For ultra-precision Numerical Control (NC) machine tools in advanced manufacturing industry, a high precision measuring method for the taper of a taper gauge was proposed by combining the integral and fractional parts together. The primary principle and the composition of the system were described, and the experimental measurement of a taper gauge was implemented. Moreover, the measurement uncertainty of the system was analyzed. In this method, the high precision indexing system containing a multi-tooth indexing table with an accuracy of 0.10″ was used to ensure the precision and reliability of the integral part, and the laser polarization interference device and a precision rotary table was combined into an angle measuring system to accurately measure the fractional part. Then, the dual-beam white light interference system was used for an accurate positioning. The experiment and analysis results indicate that the measurement uncertainty can be less than 0.3″ when the uncertainty confidence level (k) is 2. The method can realize the high precision taper measurement for any taper gauges.

Based on the Y-shaped cavity frequency splitting method, an orthogonal polarized He-Ne laser was designed and implemented. The structure and basic principle of the laser were introduced in detail, and the transverse and longitudinal models, output powers, the ellipticity of output light and frequency difference lock-in phenomenon and frequency difference tuning characteristics were experimentally investigated. By changing the voltage of piezoelectric transducer applied to the common cavity mirror, the power tuning characteristics of S polarized and P polarized lights were observed and summarized. Furthermore, the interaction mechanism of S polarized and P polarized lights as above-mentioned was also analyzed. Compared with the traditional dual frequency lasers, the characters of the laser, as well as its merits and demerits on the frequency difference stability were presented. Experiments show that the maximum output power of the laser is about 235.2 μW. By adjusting the voltage of the piezoelectric transducer in the P cavity, the frequency difference between the S polarized and P polarized lights is linearly varied from 22 MHz to 1 018 MHz and lock-in frequency difference of the laser is about 22 MHz.

According to the surface scattering of a mirror surface at the short wavelength band, a resolution calculation method based on the surface roughness of mirror was presented for an Extreme Ultraviolet(EUV) telescope. First, the propagation of scatter light in two mirror systems was analyzed in details, and the relationship between the distribution of light energy on focal plane and the ratio of mirror surface roughness to wavelength was discussed. The surface roughness of the mirror was measured in different wavelength bands and the one-dimensional Power Spectral Densities(PSDs)on all relative spatial frequency ranges were fitted by a k-relevant model. The numerical results show that the effective RMS surface roughnesses of primary and secondary mirrors are 0.59 nm and 0.77 nm in the spatial frequency from 1 / D (D is the diameter of the mirror)to 1 / λ(λ is an incident wavelength). The optical design software Zemax was used to build a non-sequence model of EUV telescope which contains the information about mirror surface roughness. This model can reveal the behavior of scatter light on the mirror surface and can calculate the resolution of the image plane. The results indicate that the radius of 80% energy changes from 3.9 μm to 4.3 μm and the corresponding angular resolution of the telescope is 0.25″. This method is convenient, effective and suitable for imaging evaluation.

As the mismatch of thermal expansion coefficients among the optics, frames and the bond layer can bring a radial stress on the optics and can affect the properties of optical systems, this paper researches how to select the athermal bond thickness to minimize or cancel the radial stress. First, the existing derivations for the athermal bond thickness based on Hooke's Law are compared, then a novel way for constraining the bond is proposed and a simplified approximation formula for the athermal bond thickness is derived. By modifying the axial constrain of the bond layer, a modified approximation formula is also derived. Furthermore, a series of bonded lenes with different bond aspect ratios are built, the simulation value of athermal bond thickness of each assembly is solved by the finite element analysis. Finally, the simulation and analytic solutions of athermal bond thickness are compared, and the application scopes of the analytic equations are pointed out . According to the comparison and analysis, it indicates that the modified approximation formula is applicable to the bond layer with a aspect less than 1/10, the approximation formula is suitable for that with a aspect in 1/10~1/3; and when respect ratio is more than 1/3, the simplified approximation formula is more appropriate. In their application scopes, athermal bond thickness equations have adequate accuracy, and can satisfy the requirement of engineering applications.

Based on the Extended Nijboer-Zernike theory, the effect of different amplitudes for exit pupils on the image intensity in the focal plane was analyzed. A novel approach was applied to testing the wavefront according to the actual condition of the amplitude in the exit pupil, which can help eliminating the error caused by the nonuniformity illuminated pupil and the Fast Fourier Transform in the original phase retrieval algorithms. A testing experiment was performed on an imaging optical system, and obtained results show that the tested wave fronts in the exit pupil of a camera lens are 0.196 5λ in PV and 0.022 4λ in RMS (the testing wavelength λ is 6328 nm). The aberrations in the wavefront are mainly astigmatism, coma and high order astigmatism. Furthermore, the approach can also be used to analyze the amplitude in the exit pupil of a camera lens and calculate the light intensity distribution on other focal planes. The experiment proves this approach available.

To meet the need of soft X-ray self-supported transmission gratings for space detecting and plasma diagnostics, self-supported gold transmission gratings were successfully fabricated with the holographic lithography and plating technology. To obtain photo-resist mask patterns with steep sidewalls, a layer of Anti-reflection Coating (ARC) was used to attenuate standing wave resulting from the substrate with high reflectivity. Then,a plating mask was obtained by coating a protective chrome on the mask and using Reactive Ion Etching (RIE). In this process, the plating mask with desired duty cycle could be fabricated via changing the thickness of the protective coating. Finally, a self-supported transmission grating with a line density of 3 450 gr/mm (290 nm period), a duty cycle of 0.55 and 400 nm thick gold bars was fabricated by plating gold in grooves of plating mask. The grating size is up to 10 mm×15 mm, and its +1 order diffraction efficiency is about 5% in the soft X-ray band between 5 and 12 nm in the National Synchrotron Radiation Laboratory (NSRL).

As the Quantum Detection Efficiency (QDE) of a spherical micro-channel plate in a photo-counting imaging detector is relative to the sensitivity of an Extreme Ultravialet(EUV)camera, this paper researches its characteristics and measurement. The secondary electron yield of the spherical MCP in EUV region is calculated based on the theoretical model describing secondary electron yield of the MCP, Then the QDE of the spherical MCP is measured by using a laser-produced plasma source, and an instrument for measuring the QDE of spherical MCP is built. The QDE of spherical MCP for different positions on a MCP surface is measured at different incident angles and wavelengths. The experimental results show that the QDE at an incident angle of 15° is five times of that at incident angle of 0°。

Stray light characteristics of an off-axis three-mirror optical system with a middle image and long-focus was simulated and analyzed, and corresponding suppression method for the stray light was proposed.however, little is researched on the stray light analysis of this structure. In order to satisfy the stray light suppression requirements of modern optical instruments, a simulation analysis is executed for the designed optical system. By a forward and reverse ray tracing method, the main source of stray light was identified and an important surface of the system was found out. According to the analysis results, some corresponding suppression measures were proposed to reduce the stray light of the optical system. The effect of stray light reduction were evaluated by comparing the number of important surfaces before and after the measures. The analysis results show that the first-order scattering stray light has been well suppressed. Furthermore, the Point Source Transmission (PST) was used to evaluate the stray light suppression capability, and its quantitative analysis was impletmented by calculating the PST of ±40° off-axis angles in EW and NS,respectively.

Fabrication techniques of lightweight silicon carbide (SiC) mirror blank were researched and the key problems in the fabricating process were discussed. An advanced lost mould method was proposed to fabricate the excellent lightweight structure with a partly open-back. In consideration of the difficulty on fabricating SiC ceramics with larger sizes and complex shapes , the gelcasting for a SiC green body was studied and the ingredient of the SiC slurry with high solid phase and low viscosity,control of slurry time and the drying processing for the wed bland in the gelcasting were discussed. Thermogravimetic and Differential Thermal Analysis TG/DTA curves for the green body were analyzed, then the best sintering process was established. By the proposed technologies, RB-SiC mirror blanks with different shapes and lightweight constructions were prepared and the largest one is 1 100 mm×820 mm.

A type of ultrathick double metal cladding optical waveguide with a submillimeter scale was used to investigate the narrow bandwidth filter of magnetic fluid, in which the magnetic fluid were injected into the sample as a guided layer. While the 1 550 nm laser beam was incident on the waveguide with resonance angle, ultrahigh order modes would be excited; the magnetic fluid was in the oscillating field. Then based on ultrahigh order modes, the refractive index of magnetic fluid was changed owing to the strong concentration of the magnetic field. The characteristics of narrow bandwidth filter were obtained by detecting the reflected beam on waveguide. In the experiment, the wavelength of the incidence beam was to be form 1 549.440 nm to 1 549.585 nm, and the bandwidth filter to be 0.06 nm. Obtained results show that the change of reflected beam is 11.9 dB when the magnetic field changed from 0 to 519 kA/m in the wavelength of 1 549516 nm. Experimental results indicate that the optical filter has narrow waveband and higher sensitivity.

A helical undulator introduced by National Synchrotron Radiation Laboratory was reformed and its performance was researched. The helical undulator was installed at the linear part of a storage ring to produce coherent synchrotron radiation with high flux and circular polarization and to provide a VUV beamline. Based on improved source parameters, the polarization radiation characteristics of the helical undulator were analyzed and the power density distribution radiated by the helical undulator, thermal distortion and the slope error on the premirror surface were calculated with ANSYS. The results show that the maximal thermal distortion of the premirror is less than 2 μm in a photon energy range desired by a burning experiment and the maximal slope error is about 1 μrad. Obtained results satisfy the requirement of SR experiment very well.

A new large format focal-plane shutter with high frame frequency is designed to the space camera with a transmission optical system and a focal length of 1 000 mm. The focal plane shutter adopts a single curtain with multiple slits in structure, and two shutters are orthogonal layout and are driven synchronously by a motor with gears. The design allows to exposure of two stitched area CCD simultaneously. A Digital Signal Processor (DSP) is used to control the motor velocity according to the curtain encoder, so as to ensure the exposure uniformity of the CCD. The experiment results show that this new shutter has a higher stability of movement, a long life span more than 2×105 operating times, a wide range of exposure time in 1/ 200 s to 1/ 1 000 s and a exposure error within 8%.Moreover, the max frame frequency can be 4 frame/s. The performance of this shutter is fully satisfied with the requirements of area scan CCD aerial cameras.

An Atomic Force Microscope (AFM) system with high-speed and large-area was constructed to measure the micro-structure surface of an optical grating. The effects of scanning speed on measuring results under different scanning modes were researched. First, the spectra of the micro-probe under constant-height and constant-force modes were measured, respectively, and the effective bandwidths of the probe were obtained under the two modes. Then, based on the constant-height and constant-force modes, the effects of scanning speed on measuring results were analyzed via measuring a line and a circle on the surface of the optical grating at different scanning speeds. By employing this AFM system, the 3-dimensional profile of the large-area micro-structure surface on the optical grating was measured at a distortionless scanning-speed in a constant-height mode. The results show that it takes only 40 s to measure a circle area with a diameter of 40 mm on the grating surface. When the scanning speed is no more than the speed that is corresponding to the effective bandwidth of the micro-probe, the AFM system can achieve the high-speed, large-area and distortionless measurement for micro-structure surfaces.

To improve the precision of a bending mirror and to reduce the influence of a horizontal deflected bender and the mirror weight on the bending mirror,the factors that caused the slope errors were studied and a gravity compensation scheme for the horizontal deflected bender was presented. Firstly, the slope errors were studied on the basis of bending theory. Aiming at the bending focusing mirror used in the beamline, the main factors that affected its precision were analyzed by finite element method. Then, according to the characteristics of the horizontal deflected bender and the results of slope error analysis，a scheme was carried out to compensate the slope errors. The results indicate that the gravity of the mirror and a cooling device mainly had influence on the cylinder mirror generatrix, while the gravity of the driven-rod mainly affected the tangential slope error. After compensation，the slope error rms of cylinder mirror generatrix is reduced from 13.14 μrad to 015 μrad and the tangential slope error rms is reduced from 8.21 μrad to 0.86 μrad. It verifies that the slope error analysis and gravity compensation scheme proposed is effective, which improves the precision of the bending mirror greatly.

As current high-speed scanning Atomic Force Microscope (AFM) is mainly designed for the biological imaging application and its scanning speed and scanning range should be improved, a novel high speed AFM was designed. Based on the flexure guide structure driven by piezo actuators, the AFM scanner with a large range was proposed, by which the AFM scanning range is expanded to 100 μm×100 μm in the x-y directions. By the Fourier expansion, the high harmonic characteristics of the common triangle and sinusoidal driving signals were analyzed, and their effects on the high speed scanning image were discussed. To avoid the mechanical self-oscillation of the stage during scanning, the sinusoidal driving signal was taken to drive the high speed scanning and the line-scan speed was up to 50 line/s. Finally, a new method to eliminate AFM nonlinearity error based on positioning sampling was designed. This method effectively reduces the image distortion resulted from nonlinear errors of the AFM scanner.

In order to decrease the nonlinearity of a Micromechanical Tunneling Gyroscope(MTG), enlarge its band width and raise the signal-to-noise ratio of the system, the tunneling gap between the tunneling tip and the corresponding tunneling electrode should be controlled in 1 nm, moreover the MTG must operate in the closed-loop mode. Based on the anti-interference and robustness of the Linear Quadratic Gauss(LQG) control theory and time-varying characteristics of Coriolis acceleration, this paper uses the predictive-LQG control strategy formed by a series of time-varying Kalman filter and an optimal LQG controller to design a closed-loop control system for the MTG. First, the overall control scheme was designed according to the operating principle of the MTG. Then, the two series of the predictive-LQG controllers were designed on the basis of the establishment of the extended dynamic equation of the MTG. Finally, the predictive-LQG controller system of the MTG was built up by Simulink, and a numerical simulation was performed. The simulation results prove that the tunneling gap has been controlled in 1 nm and the measured input angular rate could be estimated accurately by the predictive-LQG controller even if the input angular rate is a slowly varying random signal. And the estimation accuracy can reach 10-4rad/s.

To assure the pointing accuracy in a shooting target to be better than 30 μm for the Shen Guang Ⅲ(SGⅢ)facility,the switchyard of target area for the facility was designed and the rigid frame and transport mirrors in the switchyard were optimized to improve their stability. In design, the reinforced concrete was used to support the switchyard, and the tilted supports were added in the steel frame of switchyard to decrease and distribute the mass load reasonably.Furthermore, the herringbone supports were firmed to local horizontal beams. The transport mirror supports were also designed, in which their foundational frequencies were much higher than that of the swichyard. By above ways, the angle shifts of transmission mirrors in random broadband vibration can be suppressed successfully. The mechanical analysis of range structure by Finite Element Analysis(FEA) indicates that 87% of transport mirrors show their angle shifts less than 0.48 μrad. According to the relation between the rotation drifts of the optics components and the positioning errors of optical path, it can be seen that the rotation drifts of the transmission mirrors could meet the stability requirements for the positioning error of each single optical path not more than 12.9 μm.

A tracking and guiding system based on machine vision for a laser-electronic theodolite was designed to improve its work efficiency and implement the dynamic measuring in a large space. Systematic architectures of the system was introduced and the motion characteristics of outrigger scanning instruments and laser-electronic theodolite were analyzed. Based on establishing a motion model of scanning instruments, the strategy of automation guidance for laser spots of the theodolite was realized by coordinate transformation and the process of searching the optical axis of a lens. Then, the method to focus and control the lens was given. Finally, the dynamic tracking method using extended Kalman filtering estimation for a moving object was proposed to locate the target and laser spots in a center region of the image simultaneously. The experimental results indicate that the status of the moving object can be predicted accurately and the final guiding accuracy for laser spots has been reached 008(°)/m. In conclusion, the proposed system combines the advantages of the dynamic measurement process for laser tracking and no limitation in measurement styles for theodolite, which allows the laser-electronic theodolite systems based on traditional angle forward intersection to dynamic measuring fields.

A TiN coating was modificated by microscale Laser Shock Peening (μLSP) based on the laser spot with a diameter of 300 μm to further improve its surface properties. An optical profiler and a nanomechanical test system were used to observe and measure the surface morphology, nano-hardness and the elastic modulus of the TiN coating in a peening area with different μLSP parameters. Experimental results show that a conical dent is formed in the peening area after μLSP, and the biggest depth is about 1.2 μm. The nano-hardness and elastic modulus are significantly changed in the center of peening area, and they are enhanced with the increase of laser energy in a certain range. When the laser energy is 250 mJ, both the nano-hardness and elastic modulus reach the maximum, which are 50.21 GPa and 402.8 GPa, respectively. Compared with the initial state of the TiN coating, they have been increased by 140% and 43.9%,respectively.

According to the requirement of the assembly of a shaft and a sleeve for coaxiality, a coaxial positioning system was presented based on the flexible decoupling beams and micro-vision technology. To implement the online adjustments and maintain posture of the shaft and the sleeve, a flexible beam structure with a square pattern and two degrees of freedom was designed. Combined with a micro-platform and a clamp, this structure has achieved an exact coaxial precision in the positioning. According to mechanics, a stiffness model of the flexible beam was established. Then the parameters of the beams were determined by a simulation in finite element method. A micro-vision system was designed to be used in the system. Through the Hough algorithm, the precise coaxial positioning of the real-time detection was realized and the coaxial deviation computed was used as the value of the adjustment. Through lots of coaxial experiments, it shows that the coaxial positioning accuracy has been less than 8 μm. The results indicate that the positioning device can provide precise positioning of parameters, and can guarantee the realization of positioning assembly.

A concept of Asynchronous Bundle of Linear Ultrasonic Motor (ABLUSM) was proposed to improve the resolution, thrust and the speed of driving unit for a precision positioning stage. It points out that a numbers of stators (vibrators) alternately drive a mover (rotor) in a vibration cycle, the actual driving frequency is increased doubly, and better output performance can be expected. An analysis shows that increasing the contact frequency between stator and mover is able to improve the output force, speed and transient response characteristics of a Linear Ultrasonic Motor (LUSM). Meanwhile, it can avoid the mutual interference of different stators in Synchronous Bundling of LUSM (SBLUSM). Experimental results of two types of motors show that the maximum load force of the ABLUSM is 30% upon that of the SBLUSM or 25% over that of the sum of two motors when they drive the slider respectively. Results of minimum resolution test show that the minimum displacement resolution of the ABLUSM is 32.6 nm, which is 62.8% that of a single LUSM or 38.5% that of the SBLUSM. The mechanical characteristic experiments of the LUSM with one driving tip also prove that the no-load speed of ABLUSM has increased by 10% as compared with that both of the single stator and the SBLUSM. Moreover, the maximum efficiency has by 50% over that of the SBLUSM. Transient response time is shorter than SBLUSM, which is about 50% of a single stator.

On the basis of a design idea integrated with optical, mechanics and electrinics, this paper presents the techniques of a fine scanning mirror assembly with high resolution for space optics remote sensors. Firstly, the driving scheme of the system is discussed according to the configuration of a camera, and the step motor is chosen as a core of the control system. Then,three parts of mirror assemblies: integrated driving module, fine shafting and SiC pointing mirror subassembly are introduced. Furthermore, the accuracy of the pointing mirror assembly is tested. According to the test results, the angle precision of is better than 40″, the precision relative angle rate (rms error) is under 7% at the working speed and the main shafting radial run-out error is 2 μm. The results prove that the pointing mirror assembly satisfies all the requirements of the optical remote sensor.

For reducing or avoiding the iatrogenic nerve function injury due to the excessive retraction for a nerve root during a spine surgery, a silicon piezoresistive pressure sensor was employed in the nerve root retractor to quantitate the retraction for the nerve root and to improve the performance of spine surgery. To measure the pressure exerted upon the retracted nerve root, a pressure sensor was adhered to the nerve engagement surface of a nerve retractor. The rabbit and goat experiments were performed on nerve root retraction by different time durations and pressure extendings,while the nerve function injury was assessed by monitoring the electromyogram responses. Experiments show when the increase of average pressure is about 14 kPa, the change rates of latency and amplitude are 112% and 21.21%, respectively, which reveals that the retracted nerve root axonal is damaged. Furthermore, loading similar average pressare, the more the retraction time applies to the nerve root, the worse the nerve function is hurt. In conclusion, the relationship between the pressure exerted on the nerve root and the corresponding neuromuscular responses has been founded via the results of animal experiments, which shows the degree of nerve root impairment is positively correlated with the duration and extent of nerve root retraction, and can be used in the study of retraction threshold values.

This paper focuses mainly on the mathematical model of ground target tracking of a video satellite and the design of an attitude tracking controller. Firstly, the desired attitude quaternion and the desired attitude angular velocity are presented respectively according to the orbit kinematic and attitude kinematic. Then, a feedback PD controller based on the error quaternion and error attitude angular velocity is designed, and the stability of the designed controller is analyzed by using the Lyapunov stability theory. By setting the deviation of area CCD in the exposal time to be smaller than 03 pixel, the minimum attitude pointing error and attitude velocity error are induced. Finally, physical simulation of staring-imaging by a video camera is conducted on the minitype three-axis air bearing table. Simulation results suggest that the pointing error is less than 0.1°, and that the attitude velocity error is less than 0.012(°)/s, which meets the requirements of staring-imaging of area CCDs.

To satisfy the requirement of assembling a capsule and a fill tube for the cryogenic targets in ICF experiments for the 3 pL adhesive volume, the time/pressure micro-bonding technology was studied. A microinjector for life science research was changed to a time/pressure pL micro-bonding device, and a capillary needle made of quartz with inside diameter of 0.5 μm was used for dispersing. A micro-bonding station was built for experiments. Two microscopes interfaced with CCD and resolution of 1.57 μm were designed to detect the position of the capillary tip and the size of the glue drop, and a XYZ motion stage was employed to control the position of the capillary tip. The hardware composition, control system, dispensing process of the micro bonding station was introduced, and the dispensing control, adhesive volume detection, and glue drop positioning experiments were performed to research the key dispensing processes, such as the main factors that affects the volume control and the minimum volume that can be controlled. Experimental results indicate that the designed station can control the minimum adhesive drop in 2 pL using UV Epoxy adhesive with viscosity of 5P in the normal temperature environment, which satisfies pL dispensing requirements in the micro assemblies.

It is difficult to obtain a nanoimprint stamp with low cost and high quality in fabrication of Distributed Feedback(DFB)gratings. In this paper, the quartz stamp for semiconductor DFB gratings was fabricated by using double-layer metal mask and lift-off lithography. Firstly, the photoresist pattern of a DFB grating was written on a Ti coated quartz substrate by electron beam lithography, then a layer of metal Ni was sputtered on it to obtain a corresponding pattern by metal lift-off technology. Finally, the Inductively Coupled Plasma(ICP)dry etching was used to transfer the pattern on the quartz substrate. The scanning electron microscope image of the fabricated stamp shows good uniformity with sufficiently low line edge roughness. The stamp is suitable to fabricate DFB grating for 1 310 nm semiconductor laser diodes using UV nanoimprint lithography.

To decrease the influence of friction on the angular speed tracking accuracy of a precision servo turning table, the methods of friction parameter identification and friction compensation based on the LuGre model was put forward. Firstly, the angular speed over zero phenomenon occurred in the angular speed free reduction was analyzed, and the genetic algorithm was used to fit the angular speed reduction curve to obtain LuGre model parameters and rotational inertia. Then, the parameter identification method was tested and verified by a simulation experiment. Finally, the friction compensation was computed by identfication parameters and was added to the angular speed tracking system of the turning table . The actual experimental system achieves a angular speed loop and a current loop and its data update rate is 1 kHz. In the actual angular speed measuring experiment, the mean square deviation of error between actual angular speed curve and fitting curve is 2.9×10-3rad/s. In the angular speed tracking experiment, when the given signal is sine wave with an amplitude of 0.032 4 rad/s and frequency of 0.5 Hz, friction compensation can eliminate the waveform distortion, and the mean square deviation of angular speed tracking error has decreased by 27% as compared with that of double loops control system without friction compensation. It is concluded that the friction parameter identification method can obtain high accurate parameters, and can improve the angular speed tracking accuracy of servo turning tables prominently.

In order to improve the real-time performance of the Gradient Vector Flow Snake(GVF Snake) algorithm for face contour tracking in a dynamic image sequence and to overcome the occlusion problem in face tracking, a novel image extraction method combining the GVF Snake algorithm and the single variable first-order grey model GM(1,1) is proposed to extract the face contour. In this method, the moving face contour is roughly detected out firstly by using human motion information and the skin-color model, and then the accurate face contour is extracted by using the GVF Snake algorithm, by which the initialization problem of the GVF Snake algorithm is solued. For the integrity feature of face contour motion, the GM(1,1) model is used to predict the centroid position of face contour and then the position is used as the iteration basis of the GVF Snake algorithm. Meanwhile, the centroid position of face contour extracted with GVF Snake is taken as the prediction basis of the GM(1,1) model for the next frame. When the occlusion exists, the continuity of tracking can be held with the prediction of GM(1,1) model. Experimental results show that by proposed method, the average tracking time and the average tracking error are only 8.0% and 312% of those of the GVF Snake algorithm respectively. It can be concluded that this method can better reflect the motion law of face contour, and has strong real-time performance and good robustness.

According to the low measuring accuracy from a flexible coordinate measurement machine, the error correction and parameter calibration methods for the coording measarement machine were researched. An improving genetic algorithm based on optimization least square method was proposed to implement the parameter calibration. The kinematic model and the error model of the flexible coordinate measurement machine were established by Denavit-Hartenberg(DH) method. Firstly, a variable factor was used in least square method to evaluate the convergence speed. Then, the suboptimal parameter was regarded as the initial population of optimized genetic algorithm while the variable factor became steady. Finally, the improving genetic algorithm was used to search and calculate to obtain the optimal parameter and the parameter calibration was finished. The experiment shows that the proposed algorithm has fast convergence speeds and good robustness.

A fast and robust fundamental matrix estimation method based on Sequential Similarity Detection Algorithm (SSDA) is presented to estimate the fundamental matrix rapidly and accurately. The SSDA is introduced into the Maximum a Posteriori Sample Consensus (MAPSAC) to search the optimum model parameters and the accumulation times of computing a cost function are cut down by eliminating the false model as soon as possible, which not only keeps the better robustness of MAPSAC, but also reduces its computation effectively. Then, the initial inliers obtained by the improved MAPSAC are optimized with a M-estimator. Those inliers with larger residual errors are removed and the optimized inliers are used to compute the fundamental matrix to enhance the precision and improve the robustness of the algorithm. Experiment results demonstrate that the proposed algorithm performs better in accuracy and robustness, and its average speed has increased at least 30% as compared with that of the MAPSAC. The proposed algorithm can satisfy the requirements for real-time, precision and robustness in the fields such as three-dimensional reconstruction, image matching, image tracking and camera self-calibration.

An improved Canny algorithm is proposed to remove ring artifacts in Computerized Tomography (CT) images and to improve their qualitative analysis ability and quantitative measurement precision. S-L filter is adopted to filter original projection data and enhance artifact information before using Gaussian filter. Then, the angle thresholds are set to limit gradient direction, and to exclude diagonal direction edge detection and implement vertical direction edge detection. Furthermore, a gradient threshold and an angle threshold are set to detect and connect the marginal point of artifacts. Finally, the piecewise B-spline fitting method is used to correct the projection data and eliminate multiple continuous ring artifacts. Correction experiments are carried out for the actual CT images including ring artifacts. The results demonstrate that the proposed approach can remove all of the ring artifacts of reconstructed images, keep the standard deviation in detail areas basically unchanged, reduce the standard deviation in uniform areas and increase the Signal to Noise Ratio(SNR) by 2.182 dB. The proposed method is effective in correcting not only the multiple continuous ring artifacts but also the single dispersed ring artifacts, at the same time the CT image details and resolution are reserved.

In this paper, a parallel stereo matching algorithm is proposed based on belief propagation algorithm to accelerate the matching speed. Firstly, the edge of the image is calculated using an optimized edge computing method, then the image to be matched is cut into several segments by the edge information, and every segment is matched based on belief propagation algorithm in parallel time. Finally, all the segment are assembled into a whole disparity image. The experimental results show that the matching speed has accelerated by 3.51 times that of traditional one in the same disparity quantity, and this method can lay the foundation for the further real-time application.

According to the intrinsic and extrinsic parameters of a camera to be needed in some applications for machine vision, a method based on the neural network with embedded orthogonal weight matrix is proposed for camera calibration. Firstly, the neural network with the embedded orthogonal weight matrix is structured, whose weights are corresponding to the camera extrinsic parameters and intrinsic parameters. Thus, the structured neural network coincides with the model of camera. The generation of orthogonal weight matrix is served as the last mutation operator in the iteration of genetic algorithm, and the performance index is the square of 2-norm of the difference between vector consisting of network's outputs and homogeneous coordinate of corresponding feature point projected in image plane. Meanwhile, a hybrid genetic-simulation annealing algorithm is introduced into the solving-programming. When the system comes to the equilibrium, the intrinsic and extrinsic parameters of the camera can be obtained in the light of network's weights. The simulated experiments illustrate that the proposed algorithm is robust, and has the advantages of high calibration precision, simplicity and clarity according to real experiments. It provides an effective solution scheme for camera calibration of intrinsic and extrinsic parameters.

A camera system was designed by using conventional CCDs and several problems in a high-speed mode were saved. After introducing to the working principle of a interline array CCD KAI-0340DM, a highly integrated CCD signal processor was used to produce the high-speed timing signals. By improving the heat conduction efficiency between chip and board，the heat conduction area was enlarged so that temperature rising was reduced. A high-speed op-amp self-oscillation model was established, and the self-oscillation was overcome. Furthermore, the difficulty of data processing was reduced by the serial to parallel conversion. Finally, data was transmitted through Camera Link Interface,and output data were asynchronous depressed in frequence by compressing the blanking interval. The results show that the temperature has only gone up 5.2℃ when the camera system works; its signal to noise ratio is greater than 40 dB, and dynamic range is not less than 60 dB. The camera can work under the four resolutions. When the cameras resolution is 640×480,it can work in the Time Delay Integral(TDI)mode. When it is 228×164, frame frequency is 1 000 frame/s, it can meet the camera’s basic application requirements.

To obtain high quality images for remote sensing cameras in a wider temperature change or in the temperature diversity caused by different circuit parts, an adaptive method for imaging electronics was proposed. The time-delay factors of temperature in the imaging system was analyzed, and design methods for temperature adaptability and the initial alignment of Corrlated Double Sampling(CDS) positions were given. Several experiments were performed to validate the feasibility of this method. By proposed method, the output of CCD driver was divided and was taken as a benchmark for the CDS signals, so that CDS signals can be used to track the driving signal adaptively. Then, initial positions of CDS were precisely aligned by using the lowpass filtering of a RC net. Finally, a temperature experiment by proposed compensating method was carried out for the imaging system of remote camera in -35~75 ℃. Results indicate that the position precision of CDS signals after adjusting is up to 0.2 ns. In the experimental temperature range, the maximal deviation of the CDS signals is only 072 ns as comparing with that of ideal one, which satisfies the indicators of image quality and the requirements of space application.

By taking optical acquiring 3D objects with a lens array for an example, a method to obtain the optimal pick-up distance was proposed to reduce the elemental image overlap and the crosstalk in reconstructed three-dimensional images for 3D integral imaging and display technology. By analyzing the effective imaging rates of object points and the image point utilization of objects with various sizes at different pick-up distances, the optimal pick-up distance was achieved. Compared with conventional methods for a specific integral imaging system, the product of the effective imaging rates of object points and image point utilization has increased at the optimal pick-up distance and the maximum one can increased by 34 times. Optical experimental results show that the reconstructed images has full 3D information and wealth detail, which verifies the feasibility of the proposed method. Moreover, the method is suitable for both the lens array and the camera array to acquire 3D information.