An Internal Model Control (IMC) method is proposed for electronic-hydraulic servo system in this paper.As an advanced control algorithm,the design idea of IMC is that the object model is connected to the actual object in parallel,and the controller approaches the dynamic inverse of the model.Considering the advantages of IMC such as being insensitive to the model error,the electronic-hydraulic servo system model is simplified by reducing orders.The controller based on the simplified model has only one adjustable parameter which is directly related to system dynamic performance and robustness.Therefore,the parameter is regulated conveniently,and the randomicity and complexity of the controller regulation are avoided.Simulation and experiment results show that the IMC method overcomes parameter perturbation and improves the tracking performance and robustness of system.Furthermore,the artillery tracking error is within ±0.36 mrad.

To reduce the linearization errors of the Conventional Extended Kalman Filter (EKF) algorithm and the Converted Measurement Kalman Filter (CMKF) algorithm,the Second-order Converted Measurement Kalman Filter (SCMKF) algorithm is proposed in 3-dimensional space.The mean and the covariance of the converted measurements errors in Cartesian coordinates are inferred by the means of second-order Taylor series expansion.A more accurate and faster Kalman filter algorithm with debiased converted measurements is presented.Simulation results indicate that the SCMKF algorithm has higher tracking accuracy and faster convergence rate than the CMKF,the EKF,and the unscented Kalman filter,and the computation process of the SCMKF is more efficient than that of Debiased Converted Measurement Kalman Filter (DCMKF).

A method of target detection in infrared image sequences is proposed based on the least absolute deviation background predication.Firstly,prediction model of the background signal based on the least absolute deviation criterion was built.Based on characters of the least absolute deviation estimation,the extreme value was extracted by the linear programming.Then,the estimated image was subtracted from the source image to get the residual image.Finally,the fast threshold selection algorithm based on the two-dimensional exponent entropy was used to segment the residual image.The experimental results were given and analyzed.They were compared with those of the method based on least squares estimation.The results show that this approach can precisely detect the small infrared target and the method is better than the method based on least squares estimation.

In order to achieve robust and efficient tracking of a battleplane with strong flexibility,a novel tracking method based on a new similarity measure of gradient features and adaptive sizing of target is proposed in this paper.The similarity measure of gradient features is based on Mean Absolute Difference (MAD) by the introduction of matching power template,which is obtained by the transformation of target template.Using the characteristic of local superiority of target gradient feature,the probability distributing of gradient transformation is used to estimate target size.And then adjusting the size of target template is performed efficiently according to the estimate result while tracking.The results of experiment over real-world sequences show that the proposed tracker is robust to the rapid changes in viewpoint,pose and scale by the battleplane while moving flexibly.

To track the moving target in the cluttering backgrounds is one of the most important research fields in computer vision.A new tracking method based on the significance weighted least square image matching was proposed.According to the variance between the target and the background,the new method set the different weights to every parts of the target and tracks the target by the Significance Weighted Least Square Image Matching Method (SWLSIM).We compared our new method with the three classical tracking methods.Experimental results show that the new method can track moving targets in the cluttering background effectively and accurately.However,the tracking results of the other three methods are unacceptable.And the new method is satisfied for real-time tracking processing.

For the space-time registration errors of radar and IR sensor,a registration model for space and time biases is given and a registration algorithm for on-line biases estimation is presented.By incorporating sensor misalignments and target states into an augmented dynamic model,the dynamic equation of the augmented state and the augmented measurement equation are constructed.Then through the analysis for nonlinearity of the augmented measurement equation,a method of two stage filtering of Unscented Kalman Filter (UKF) and KF is proposed to estimate target states and register radar and IR sensor simultaneously.Furthermore,by comparison with the method of only using UKF,the proposed method has higher accuracy of estimation and less computation.The simulation results show the availability and practicability of the proposed algorithm.

It is a difficult problem of pattern recognition based on clustering to solve the distance between a point in l-dimension space and a quadric surface.The latter has an important application in spacecraft laser-warning.The paper shows that the algebra equation stated above exits single root if the eigenvalues of A are all equal,and gives a recursion algorithm searching for the root.The upper or down boundaries of a Lagrangian multiplier are given if the eigenvalues of A take the same sign,which provides possibility and convenience for dichotomy solving Lagrangian multiplier.Resolving the problem will provide a new method for laser detecting and laser-warning.

Wave-front reconstruction is the core unit of wave-front signal processing of adaptive optics.Based on the numbers of calculations and simple arithmetic of wave-front reconstruction,the discussion for the parallel arithmetic of wave-front reconstruction,sub-aperture row-coefficient translation and matrix cycle-division of reconstruction matrix is proposed in this article.On the basis of above,a real-time wave-front reconstruction on a control flow systolic array is built.In view of a 127-element adaptive optics system,experiments for the control flow systolic array and the traditional systolic array were carried out,and the experimental results show that the systolic array is slightly less than the traditional systolic array on real-time trait,but the resources occupied is much less than that of the traditional systolic array.

Space-based Visible (SBV) was the typical space-based optical sensor which was used for space surveillance successfully.Via a visible sensor installed on a satellite to detect the sunlight radiation reflected by space objects,SBV could achieve the goals of detecting,capturing and tracking these objects.At the same time,abundant observation data of various space objects were collected.In this paper,the characters and performance of SBV were introduced firstly.Then,taking the SBV as the prototype system,the information flow and technology characters in optical space-based space surveillance were outlined and analyzed,and the corresponding key technologies were abstracted,including target and star detection,Resident Space Object (RSO) catalog maintenance and update,analysis of objects brightness characters.The technology choke point and future development directions were discussed finally.

A novel method of Non-dispersion Infrared (NDIR) based on correlative technology for measuring CH4 trace gas and its application in the field measurement are described in the paper.Because of low concentration and weak absorption,atmosphere methane detection with Non-dispersion Infrared (NDIR) system faces the problem of high noise interference.To improve detection ability,Gas Filter Correlation (GFC) technique is put forward to reduce background noise.Correlative technology plays an important role in detection of feeble signal.The method is based on the gas characteristic absorption in infrared region and the correlation operation principle.Feeble spectrum signal is modulated and detected by combining advanced Gas Filter Correlation (GFC) technique and correlation technique.Methane is detected in real time and fast.Experiment results show that the system detection precision can reach 10ppb.

Visible light transmittance of coated glass is one of the important indexes of optical properties.A novel photoelectric integral measurement method for coated glass light transmittance is proposed.Compared with traditional photoelectric integral measurement method,the measurement method matches not only experiment illuminant to CIE standard illuminant,but photodetector spectral sensitivity to CIE human eye sensitivity.The basic formulas to calculate spectral transmittance for designing color filter are given,and the correcting color filter is optimized using evaluation function of minimizing relative error.Some satisfactory results are obtained.Experiments show that the measurement error is ±1%,which is less than ±2% GB request.The measurement instrument designed using the method has some advantages,such as simple structure,high speed,low cost,etc.It may supply references to rapid quality control and online detection.

On the basis of analyzing complete camera distortion model,a new calibration algorithm is proposed.The new algorithm consists of three steps.In the first step,the calibration parameters are estimated using a close-form solution based on a distortion-free camera model.In the second step,estimating the set of distortion parameters with the other parameters fixed.In the third step,the parameters estimated in the first step are improved iteratively through a nonlinear optimization,and makes all the parameters globalization.Very good results are obtained with real data calibration.The results show the precision of the new proposed method is up to 0.0367 mm,which meets the requirements of high precision 3D measurement and other application.

Aiming at disadvantages of all kinds of current ranging methods,a new ranging method based on similarity principle was put forward by using two cameras.We introduced principle of this method,and deduced ranging formula.But the formula can’t be applied to engineering practice directly.In order to overcome the difficulty,we proposed an experimental method which could realize the engineering value of this new ranging method.In order to verify the practicability of this method,two grope of experiments have been done.The results show that the ranging method can measure distance accurately at short or middle distance.Compared with other ranging methods,this method is low-cost and suitable for the visual ranging system of mobile robots.Moreover,it can make the mobile robot get ability of telescope and wide-angle lens at the same time.

Aiming to read data automatically with digital level,the localization algorithm of the double sinusoid barcode staff was studied according to coding principle that different widths of bar code represent different altitudes of staff,which was used in digital leveling to accomplish automatic staff reading.Based on the relationship of the phase difference and sampling value combination of two sinusoidal signals with position on ruler,two automatic measurements were researched and their numeration formulae were given.One was decoded by Fourier analysis and linear conversion,and the other was decoded by sinusoidal image breadth combination matching and geometrical positioning.Appropriate parameters of sinusoidal signs were designed according to respective decoding means.The results of the simulative barcode reading experiments indicate the feasibility of the algorithms.

The computation complexity of Lapped Biorthogonal Transform (LBT) is high and the precision of its integer algorithm is lower.To overcome these shortcuts,a fixed point algorithm of LBT is proposed.Based on the approximation of numbers in Q format,fixed point is used to approximate the value of trigonometric function of the rotation matrix in the algorithm.And the parallel computing of the algorithm in TMS320C6000 DSP is studied by using assignment partition,complex instruction,software pipeline and optimization method to improve the processing speed.The algorithm has been used in prototype data compression equipment of a spy satellite.Image coding experiments show that coding performance of the algorithm is very competitive with CDF9/7 wavelet and the complexity of the algorithm is far low.

Low-complexity and high-compression video coding play an important role in high resolution video surveillance.In indoor video,moving objects are Region of Interest (ROI) under surveillance.Based on detecting and restoring ROI,a novel video coding scheme for indoor high-resolution video surveillance is proposed in this paper.Both base-layer and ROI-layer bit streams are obtained and transmitted based on object extraction.At the decoder,low resolution video image can be decoded from the base-layer,while ROI-layer is used to restore region of interest with high resolution.Experimental results show that the proposed scheme reduces coding complexity and improves coding efficiency.

A method,called directed gradient power fractal feature extraction,was presented for object representation.In the method,we proposed the definitions and algorithms of directed gradient power,directed gradient total power,total directed gradient power,directed gradient power fractal feature,and complex number fractal feature.The properties of directed gradient power fractal features were analyzed and concluded.Two-dimensional fractal features were extracted for SAR image object detection.Theoretical analysis and experimental results show that by using the method,the objects of different shapes and scales can be detected efficiently.The method has the advantages of simple programming,faster computation,etc.

Base on investigate a great many of relevant arithmetic home and abroad.Therefore this paper classify image mosaic arithmetic and give a new arithmetic of image mosaic base on pixels grayscale curve base on pixel.This method do a cursory mosaic base on matching points at first,and then do a new arithmetic of image mosaic base on pixels grayscale curve.It is not only keep veracity similar to essence arithmetic,but also accelerate.In the end this arithmetic is tested in the experiment.The outcome indicate that in the condition that precision is similar to essence arithmetic,the speed increase to about160% than essence arithmetic when the testing pictures are different.

An adaptive calculation method for realizing the threshold segmentation algorithm is put forward by using threshold segmentation algorithm based on transitional region as a theoretical foundation.This kind of non-interactive method can not only avoid the manual intervention of choosing threshold value in image segmentation but also gain ideal digital edge of Tooth Contact Zone (TCZ).During the pretreatment process for changing true color image of TCZ to gray image,the adaptive calculation for trichroism coefficients is realized by using the statistics as a foundation and focusing on concrete true color image.The effect of the pretreatment provides entire image information for the subsequent image segmentation.Experiments show that the method developed in the paper for extracting digital edge of TCZ of bevel gear features is applicable.

A novel automatic ear anatomy feature extraction method was proposed.Firstly,the side face image including an ear was preprocessed to eliminate the noise and spur.Secondly,we regarded the 2D gray image as a surface because the characteristics of surface are that gray level changes in edge area is sharp and the curvature is larger than that of the smooth area.In accordance with the characteristics,the maximum and minimum principal curvatures of surface were calculated,and then,automatic threshold was used to segment curvature images.Finally,two thresholded images were combined to get the feature edge image.The experiments demonstrate that extracted feature edge curve is smooth and connected.Experiment shows that our method is robust to illumination intensity and edge curve can meet the requirements of next process.

In order to analyze working principle and stability error of double longitudinal-mode He-Ne laser with stabilization frequency,a requirement is put forward,which is that an intracavity laser with cavity of 100～300 mm length can produce double longitudinal-mode light.According to laser working process,stabilization frequency includes mode hopping,transition stage and stabilized mode.With laser principle and thermodynamic theory,mechanism of stabilization frequency of the laser is interpreted on the basis of output voltages of the two photodiodes.Beat frequency test shows that frequency stability of He-Ne laser is 5×10-10.Based on error analysis,the frequency stability of He-Ne laser is only influenced by change of resonant cavity length.Furthermore,increasing length of the resonant cavity can lead to higher frequency stability.The laser with double longitudinal-mode stabilized frequency has a shorter locked time.

Aiming at the problem that there is no comprehensive model of coherent beam combining through superposed volume gratings,we establish a comprehensive model for two-beams combining and its analytic formula is derived using coupled wave theory.It shows that the intensity of combined beam is influenced by many factors such as the amplitudes and phases of the two incident beams,the phase shift between gratings,the amplitude of modulated refraction index,the thickness of gratings and the detuning et al.The effect of each optical parameter on combined beam is numerically analyzed in detail.And the sufficient and necessary conditions to realize efficient beam combination were summarized,such as the thickness of gratings must be optimized,the index modulation of each grating must be equal and the Bragg mismatching must be small enough et al.

The third harmonic generation of the super intense ultrashort pulse with a single β-barium borate (BBO) crystal was studied theoretically and experimentally.The effects of the intensity of the fundamental field,the thickness,the detuned angle and the azimuth angle of the crystal on the conversion efficiency and the pulse shape were analyzed.The parameters of the experiment were optimized.The corresponding experiment was performed for the ultrashort pulse with center wavelength of 810 nm,the pulse width (FWHM) of 100 fs,the bandwidth of 25 nm and the pulse energy of about 6 mJ by using a single BBO crystal.The conversion efficiency of the third harmonic generation can approach 0.8%.Analyzing the experimental results,we proposed the methods for improving the conversion efficiency of the third harmonic generation with a single BBO crystal.

Aiming at having large error in the linear nonuniformity correction method of Infrared Focal Plane Arrays (IRFPA) and complexity algorithm of nonlinear nonuniformity correction which limits its application in actual project,a new nonlinear nonuniformity correction method,which is easy to realize real-time correction and has higher precision,is proposed.Based on principle of new method,the mathematical model is established.At last,experimental results are given out.The results show that it has higher correction precision,less parameters needed to store and easy to implement real-time correction by hardware circuit.

To overcome the difficulty of optimizing the amplitude of millimeter-wave generated by optical frequency multiplication,a novel millimeter-wave generation technique using dual-electrode Mach-Zehnder (MZ) modulator is proposed.Light-wave injected into dual-electrode MZ modulator is split into two optical carriers,which are phase modulated by two microwave signals respectively and then combined together.Then the light-wave has effect of FM-IM conversion and is photodetected to generate electrical signal consisting of harmonics of the microwave signal.The desired millimeter-wave signal is included in the harmonics and can be maximized just by adjusting phase modulation index.Remote local-oscillator delivery for uplink utilizing subcarrier modulation and the nonlinearity of MZ intensity modulator is also proposed.Data transmission is successfully implemented on downlink and uplink.It shows that millimeter-wave Radio over Fiber (ROF) system based on the proposed generation technique and the remote local-oscillator delivery is feasible.

Since the all-optical wavelength conversion at present is realized by using Semiconductor Optical Amplifier (SOA),the all-optical wavelength conversion based on Erbium-doped Fiber Amplifier and Nonlinear Dispersion Fiber (EDFA+NDF) is put forward in the paper.When the pump and signal lights input the NDF,Four Wave Mixing (FWM) in NDF can produce new light in the nonlinear dispersion fiber.By using the OptiSystem software,the simulation of the all-optical wavelength conversion is achieved.And then,the all-optical wavelength conversion is realized through the EDFA and the G655.The conversion efficiency of -16.7dB is acquired.The experiment result indicates that sharp wavelength conversion picture is obtained both in theoretical simulation and lab.

In order to provide a straight,natural and bi-directional interaction method for the Augmented Reality (AR) system,a prototype system is proposed by integrating systems of eye gaze tracking and AR.The system can track user’s fixation region of interest in real time by eye gaze tracking,and reflect the features of AR system.In accordance with the basic demands of integration,combined with the virtual camera model of optical see-through Helmet Mounted Display (HMD),the principle of the system and the method of optical design and registration are introduced in detail.The experimental result shows that the integrated prototype system is effective in human computer interaction,which establishes a sound foundation for the applications of eye gaze tracking and AR.

In order to correct the barrel distortion of wide-angle lens,a new method based on spline function was proposed.Firstly,according to the physical principle of barrel distortion,the wide-angle lens was relatively calibrated by template based on concentric circles.Secondly,depending on the slickness of spline function and the continuity along radius’ direction of barrel distortion,the cubic spline function was used to approximate the rules of barrel distortion.Lastly,the distortion’s rules fitted by cubic spline function were used to restore the distortion image captured by the same visual system.In experiment,the new method could restore the image with barrel distortion rapidly and get well performance.So,for vision system with fixed focal length,the new method can correct the barrel distortion rapidly and efficiently,which needn’t take into account the parameters of vision system.