Multiclass classification is an important problem in pattern recognition. Various classification methods have been proposed in the past few decades. However, most of these classification methods neglect the errors or the noises that exist in samples. As a result, classification accuracy is badly influenced by the errors or noises. In this paper, we propose a joint sparsity matrix learning method, which exploits l2,1-norm minimization to perform multiclass classification. In order to overcome the influence of the errors or noises, we introduce a sparse matrix to explicitly model the errors or noises and apply an iterative procedure to solve the l2,1-norm regularized problem. We perform experiments on four face databases to verify the effectiveness of the proposed method.
The generation of a variable diameter annular flat-top laser beam in the far field based on an adaptive weight FFT-based iterative algorithm (AWFFT-IA) and a phase only liquid crystal spatial light modulators (poLCSLM) is demonstrated. The iterative algorithm is used to design the needed phase distribution written on the poLCSLM for the target diameter annular flat-top laser beam shaping. The experimental results show that the method proposed can concentrates about 71% of the incident laser energy into the desired region and the root mean square error (RMSE) of the tailored flat-top intensity profile is more or less 12%.
An adaptive-weight fast-Fourier-transform-based iterative algorithm is proposed for far-field flat-top beam shaping. This algorithm inherits the projection optimization idea from the Gerchberg-Saxton algorithm, but the far-field amplitude for inverse fast Fourier transform is adaptively modified by using a novel optimized adaptive-weight strategy. First, the application of this method to square flat-top intensity-profile beam shaping is discussed as an example. The pure-phase distribution simulated by 100 iterations of this method concentrates 93.89% of the incident laser energy into the desired region and the root mean square error (RMSE) of the tailored flat-top intensity profile is 0.0094. Less than 20 iterations of this method concentrate more than 90% of the incident laser energy into the desired region and the RMSE of the tailored flat-top intensity profile is under 0.05. Then, the applicability of the method to designing the phase distributions for variable-shape or variable-diameter flat-top beam shaping is demonstrated.
Multi-beam technology is one of the key technologies in optical phased array systems for multi-object treatment and
multi-task operation. A multi-beam forming and steering method was proposed. This method uses isosceles triangle
multilevel phase grating (ITMPG) to form multiple beams simultaneously. Phase profile of the grating is a quantized
isosceles triangle with stairs. By changing the phase difference corresponding to the triangle height, multiple beams can
be steered symmetrically. It took 34 ms to calculate a set of parameters for one ITMPG, namely one steering. A liquid
crystal spatial light modulator was used for the experiment, which formed 6 gratings. The distortion of which had been
compensated with the accuracy of 0.0408 λ. Each grating included 16 phase elements with the same period. Steering
angle corresponded to the triangle height, which is the phase difference. Relative diffraction efficiency for multiple
beams was greater than 81%, intensity nonuniformity was less than 0.134, and the deflection resolution was 2.263 mrad.
Experimental results demonstrate that the proposed method can be used to form and steer symmetrical multiple beams
simultaneously with the same intensity and high diffraction efficiency in the far field, the deflection resolution is related
to the reciprocal of grating period.
The discreteness of driving electrodes in liquid crystal optical phased array (LCOPA) device causes phase valley in the
region between two adjacent electrodes. When a one-dimensional transmission-type LCOPA device was used, there was
a pair of diffraction sidelobes with considerable intensity on either side of the deflected beam, which decreased the
diffraction efficiency of deflected beam greatly. An analytical numerical model of phase valley was proposed for the
quantitative analytic analysis on diffraction efficiency. Results showed that phase valley is the cause for diffraction
sidelobes. The diffraction efficiency of deflected beam decreases steeply as the phase valley depth increases. When
valley depth decreases close to zero, the main factor having affect on diffraction efficiency turns to be flyback region
size. The influence of electrode space on diffraction efficiency was quantitatively analyzed as well. An effective way to
reduce or even eliminate the phase valley is to reduce the space between electrodes.
The structure of a storage capacitor in nematic liquid crystal optical phased array causes the data refresh rate to be higher
than the liquid crystal switching frequency, which demands a high data bandwidth. A novel driving method was
therefore proposed to reduce the data refresh rate. Without using a storage capacitor, the proposed method uses digital
scanning instead of conventional analog scanning. Furthermore, digital scanning is able to drive all pixels in parallel and
transmit driver data only once for one frame. 1-D nematic liquid crystal optical phased array was used to test the novel
driving method. During the experiment two frames of driver data were alternately transmitted at a refresh rate of 10 Hz,
and diffraction patterns were acquired. Experimental results demonstrate that the proposed method can be used to keep
the data refresh rate lower than the liquid crystal switching frequency.
In order to realize laser beam steering, a driving and control system based on Voice Coil Actuator (VCA) is designed for
linear motion of micro-lens array. According to the feedback displacement signals acquired from displacement sensor,
controller of the system sends control signals to Purse Width Modulator (PWM) driver which is used to power the VCA.
VCA provides driving force for the support structure of the system which connects the micro-lens array, and then the
linear motion of micro-lens array is achieved with the movement of support structure. The effectiveness of the system
was verified through experiment. Experiments results show that the displacement resolution of the linear driving and
control system is 10&mgr;m , which satisfies the requirement for laser beam steering.
Wave front precision distortion correction and transformation using a Liquid Crystal Spatial Light Modulator (LC
SLM) is discussed. After passing through non perfect elements, a coherent beam becomes a distortional beam. The
distortion is expressed in either optical paths difference or phase difference way. The difference is measured
quantitatively. SLM modulates the phase of the beam in a pixel resolution. The beam is recovered to a plane wave
front again after the correction. The accuracy of the modulation is 0.0165λ.
Diffractive beam direction needs to be controlled agilely and precisely in beam steering and some diffraction
applications. Far field fringe generation, wave front distortion correction by employing programmable diffractive element
is a new interesting research area. A phase-only liquid crystal spatial light modulator is used to produce arbitrary
fringes expected.
For the practical application, the main technology we tried includes: 1) the fast recovering speed from arbitrary fringe
to the phase distribution of liquid crystal spatial light modulator, 2) the expected fringe light intensity information is
delivered by phase information. The previous G-S phase recovering algorithm demands more iterative calculations,
which is not suitable for real time purpose. According to the statistics of light field distribution of expected fringes, a
pseudorandom phase encoding method has been used in this work. The expected fringe light intensity information is used
to limit the initial value for phase recovering calculation. By this way, the iterative time has been reduced largely, and
the phase recovering is optimized as well. The expected diffraction fringe can be obtained through 3-5 iterative
calculations.
The experiment results show that the technology is satisfactory for fast applications.
A practical numerical model for liquid crystal cell is set up based on the geometry structure of liquid crystal optical phased arrays (LCOPA). Model parameters include width and space of electrodes, thickness of liquid crystal layer, alignment layers, electrodes and glass substrates, pre-tilted angles, dielectric constants, elastic constants, the refractive indexes and so on. Especially, the thickness of alignment layer is first considered to the best of our knowledge. According to electrostatic field theory and Frank-Oseen elastic continuum theory, two dimension (2D) electric potential distribution and 2D director distribution are calculated by means of the finite difference method on non-uniform grids. And the phase delay of LCOPA is derived from crystal optics. The influence of cell sizes on phase delay distribution is analyzed. The evaluation function of fringing field effect is provided. And the methods to decrease fringing field effect between electrodes are also discussed.
In order to obtain the phase modulation characteristics of liquid crystal spatial light modulator (LC SLM) a polarization interference method is introduced to measure the phase shift. In this method the polarized components of the light reflected from the LC SLM will interfere with each other after crossing two polarizers. The phase shift can be determined by analyzing the interference intensity. Since the two interference components of the light travel through the same path, the measurement setup is insensitive to vibrations and air currents, and then the accuracy of the measured data can be improved. The experiment results show that the phase modulations corresponding to different pixels of the LC SLM are relatively uniform across the aperture. The phase shifts will change linearly over 0 to 2π with gray levels when the custom look-up table (LUT) is used.
The performance of space mirror lied on the properties of the mirror surface material to a great extent. In this paper, the silicon carbide thin film deposited on reaction-bonded silicon carbide (RBSC) space mirror blank was produced by Chemical Vapor Deposition (CVD) process. Some mechanical and physical properties of the SiC thin film were tested because they were important to study ability to work of space mirror. The result of XRD phase analysis indicated that the component of the SiC thin film was β-SiC. The micro hardness of the film was tested. The thickness of SiC thin film was tested using needle touch contour graph. The results showed that the thickness of SiC film was about 20 μm and film was even in the macro scope. The adhesion strength of SiC thin film and RBSC substrate was tested by scratch method and the results showed that was excellent. The residual stress of SiC thin film was tested by X-ray, at the same time; the origin of residual stress and the calculation of thermal stress were discussed. In the room temperature, the residual stress of the SiC film was compressive. After precision grinding, the surface topography and roughness of the SiC thin film was tested by Atomic Force Microscope (AFM). The results showed that the surface of SiC thin film had extremely low surface roughness and high surface form accuracy. The thermal shock resistance of SiC film was fine by tested.
Reaction Bonded Silicon Carbide (RBSiC) has long been recognized as a promising material for optical applications because of its unique combination of favorable properties and low-cost fabrication. Grinding of silicon carbide is difficult because of its high hardness and brittleness. Grinding often induces surface and subsurface damage, residual stress and other types of damage, which have great influence on the ceramic components for optical application. In this paper, surface integrity, subsurface damage and material removal mechanisms of RBSiC ground using diamond grinding wheel on creep-feed surface grinding machine are investigated. The surface and subsurface are studied with scanning electron microscopy (SEM) and optical microscopy. The effects of grinding conditions on surface and subsurface damage are discussed. This research links the surface roughness, surface and subsurface cracks to grinding parameters and provides valuable insights into the material removal mechanism and the dependence of grind induced damage on grinding conditions.
A 256×256 pixels commercial phase-only Liquid Crystal Spatial Light Modulator (LCSLM) is adopted to generate arbitrary spot patterns successfully. An improved Fourier iterative algorithm combining with pseudorandom phase encoding is proposed to satisfy the real-time processing requirement. The improved algorithm based on the (Gerchberg-Saxton) G-S algorithm with the initial phases generated by the pseudorandom phase encoding method. By employing the improved algorithm, the intensity error of a 4×4 spots array pattern is less than 7%.
On-line real-time detection method for the defect of TFT-LCD is becoming increasingly important as TFT-LCD has replaced CRT displays and become the first choice in many applications. Traditional defect inspection methods of TFT-LCD are based on clear features and exact mathematic models. However, the defects of TFT-LCD are of strong complexity and vagueness. Moreover, determining the defects is a complicated process, which is influenced by the objective characteristics of the defects as well as the subjective factors of the observer. Therefore, it is very difficult to establish the accurate mathematical models for the defects. A fuzzy expert system approach is proposed for the defect inspection of TFT-LCD. Tests indicate that this system could emulate the experts or experienced operators to realize the automatization of the defect inspectin of TFT-LCD.
The optical properties of the Ce:Cu:LiNbO3 co-doped with different concentrations of In were measured, including absorption spectra, infrared transmittance spectra and holographic storage properties. The doped threshold of In was determined via the results of spectrum measurement. When the doped concentration of In is over its threshold, the photo-damage resistance ability of In:Ce:Cu:LiNbO3 is improved dramatically. In the measurement of holographic storage properties, the response time of 34s, diffraction efficiency of 33%, and exponential gain coefficient of 18cm-1 were obtained for In(3mol%):Ce(0.20wt%):Cu(0.015wt%):LiNbO3 crystal.
In:Er:LiNbO3 crystals with fixed Er concentration of 1mol% and variable In concentration of 1mol%, 2mol% and 3mol% have been grown by the Czochralski method, and then was made into waveguide substrates. The directly observing facula method was employed to measure the photo-damage resistance ability of the substrates, and it is found that In(3mol%):Er:LiNbO3 has the highest photo-damage resistance ability among these three substrates. The structure was determined by the infrared transmittance spectra and the UV absorption spectra. The mechanism that the photo-damage resistance ability of the substrates increases with the increase of the doped concentration of In was discussed via the structure of the crystal.
KEYWORDS: Image processing, Digital signal processing, Signal processing, Human-machine interfaces, Liquid crystal on silicon, Control systems, Video, Interfaces, Image filtering, Amplifiers
Along with the rapid development of wearable computer and multi-medium technology, many units need to be mounted on the helmet such as micro display, camera, voice input and output components, even some devices for safety and security purpose. If these units and components are controlled by the wearable computer directly, it would make the interface between helmet and wearable computer complicated. The better way is to add a controller to the helmet, then the wearable computer only need to interface with the controller. The helmet controller controls all of the functional components of helmet. Of cause, it should be noticed that the dimensions of the components must be small since the volume of helmet for the controller is very limited. The core of the helmet controller we designed is composed of a digital signal processor (DSP) and field programmable logical array (FPGA). The DSP carries out the function of encoding, decoding, compression, encryption, synthesis, and filtering of image and voice signals. FPGA drives and controls a micro display, controls the functional components, as well. All of these reduced the amount of elements, enlarged the integration level, which realized the helmet controller microminiaturization.
KEYWORDS: Digital signal processing, Signal processing, Signal attenuation, Image processing, Image transmission, Modulation, Infrared cameras, Sensors, Imaging systems, Copper
This paper focuses mainly on DMT transmission technology applying to downhole image transmission to extend bandwidth, also considers the details about using digital signal processor (DSP) to compress, process and reconfigure the image taken from infrared CCD camera and to process signals from multi-parameter sensors. The system is composed of multi-parameter sensors, infrared camera, TMS320c6x DSP for image compression and reconfiguration, ADSL modem chipset, ground computer, etc., which extends the bandwidth, decreases signal attenuation and abates the noise. Therefore system implements real time downhole imaging, multi-parameter monitoring and inspection by two-wire transmission over 7000 feet no repeaters.
At first, this paper introduces the principle of our section outline sensor using the single-stripe pattern. At second, the measurement error of our sensor is analyzed: the error factors causing the measurement error of our sensor are found out; the mathematical expressions of the measurements are deduced. At third, the measures to reduce the measurement error and the design rules of our sensor are given. At last, a design example is given.
This paper describes the principle of the spatial encoding- based rangefinder. Its parameters to be calibrate are found out. Mathematical expressions are deduced of its measurement errors caused by calibration errors. Analysis shows that it is difficult to calibrate the parameters separately, especially some parameters are unable to calibrate. So we present a new calibrating method based on table. In the normal operation of the rangefinder, it directly obtains the ranges the pixels of the image of the object from the table after giving its codes. Our calibrating method has the advantages of simple and high accuracy.
Generation of scanning structured light is an initiative triangulation-based method for acquiring range data. Unlike traditional projection of paten diaphragm, diode laser, rotating mirror and CCD camera are employed into our device. Diode laser is modulated to emit a modulated slit laser beam. Reflected by rotating mirror, the modulated slit laser beam flashes in different frequencies and generates several new type strip structured images in sequence. The strip structured images are characterized with a new pattern code which has higher resolution, speed and robustness than simply binary code. The new, highly efficient, binary code can shorten the measuring time greatly. In order to raise the resolution, the modulated silt laser beam, reflected by rotating mirror, should scanning all the measured space for eight times. As sweeping the measured space, the scanning step of modulated slit laser beam must be synchronized with rotating mirror and CCD camera. When the scanning is completed, the CCD camera take the images and store them in image-storage timely. All these work coordinately under the control of phase-locked loop. By applying the new scanning technique, a depth image with resolution of 1/256 can be obtained in just 0.3 second.
The opto-contact mini-displacement measuring instrument is mainly for mini-displacement measurement. It not only can measure small geometry size, for example silicon chip thickness, but also can measure some other parameters such as small translation, waviness, diameter run-out, vibration. Combining optical, mechanical, electronic and computer technologies, this instrument can do some work automatically like dynamic sampling, real-time processing, on-line measuring. The instrument is based on laser triangulation, it is composed of a CCD, an optical system and a computer, which can sample and process data with high speed. The measuring principle is described as follows: reflected or scattered light by measured surface are received by CCD, since a good relationship between the offset of image point on CCD and the mini-displacement of object located on a reference plane, a mathematical model can be founded, then the mini-displacement may be calculated according to the offset of image point position on CCD. It's measuring range is +/- 500 micrometers , and precision resolution is +/- 0.1 micrometers .
Space encoding is the effective method saving image- acquisition time in structure light 3D vision. In this paper, we propose calibration parameters, which do not include camera inner parameters, of the rangefinder using scanning-based spatial code pattern. We discuss the influence of calibration errors on rangefinder measuring. A model of the systematic calibration errors is developed based on the geometry of the rangefinder. According to given rangefinder accuracy, the limits to calibration errors are set from the model. An example is cited. Moreover, we discuss the influence of the shifting of the camera lens center and the polyhedral mirror reflecting point on the rangefinder precision.
This paper describes a kind of profiler with the above features. It uses the white-light interfering method and adopts the structure of Micelson interferometer. It includes the light bulb used as the light resource, CCD used as the sensor, PZT providing microscopic translating and the computer sampling and processing data with high speed. When the reference-mirror translating, computing and comparing the modulation M of each point of the surface will give their relative surface height. Then we obtain the surface profile of object to be measured.
In this paper, the principle of range-image based on image coding is introduced. According to the encoded images obtained by our laser scan-based image-encoder, the algorithms are presented to produce the threshold value, turn the encoded images into binary-state, calculate the scanning angle and find the range image. The corresponding utility programs are completed, and verified by our experiments.
Classical machine vision is difficult to be used as practical 3D vision because it loses depth when projecting 3D objects into 2D images and then has ambiguity when reconstructing 3D object. In contrast, range images, i.e. 3D images, keep the invariance properties of 3D objects, which makes the image processing tasks easier. This paper develops the prototype of spatial encoder using laser scanning method. Range images are obtained based on encoded images produced by our encoder.
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.
This paper deals with a fast rangefinder based on laser scanning space-encoding method. In the first place the principle of image-encoding is introduced and the calculation formulas of measurement values in 3D rectangular coordinate system are given. In the second place the design rules are put forward at the angle of the measuring volume, shading light and received light power properties. At last the measurement errors caused by quantization errors, the resolutions and their errors are discussed through error analysis, and then the design rules to reduce effect of quantization errors are presented.
This paper deals with a laser scanning device used in a space-encoding rangefinder. At first the principle of space- encoding is introduced. And then the error expressions of the rangefinder are given on the basis of establishing its mathematical model. Furthermore are deduced the functional relations between the measurement errors and parameters of the laser scanning device as well as errors of the parameters. At last approaches to reduce the measurement errors caused by the laser scanning device are found out according to the above said error expressions and functional relations.
In 2D intensity image, the problems often cause wrong object recognition such as: los of depth information, shadows, overlapping objects. The main advantages of range image are the acquisition of depth information, which makes the separation easy, but it has blind areas; the boundary with special direction cannot be found. Using both intensity image and range image to do 3D object recognition can make full use of the advantages to avoid the shortcomings of each kind of image. Edge detection, grouping and linking, image segmentation, and outline determination are performed in both images. If the match between the two results is good, either one may be accepted, but if the match error is big, we need to confirm which one will be accepted or rejected. After examining each step, the criteria are used to form the outline of the object, then, the type, size, position, and orientation of each object can be calculated and determined.
An instrument for fine-machined pipe (longer than 10 m) inspection and measurement is introduced in this paper. The non-contact interior wall roughness measurement system based on light reflecting and scattering theory has a wide range (Ra from 0.012 micrometer to 4 micrometer) and is suitable for different types of machined surface measurements. The interior diameter is measured by a pair of gratings which produce Moire fringes. The accuracy of which is better than 5 micrometer for a pipe diameter greater than 100 mm. A CCD video camera inspection system is also installed on the measuring head to check the interior wall quality.
This paper describes a laser scanning range finder that may produce a real time space encoding pattern. Laser diode is modulated by a programmable controller. So a space encoding sequential constructive laser beam is employed as an active illuminating source. Using less frames of space encoding patterns to produce a range image with high resolution is the main goal of this study. The calculation of 3D coordinates of each point on the measured object is explained briefly.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.