polygonal mirror surface by the centrifugal force created by high-speed spinning of the polygonal mirror. Finite element analysis is used to obtain the deformation pattern of the polygonal mirror by centrifugal force under various constraints governing the attachment of the polygonal mirror to the spindle motor. It is found that the amount of the deformation of mirrors and the deformation patterns greatly depend on the bolting diameter and the number of bolting positions; in particular, when the uniform constraint is applied, there is an R value which describes the minimum deformation.

A new optical scanner is described which serves as a monogon, or single-facet device, providing one scan per shaft rotation. It cancels cross-scan line placement errors automatically, yielding scan lines which are spaced precisely, independent of drive shaft wobble. This scanner is configured for simple fabrication, of low mass and size, allowing convenient dynamic balance for high-speed operation. This new scanner is identified as an open-mirror monogon.

This paper analyzes the jitter components due to the mirror in a polygon laser scanner. Before the start of analysis, the following hypothesis was adopted. There are two factors in the condition of mirror facets which affect the high-frequency components of jitter; one is the 'radius variation of the mirror' and the other is the 'curve of the mirror facet.' These two factors can possibly be superimposed. Based on this hypothesis, the following experiments were conducted. The 'radius variation of the mirror' and the 'curve of the mirror facet' of a polygon mirror were independently measured for each mirror facet, and the data were converted to jitter values and then added. Also, a general-purpose jitter measuring optical system was constructed, actual jitter was measured, and the high-frequency components were analyzed. From these experimental results, the jitter values for each mirror facet agreed well with synthetic value. It was concluded that the above-mentioned hypothesis is correct. Here, the 'radius variation of the mirror' represents the dynamic indentation and protrusion during the rotation of the polygon scanner. Therefore, it is the overall 'radius variation of the mirror' which includes the variation produced by the deviation of the mirror which is attached to the spindle motor and axial vibration due to rotation in addition to the polygon mirror's static indentation and protrusion. In the present study, in order to measure the 'radius variation of the mirror,' a new 'radius variation measuring device' has been developed. Also, generally, laser printers contain imaging optical systems such as an F(theta) lens, and the 'radius variation of the mirror' is not directly related to jittering. However, the 'radius variation of the mirror' is one of the important parameters of a polygon scanner. By the present analysis and measurement method, one of the factors of jitter, 'the radius variation of the mirror,' has been clearly evaluated.

The characteristics of a rotating polygon scanner in relation to its flat field lens are expressed as a function of relatively fixed system parameters. They include the desired resolution, number of facets, scan duty cycle, and desired pupil relief distance (to the flat field lens or to the first physical obstruction in the optical path). Account is taken of beam enlargement on the facet due to non-normal landing, and of provision to avoid excessive beam truncation at the ends of scan. Also, the location of the polygon with respect to the optical axis is determined from system parameters. Equations are provided which govern the polygon dimensions and its relationship to associated components, and application procedure is summarized.

This paper briefly reviews the history of the development of holographic deflectors for graphic arts applications with emphasis on the current status of the technology and products which serve this market. In particular, a number of commercial products and the hologon used in them are described.

This paper describes the research and development efforts undertaken to develop holographic deflectors that could achieve rotation speeds of 28,000 RPM or higher. Major areas of investigation undertaken to achieve proposed operating conditions are the evaluation of both ac and dc spinner motors and the accompanying drive circuitry, motor bearing life requirements, and hologon structural considerations.

Excellent performance of hologon deflectors has resulted in the design and manufacture of high-resolution laser scanning systems which utilize simple spherical optics, and have very high output accuracy. However, for flat-field systems an f-theta scan lens is needed, and cost- effective production designs force the f-theta distortion figures to 0.05% and above. Hence, for higher accuracy requirements, the residual spot placement errors must be corrected differently. In this paper, the authors describe the work undertaken to compensate for the f- theta distortion of a scan lens in a holographic deflector system used for master artwork generation. This system employs a monofacet hologon operating at the green HeNe wavelength (544 nm), produces 200 scans per second, and is capable to address >4000 dots per inch. The authors also present the results of the compensation technique which employs pixel position modulation, and reduces the 0.1% residual scan lens f-theta distortion to a level below 0.003%, as required for a <15 micrometers dot placement error at any point along the 470 mm scan field.

In the present paper, a system for determining longitudinal displacement of an object is described. In this system the laser radiation is modulated by a rotating diffuser and it creates a dynamic specklefield whose intensity is analyzed by a low-frequency spectrum analyzer. A change in the geometrical parameters of the system due to varying the distance between the radiating head fixed on the object and the receiving unit with a diffuser changes the intensity fluctuation spectrum, which permits determining the displacement. The system is simple in design.

A method is described for acquiring remote images from a reciprocating axisymmetric aircraft seal ring. The method relies on orbital scanning technique which allows a CCD camera and fiber optic probe to be positioned on azimuth plane within the 2(pi) radian of the inner circumference of the seal ring. The orbital motion which can be controlled by a computer allows the wider coverage of the seal ring surface but in incremental sector bases. The motion control which is based on a stepper motor and a CY500 controller is described. Position accuracy of the probe on the azimuth plane was analyzed. Errors were identified based on pure rotation of the probe, and also for rotation of the probe and the seal ring. Some experimental results obtained with an orbital motion of the fiber optic probe show that image acquisition is possible with the technique.

The purpose of this paper is to illustrate to systems design engineers how to determine a first- order approximation of the manufacturing specifications for a high performance polygonal scanner sub-system. The polygonal scanner, motor, encoder, and controller are treated as elements of the sub-system, thereby enabling specification tradeoffs to be made at the engineering and manufacturing levels. A number of basic scanner configurations are reviewed, and followed with an introduction to a film recorder scanning system. The sub-system's performance requirements are established, from which the scanner tolerances are determined. Furthermore, the manufacturing and engineering considerations are depicted in terms of the in-house disciplines and facilities that are needed to engineer and manufacture a high performance polygonal scanner sub-system.

The extensive review by Shepherd (1991) of bearings for rotary scanners is summarized. Consideration is given to bearing choice, spin axis error, speed, cost, power, life, contamination, vibration and shock, and atmosphere. Alternative bearing types are examined. It is shown that the bearing selection process is an integral part of the scanner design process. The choice of bearings has a dramatic effect on the cost and performance of the scanner.

GEC Ferranti have been manufacturing rotating instruments such as inertial quality gyroscopes and rotating polygon scanners for many years. Markets for scanning devices vary widely and include cameras, thermal imaging, laser printing and projection. The advent of high resolution laser video projection brought about new challenges in scanning technology. Initially, these challenges could only be met with complex and expensive devices and ancillary equipment. Based on our experience and success in the markets already mentioned, we have developed a range of compact line scanning devices which can be manufactured on a production basis, and which meet the requirements of scan rate, accuracy and reliability demanded by HDTV systems.

Hydrodynamic bearings utilizing ferrofluids are a new class of high performance bearings. These bearings are self-contained with ferrofluid acting both as a hydrodynamic pressure film and a sealant. Although relatively unknown in the laser scanner industry, ferrofluids have been widely used over the last two decades for applications in the semiconductor, computer and audio marketplaces. In this paper, performance features of Ferrofluid Film Bearings are also discussed, and experimental data such as power consumption, rotational accuracy and audible noise are presented to show that Ferrofluid Film Bearings enhance the performance of high resolution scanners and in this application are superior to ball bearing and gas bearing performance.

Instrument ball bearings are used in pairs, loaded against each other, mounted on a shaft, and contained in a housing. Their most important function is to define a spin axis. We develop a vector manufacturing error characterization for each bearing component. We calculate the combined misalignment of a loaded bearing pair, coupled by shaft/housing elasticity, to give a time-dependent resultant displacement/direction of the spin axis with respect to a spin reference axis. Effects of initial placement, contact angle mismatch, and total rotation angle are considered. We suggest that selective assembly with error cancellation is possible for some applications.

Random jitter of the rotational axis in scanners produces scan-line tracking errors that are unacceptable in the laser typesetting and printing industry. The design of the Butterfly scanning device nullifies scan-line cross-scan tracking errors stemming, in part, from bearing wobble. Terms and expressions associated with types of line scanning systems and reflective rotary scanning devices are defined. The essential features of an Axe-Blade scanner, a two- facet scanning device, and the principles of a Butterfly scanner, a two piece, four-facet scanning device, are described. The symmetry of the Butterfly scanner simplifies its design and construction, its alignment and dynamic balancing; and leads to a compact unit. The design uses commercially available bearings, and achieves scanned beam accuracies in the arcsecond and the subarcsecond ranges that, in general, require gas bearings.

An acousto-optic system for projection of high resolution color video is proposed. The proposed system consists of many low resolution monochrome projectors operating in parallel. The final image is a matrix of smaller pictures. A method for maintaining precise registration of color and continuity across adjacent portions of the complete picture is described.