To realize the high accuracy radiance calibration for a ultraviolet detector, a radiance calibration system for ultraviolet detector was established to meet the requirement of ultraviolet remote quantification. On the basis of ultraviolet standard detector in National Institute of Standard Technology(NIST,USA), the absolute criterion was transferred to the responsivity unknown ultraviolet silicon detector, ultraviolet multiplier phototube and the ultraviolet solar blind photomultiplier. The spectral responsivity and quantum efficiency of the ultraviolet silicon detector was calibrated at 200~400 nm and 200~300 nm, and the spatial uniformity was calibrated at 350 nm. According to the guidelines for evaluating and expressing the uncertainty of NIST measurement, the relative uncertainty of transference is less than 2%. Experimental results reported here prove the feasibility establishing a radiance calibration system for ultraviolet detector by a replacement method based on standard detector.

When a deuterium lamp with MgF2 window is used as the UV-VUV radiation transfer standard source,the radiation intensity will be gradually degraduted by the window surface pollution of deuterium lamp in vacuum environment.This paper analyzes the vacuum radiation degradation mechanism of the deuterium lamp with MgF2 window. Based on this, a liquid nitrogen cooled screen device is developed to control the radiation degradation of deuterium lamp. Then,an equipment for studying the radiation characteristics of deuterium lamp in vacuum environment is established to analyze the effectiveness of the liquid nitrogen cooled screen device. The experimental results show that the average degradation rate of deuterium lamp decreases from 7%/h to 1%/h in 160~300 nm,which shows that developed liquid nitrogen cooled screen device can very effectively control the radiation degradation of deuterium lamp with MgF2 window in vacuum environment.

The application of Fiber Bragg Grating(FBG) sensor to measurement of material property is researched. The principle of Split Hopkinson Pressure Bar (SHPB) for dynamic performance measurement of low impedance material is introduced and the problems related to the application of FBG sensor in SHPB for strain testing are discussed. The relation between the strain wave propagating in elastic bar of low impedance material and the wavelength shift is derived. Then, the interrogation system for high frequency dynamic strain measurement based on FBG is set up and the surface strain of elastic bar in SHPB is measured with the interrogation system and compared with the instrument PI02-04.The experimental results show that the interrogation system has good performance during 10 kHz and its signal-to-noise is about 40 dB.Research results prove that FBG sensor has a good application prospect in the dynamic performance measurement of low impedance material.

A mechanical structure for two-dimensional fast-steering active cooling mirror was designed to control precisely the transmission direction of laser emitting and to reduce the mirror thermal distortion to improve laser beam quality. Based on the design and fabrication principles for the high power laser reflector, an active cooling mirror was adopted to improve the beam quality by reducing the mirror thermal distortion which can control the mirror distortion less than 0.1 μm. Then, an equipment was designed from three aspects in the framework, driver, controlling system and the angle measurement element. By using a x-y axis framework structure,the effect of the cooling water tube for mirror on axis system could be reduced and rotation precision of the axis system was less than 2″. The test result was analyzed and it is shown that the mechanical structure works stably and reliably, stabilization precision is less than 1″,which can satisfy the system requirements of stabilization, higher precision and rapid speed for controlling and correcting accurately the transmission direction of the beam, and can be applied widely to the high power laser system.

An only-one correlation-peak joint transform correlator based on Mach-Zehnder interferometer is presented aiming at the shortcomings in the energy dispersion of correlation peak and of Classic Joint Transform Correlator(CJTC) the input image size and relative position to be restricted. The image input of proposed method has no relative displacement between input images by Mach-Zehnder interferometer, and realizes a high intensity overlap correlation-peak in the center of detector, so that the contradiction between the input image size and relative position with the transform lens aperture and detector reception area is relaxed. Moreover,zero-order autocorrelation peak is eliminated and diffraction efficiency is enhanced by power spectrum subtraction. The theoretical analysis and experimental results show that proposed method can offer a high intensity correlation-peak and its anti-noise ability is six times that of CJTC,which enhances the recognition ability of the system. Also,this method makes good use of the space bandwidth and reduces the system costs,so that it is suitable for large size image recognition and weak signal detection in noisy environment.

提出了一种基于振幅调制纯相位滤波器(AMPOF)的旋转不变识别新算法。该算法充分利用振幅调制纯相位滤波器区分能力强的优点,结合综合鉴别函数方法,拓展了综合鉴别函数(SDF)在畸变不变模式识别领域的应用。利用各幅训练图像傅里叶变换的振幅和相位信息,逐一构建成子滤波器,并将它们加权求和得到合成滤波器;然后采用迭代算法求解出合理的权重系数。仿真实验结果表明:对于实验采用的目标训练图像和非训练图像,改进算法的相关输出峰值大约提高95%~112%;对于只有少量像素的局部训练图像,改进算法的相关输出峰值大约提高74%。这种优化算法使得振幅匹配和相位匹配更加充分,相关输出增益明显增强,能够一定程度地降低训练量。

Based on micro-interaction principle of fabricating SiC material with fixed abrasive, this paper quantitatively discussed the influence of depth of diamond abrasive formed in SiC workpiece on the material removal rate and surface roughness of optical component. The mathematical model of material removal rate and the simulation results of surface roughness were obtained, respectively. By comparison of experimental results with theory in material removal rate and surface roughness, the trend of theoretical and experimental results are basically the same and the values of results are maintained in the same quantitative level; The difference of roughness of pellets W1.5, W3.5, W5 in experiment and theory are 5.97%, 3.19%, 3.59%, respectively, which can validate the correctness of theory analysis.

A method for analyzing and evaluating stitching precision based on t distribution statistic was presented aiming at the problem of stitching precision of aspheric surface analyzed by sub-aperture interferometry based on correcting aberration. The residual error was used to implement the regression diagnosis according to the principle of stitching measurement and the feature of stitching model. Then, based on t distribution statistic,the forecast dependent variable in the regression analysis was used to assesse and analyze the stitching precision. The uncertainty in experiment result was analyzed, analyzed result indicates that the expansion uncertainty based on t distribution statistic is 0.362 3λ(0.229 μm), and the expansion uncertainty by the conventional precision evaluating method is 0.234 μm. The results of the two precison evaluating methods are idential basically,which shows the new precison evaluating algrithom overcomes the uncertainty of comparative analyzing method on the premise of ensuring the precision,and also solves the problem that the confident coefficient according to normal distribulion can not reflect the effect of sample root-mean-square reliability on the fiducial probability.

The Near Infrared Spectroscopy (NIRS) of 55 natural water samples were collected by a PbS detector with 8 nm resolution at wavelengths from 400 nm to 2 500 nm in air background and three quartz sample cells in thickness of 1 mm, 2 mm and 4 mm respectively were used to investigate the effect of different optical path difference. The nitrogen concentrations were measured by the Chinese National Standard Method (NSM), and the value of TN is used as the standard. The calibration models of the NIRS were established with Stepwise Multiple Linear Regression (SMLR) and Partial Least-squares (PLS) to analyze the TN concentration. Experimental results show that the correlation coefficient of the calibration model is 0.97 when the sample cell is 1 mm, and the Root Mean Square Error of Cross Validation (RMSECV) is 0.490 mg/L. In order to validate the feasibility of the NIRS method for TN measurement of water, the results of the NIRS method and NSM method have been compared to each other by t-test. The experimental results demonstrate that there is no significant difference between the NIRS method and the NSM by t-test on the condition of the significance level greater than 0.05;The NIRS method is promising for detecting TN in waters and wastewaters. Moreover,because of the strong influence of the water absorbance, 1 mm thickness sample cell is more suitable for near infrared spectral analysis of TN.

To improve system resolution and image quality,the characteristics of echelle grating spectrography were introduced and a method of fixing and checking for the optical system of the echelle spectrograph was researched.Proposed method could reach the precision of the design through fixing the slot, collimation lens, prism, echelle grating and focusing lens.A special fixing method for the echelle grating spectrography during the fixing process was also introduced. Finally, a calculation formula was deduced based on Reyleigh principle to check the resolution of the echelle grating spectrography. The checked result shows that the resolution of the echelle grating spectrography meet the system requerement of 10 000.These results reported here mean that propose method can offer a reference for fixing and adjusting of other precision instruments.

A focus error detection system in sub-micrometer grade is designed based on critical angle method with high resolution and low light losses. The principle of defocus detection by critical angle method is introduced and the expression of defocus is deduced with Gauss optical formula based on Fresnel formula and a reasonable hypothesis. The detection system consists of He-Ne laser, critical angle prism, four-element photodetector, signal sampling circuit and data acquisition card, etc.,the error of defocus is detected by single light path critical-angle method. Moreover,a digital filter and the normalization technology are used to extract the Focus Error Signal(FES) to obtain the defocus curve. Experimental results indicate that the focus resolution is less than 15 nm and static linearity range is ±4 μm,which satisfies the focus error detection system requirements in sub-micrometer grade and provides references for engineering application of focus error detection system with critical-angle method.

To improve the alignment precision of laser beam and target for high power laser driving device in Inertia Confinement Fusion(ICF),several kinds of laser beam aligning methods,such as inside chamber direct alignment method, outside chamber direct alignment method, and indirect alignment method with Laser and Target Alignment Sensor(LTAS) based on optical conjugate principle, were induced and their advantages and disadvantages were analyzed. Then,the precision of the last method including the image measuring error,the errors introduced by protective glass of CCD, the movement errors of guides and the calibration error of Laser and Target Alignment Sensor(LTAS) was discussed. The experiments and precision analysis indicated that the laser beam aligning precision induced by LTAS is ±10 μm, which shows proposed LTAS can rapidly and exactly align the laser and target of large-scale laser driving device.

In order to realize dynamic positioning of piezopositioning stage with high speed and high accuracy, the motion positioning model of the stage is studied. Because the dynamic positioning accuracy of piezopositioning stage is affected primarily by dynamic property and hysteresis property of the stage,this paper proposes a dynamic hysteresis model of piezopositioning stage after introducing two kinds of typical property models and their applicable scopes. Furthermore,the parameter identification method of the dynamic hysteresis model constituting of Prandtl-Ishlinskii (PI) hysteresis operator is given. Experimental studies on the TRITOR100 piezopositioning stage indicate that the dynamic hysteresis model is more accurate than both the linear dynamic model and PI hysteresis model when the stage positioning is dynamically positioned at high speed in ±900 V/s input voltage within 30 μm positioning range. The mean error of the dynamic hysteresis model is 0.16 μm, and the maximum error is 0.38 μm, which provides foundations for designing high performance control system of piezopositioning stage.

A plate-welding test and a butt-welding test for Invar 36 alloy were carried out using Nd:YAG pulsed laser as welding heat source. The effects of process parameters(laser power,welding speed,pulse duration and defocus distance) on the morphology of welding joint,welding width and the penetration were analyzed and the hardness,composition and tensile strength of the fusion zone for butt-welding of 0.85 mm thick Invar sheets were tested. The results indicated that the main factors affecting the weld bead penetration, weld width and Heat Affect Zone(HAZ) are laser power and pulse duration. Moreover,the welding speed effects on the fish-scale spacing and the defocus distance on the weld bead penetration and weld width mainly. Experimental results also indicate that selecting appropriate process parameters can realize the butt-weld for the 0.85 mm thick Invar sheets in good welding modality.The composition of the welding-line has no obviously change, its tensile strength is equal to the base body and its hardness is appreciably lower than that of the base body.

A cable tension sensor with bypass excitation structure is proposed based on Villari effect principle of cable. Magnetic circuit is deeply analyzed by calculating magnetic conductance, the cable tension stress can be obtained by measuring inductive voltage in inductive loops. The sensor output may be determined by stress changing, inductive loop denseness, magnetic field and magnetic permeability. In addition,magnetization method, magnetic field intensity and sensor structure sizes are analyzed to determine the parameters of the cable sensors. Finally,a experimental system is developed, and the pulling tension experiment is carried out. The experimental result indicates that sensor sensitivity is up to 1.6 mV/kN, which shows that the magnetic circuit structure and parameter design of proposed sensor are feasible and suitable for cable tension measurement.

A kind of elliptical flexure hinge named deep-notch elliptical flexure hinge was proposed by taking notch width as the length of minor axis and notch deepth as the length of long axis for an ellipse. Based on the bending theory of beam with variable cross-section in material mechanics, the integral formula of the elliptical flexure hinge was deduced by inducing centrifugal angle as the integral variable. By defining an intermediate parameter in the integral formula, more concise analytical equations of compliance, rotation precision and maximum stress of elliptical flexure hinges were deduced to avoid time-consuming numerical integrals. A number of elliptical flexure hinges with different shapes were analyzed by using ANSYS finite element software. The analysis results are coincident with that of analytical formula well. Where, the maximum errors of angular displacement, maximum stress and rotation precision are less than 4%, 5% and 7%, respectively. These data indicate that the analytical equations are correct and also show that deep-notch elliptical hinge is a good choice for high precision transmission.

A localizing and tracking system based on magnetic sensor array is proposed to realize the real time localizing and tracking of the micro diagnosis and treatment devices in human Gastrointestinal tract (GI tract). This localizing system uses magnetic dipole model as mathematic model for magnetic field of permanent magnet and uses a 4×4 three axis magnetic sensor array to measure the magnetic field. A control circuit and an USB adaptor are used to transport the signal to the computer,then, the position and posture parameters of the magnetic object are computed by a nonlinear optimization method. Finally,a program is developed to display and save the localization result in real time. The real experimental results show that the average localization error is less than 7 mm, the average orientation error is less than 6°when the capsule is 150 mm away the sensor array plane. At the same time, this localizing system is able to track the capsule that moves among a simulated intestinal tract.Moreover, the localization error is not sensitive to glass, plastic, clothes, water and human body. It is conclusion that this localizing system is able to realize the real time localizing and tracking of the micro diagnosis and treatment devices in human GI tract.

In order to obtain a PZT piezoelectric function structure on the silicon substrate suitable for monolithic integrated chip, the research status quo of PZT piezoelectric thin (thick) films in the field of MEMS is analyzed, and then a new type of bi-cup PZT/Si film function structure is presented. This structure is optimized and designed with ANSYS finite element analysis software. The optimized structure parameters of the PZT and up-/sub-silicon-cup are DPZT∶D1∶D2=0.75∶1.1∶1;and the resonant frequency of the first-order modality is 13.2 kHz. The bi-cup PZT piezoelectric thick film on the silicon substrate is fabricated with the oxidation, lithography, anisotropism etching and screen-printing processing, whose thickness is 80 μm, and has piezoelectric driving function. This PZT piezoelectric thick film driving structure with bi-cup on the silicone substrate has a better compatibility with MEMS technique, and is suitable to be a driving component for MEMS micro-actuator.

For the demands of wear on-line monitoring for mechanical equipment, an on-line oil monitoring system based on microscopic image analysis is constructed.According to the characteristic of system light route, the image of wear debris is converted into gray image based on its color feature, and the wear debris object is extracted by subtracting the background image from the wear debris image. The classifier for two kinds of wear debris is designed based on the least square support vector machines, and the parameters of this model are optimized by Particle Swarm Optimization(PSO) algorithm. Based on this classifier, an integrative wear debris classifier is designed according to the wear debris recognition system. The performance and recognition precision of this system are tested by the ferrography technology. The result shows that the recognition precision of this system is as high as 95%,which can meet the demand of wear debris on-line monitoring.

The disequality of cells in the same battery team is one of the main factors influencing on the performance of satellite power system. This paper discussed the disequalization of a battery,and established a mathematical model.Based on the model, a new method to control battery equalization was present to counterpoise the energy of in the same battery team in series both for charging state and discharging state. The design ideas and design schemes of the hardware and the software of this equilizer were introduced in detail and the software flow charts of equalizer in charging state and discharging state were given. Finally an experiment was carried out to testifly that the method can decrease the inequality of in the same battery team to 13.42 %(charging) and 9.39%(discharging) as compared with those of battery systems without equalization.

An optical supporting structure is studied to improve its structure stiffness and reduce the structure mass to meet the requirement of space camera. A formula to calculate fundamental frequency is proposed using Rayleigh method, and the influences of the angles between trusses on fundamental frequency are analyzed. Then, the optimization method and process are studied. Based on the analysis above, an optical supporting structure with high stiffness for a pre-research space camera is designed. The analysis results indicate that the mass of designed optical supporting structure has been reduced by 22 kg and its fundamental frequency is 146 Hz,which shows that above analysis has a certain instructional significance for the design of optical supporting structures in high-resolution off-axis three-mirror reflective space cameras.

When a space camera is effected by the changes of gravity,temperature,pressure in space environment,the image detector can not be matched with the focal plane for mirror position change. In order to improve the imaging quality,a straight-line focusing machanism driven by step motor and glided on liner guide is designed based on analysis of several kind of focusing methods and discussion on focusing mechanism.This mechanism uses ball screws to translate the rotation to the line movement, and select an absolute encoder to detect the displacement. It has the advantages of light friction and simple transmission chain and can reduce the transmission error effectively. The precision of this mechanism is checked by open-loop control and closed-loop control. Moreover,the mean value and standard deviation are calculated by statistics analysis, and the residual error curve is plotted based on the test data. The measuring results show that the focusing machanism has a high precision, the error of closed-loop control is 0.002 5 mm, and the error of open-loop control is 0.001 5 mm, which are less than the precision of permission.

In order to reduce the effect of image motion on the image quality of space camera and to improve the resolution of camera, a compensation method of image motion to space camera is researched. The producing reason of drift angle and its adjusting principle are analyzed. Based on the characteristics of the camera, a fine compensation mechanism for motion image is designed. By taking a 80C31 as controller of drift angle,a step motor as actuating mechanism, and an absolute encoder as measuring component,the system realizes the closed-loop control of the drift angle position using transmissing form of sine agencies. Because the adjusting range of drift angle is between -4°～ +4°, this paper measures and obtains ten groups of practical data by absolute encoder in choosing -4°、-2°、0°、+2°、+4° as the expectation values of drift angle. Experimental results indicate that the drift adjusting mechanism can implement the accurate image motion compensation,and the accuracy of the drift angle can reach 2′,which can meet the accuracy requirement less than 3' for control system.

A stray light removing mechanism is designed for improving image contrast and image deblurring in a spaceborne camera. The spaceborne camera researched in this paper is based on on-axis Three Mirror Anastigmatics(TMA), and the stray light removing mechanism consists of inner and outer baffles, field stop, Lyot stop and blocking board. Cassegrain system is taken as a reference in the CAD modeling for design of inner and outer baffles and the random light Moente-Carlo simulation method is adopted to evaluate the effect of stray light removing. Moreover,the Bidirectional Scattering Distribution Function(BSDF) for describing the scatter character of surfaces, choice of key surfaces and the establishment of model are introduced and the stray light ratio and Point Source Transmittance(PST) curves are acquired by computer analysis.Results show that the stay light ratio is below 0.35% when stay light incidents at +10° meridional direction and the PST is at the level of 10-6 in 0~20°. These results can offer a good reference for the further optimization design.

A new method called Local and Global Preserving Embedding (LGPE) is proposed for manifold learning. This method assumes a global embedding function in low dimensional space, then incorporates the relative compactness information of the data distributions on the global geometry to reconstruct sample data. Finally, the global low dimensional submanifold is obtained by minimizing the cost function.The LGPE preserves the local and global structures of the data points simultaneously, and can obtain better dimensionality reduction on the sparse Swiss-roll dataset with noises (N=400, SNR=10 dB) and COIL-20 multi-poses dataset.When it is used in the AT&T face dateset,the recognition rate can be improved by 15% as compared with that of other local manifold methods under condition of embedding dimension lower than 40. The experimental results on both synthetic and real data sets show that proposed method is effectiveness and robustness for noise and sparse data.

The mathematical model of a stereo vision sensor is analyzed. Aiming at the application of vision measurement in large Field of View (FOV), a novel calibration method for stereo vision sensor with large FOV is proposed. A baseline rulur is randomly placed in FOV of the vision sensor for several times and the images of the feature points on the baseline rulur are acquired by the cameras. Then, based on the foundational matrix and the constraint of the known distance between two feature points on the target, the intrinsic parameters of cameras and the structure parameters of stereo vision sensor are simultaneously estimated by the method combining linear solution with nonlinear optimization. The experimental results show that the proposed calibration method is efficient and easy to operate . All parameters of stereo vision sensor can be determined without initial value. By the measurement precision of 0.06 mm in the measurement range of 6 000 mm×4 500 mm,it is proved that the proposed method is suitable for calibrating stereo vision sensor with large FOV on the spot.

In order to improve the instability of handwritten character pattern caused by different writing styles, a novel handwritten numeral character recognition approach based on manifold learning is proposed in this paper.Based on non-supervised manifold learning,a supervised information is induced to the algorithm to ensure the map from high dimension to low dimension to retain some manifold structures and also to seperate different kinds of manifolds. By proposed method,Supervised Locally Linear Embedding (SLLE) algorithm is used to reduce the dimensionality of input feature.Then, the reduced feature is classified by simple classifier.Finally,the proposed algorithm is tested on the characters in MINST character database. The experimental results demonstrate that the method can effectively improve the recognition rate to 93.27% and can provide a new approach to the research of handwritten numeral character recognition.

A real-time image restoration algorithm is proposed for the images blurred by rotational motion during aerial imaging.According to the rotational motion with space-variant blur and different motion rates for image,this paper analyzes how the rotational motion is generated and derives a mathematical model of rotational motion.Then,the coordinate conversion based on Bresenham algorithm is applied to transform the space-variant blur on circular arc into the space-invariant blur on pixel lines and 1D Wiener filter is used to restore the linear image motion transformed by coordinate conversion. The experimental results show that the proposed algorithm can restore a 1 024×1 024, 8 bit blurred image in 3.31 ms on the Graphics Processing Unit (GPU) platform, and the details in the rotational motion blurred image is pretty well restored and the PSNR of the restoration image can reach 28.76. The restoration quality and the effectiveness of the proposed algorithm reveals it is pratical in the aerial imaging system.

In order to optimize the design and calibration of the structural parameters of single stria section contour sensor, a mathematic model that the single stria section contour sensor measures the coordinates of the spatial points is established. The measurement errors of the single stria section contour sensor are analyzed based on the mathematic model and the relationship between the structural parameters of the sensor and the measurement errors is obtained. Based on the comprehensive consideration of the effect of the structural parameters on the measurement accuracy, the criterion for designing the structural parameters of the sensor is given.A practical example of the design and calibration of the structural parameters is introduced to demonstrate its implementation,and the measurement accuracy of the sensor is proved by an experiment. The results indicate that relative depth measurement error of real system is less than 0.3%, and the complex surface reconstructed by real system has a good visual effect.

In order to extract the line-features of extended object more quickly and more accurately, a line-edge extraction method is proposed based ergodic block. With proposed algorithm,a whole image is traversed by blocks,then the blocks with double peaks is obtained according to the property of gray-histogram.After processing those blocks by gray threshold and chain-edge extraction, the blocks without line characters are eliminated, the others with line characters are sorted by slope and intercept.Finally,the lines are got by computing the mean of similar slopes and intercepts. Experiment results show that proposed algorithm can extract line in subpixel precision and the precision of angle is higher than 0.01 rad. The complexity and computation of proposed algorithm is more better than that of Hough transform,which provides a way for hardware-implementation.

To improve the data output quality of multi-channel, high data rate and high transmission distance CCD system in a remote sensing camera, a kind of parallel fiber transmission system adaptive to multi-channel high-speed CCD image data is developed to apply to the space remote sensing camera. The fiber transmission channel with effective data rate of single port up to 1.28 Gb/s is designed and realized. Based on describing the thought of design and the structure of fiber transmission channel, the key technologies among the system, such as high-speed data interface and high-speed serial photoelectric interface are analyzed in detail. A TDICCD image data parallel transmission system with five channels based on fiber is developed, total effective data rate is up to 5.7 Gb/s. The experimental results indicate that the system works stable, high effetive,and no error code,which can meet the requirements of high-speed and far distance in the transmission of multi-channel CCD image data.

A digital image generating method based on the CCD and optical parameters of sensor system was presented. According to the relation of photometry and radiometry, a model of star imaging gray value was produced to calculate the imaging intensity for target in focal plane. Then, the number and magnitude of stars in the space were counted to obtain the statistic regularity of stars in the field of view by star map software SKYMAP. Finally, the digital image generating method was set up acoording to the statistic regularity of stars in the space and star image characteristic on CCD focal plane.The digital image generating method was simulated based on CCD and optical system parameter, results show that the method can generate a star map with higher apparent magnitude (magnitude＞10) and can satisfy the requirements of space object detection and surveillance for star map.

To eliminate the image jittering from the vibration caused by the translation motion of a camera, a method of digital image stabilization based on morphological edge pattern matching is proposed. In this method, morphological edge matching is used to estimate the local movement vectors based on the 1 bit edge maximal correlation. Since correlation calculation is logical operator instead of pixel-wise gray arithmetic operators, the complexity of proposed method is greatly reduced. Moreover, in order to remove the undesired local motion vectors caused by several undesired conditions, the local motion vectors are analyzed to judge whether it is desired or not and to eliminate the effects of locate motion target,low contrast and texture pattern regions on the estimation of globe motion vectors.Finally, the reliable globe movement vector is obtained by weight median. Experiments show that this method is fast and efficient for image stabilization, and can be applied in the video vibrations without large rotations.

With the development of space technology, the data of space remote sensing image are expanded greatly. To meet the real-time processing need of satellite on-board JPEG2000 image compression system, the optimization and improvement for the compression algorithm, as well as the design of high speed hardware process system are investigated. The Discrete Wavelet Transform(DWT) algorithm based on post-scaling lifting and quantification algorithm of JPEG2000 are analysed. Then, by a new scaling-quantification factor, the two processes of scaling in DWT and quantification are combined into one process to remove the quantification process in JPEG2000. Finally, the fixed-point implementation in the quantification removing method is brought forth. Experimental results indicate that the time for implementation scaling and quantification has been decreased by 50% with the new method,and the real-time processing ability of the satellite on-board JPEG2000 image compression system can be improved greatly.