An Acceleration Feedback Control (AFC) for multi-closed loops control system is proposed to improve tracking performance in tracking control system, which can provide an inner loop with fast and wide bandwidth for the control system that is better than only velocity control loop. A feed forward controller based on acceleration feedback is important to track high velocity and acceleration targets because the control mode with AFC can suppress variations of system parameters. It is easy to implement the control algorithm in the high performance float digital signal processor. Experiments show that the proposed method is effective. The bandwidth of disturbance suppression is improved from 15 Hz to 30 Hz, and the suppression ability reaches -30dB under the frequency of 10 Hz.

A novel method for automatic measurement of target’s 3D gesture was proposed here to solve the problem of measurement of target’s 3D gesture in target trials of shooting range. Firstly, the model of imaging was analyzed. Secondly, geometric active contour and improved Hough transform were used for extracting axes of symmetry target precisely. Then, the 3D gesture parameters were corrected automatically by making use of the collinear equation in imaging model, which eliminated the duality of 3D gesture parameters and gained the actual results. Experiment results show that the method can extract the axes of axisymmetric target effectively in high precision and get correct 3D gesture parameters.

In order to meet the needs of moving object location in different video sequences, a novel moving object location method was proposed. Subtractive clustering algorithm was used for object location in video sequences. The theory of subtractive clustering algorithm was analyzed. Equations of subtractive clustering algorithm, software flowchart and realization steps of proposed method were presented. Different location results for different video sequences were studied. Subtractive clustering location algorithm was also compared with region growing location method. The location order, time consuming and its robustness against blob noises of the proposed method were discussed. Experiment results show that the proposed algorithm is fit for video sequences whose binary images have big noise blobs and bad spatial connectivity.

A new system noise covariance modification algorithm is proposed in order to avoid the problem of degraded performance of the filter due to the incorrect statistics of the system noise. Combined with the Extended Kalman Filter (EFK)、Unscented Kalman Filter (UKF) and Divided Difference Filter (DDF), adaptive nonlinear Kalman filters are developed. The algorithm is applied in nonlinear measurement electro-optical tracking system and the performances of the adaptive nonlinear Kalman filter is compared with the basic nonlinear Kalman filters. The Matlab simulation results show that the filter can modify system noise covariance in real time, efficiently avoid the above problem and the performance outperforms the basic nonlinear Kalman filters.

To realize fast and robust digital image stabilization, a new optimum motion estimation method is proposed. Firstly, an approach for block judgment is presented. Before calculating motion vectors, gradient information is analyzed to judge the blocks, and only useful reference blocks that are indispensable for accurate motion estimation are selected with its reliability and consistency on motion estimation, by which the number of blocks for subsequent calculation is reduced. Secondly, in the block matching, an improved Sequence Similarity Detection Algorithm (SSDA) is used to reduce computing cost. Finally, the LMedS method is used to eliminate wrong motion vectors and the parameters of whole motion model are obtained by least squares algorithm. Experimental results show that the method gets 1/4 pixel accuracy, and one process time is less than 7 ms.

The principle of the adaptive optics system for all-path aberration correction was analyzed in detail according to the theory of wavefront propagating in optics medium. While the wavefront sensor was calibrated by conventional method, the residual error of the adaptive optics system was deduced, and the shortcoming was indicated from the points of the residual error and the operation feasibility. Based on the characteristics of the adaptive optics system for all-path aberration correction, two kinds of new calibration methods for common mode Hartmann wavefront sensor were proposed, and the residual error of the system was analyzed in detail while the wavefront sensor was calibrated by new methods. The result indicates that the residual error of the system calibrated by the two new calibration methods is related to one of the residual errors of the system calibrated by the conventional calibration method.

On the basis of the traditional magnitude arithmetic of space targets, the computing models of star magnitude of space targets like prism were researched. First, a more detailed arithmetic was deduced to compute effective optical reflective area of space targets like prism by adopting a facet model. For the column could be approximated by prism, the effective optical reflective area of space targets like column was obtained directly. In order to simplify the solving of the model, Satellite Tool Kit (STK) was used to provide relationship among sun、space object and observation station in the magnitude models, and the process of computation was given. The proposed new computing model was verified with other model and some analyses, and the results show that the models are easily solved with STK and the new column model has higher resolution than the experience model.

A model for compensation of radial shift was established to further improve fabrication precision during polar laser direct writing. At first, exposure dose distribution model of polar laser direct writing was established , and the cause of radial shift was analyzed. And then, the linewidth broadening model and radial shift model were established corresponding to object lens with numerical aperture 0.5, and the coupling relationship between linewidth broadening and radial shift was analyzed. Based on the above mentioned, compensation models of linewidth broadening and radial shift were established. Simulation results indicate that the radial drift value was about 3 times linewidth broadening value. The compensation of radial shift reduces the system error of polar laser direct writing and improves the fabrication precision.

A model for polarized light propagating in the biological tissue was proposed and the propagating of polarized light was simulated with the model. The distribution of polarization in biological tissue and birefringence model of biological tissue were reconstructed with different input polarized light. The relationship between Degree of Polarization (DOP) and order of scatters shows that depolarization of polarized light propagating in tissue weakens the interference between two arms of interferometer in PS-OCT and brings down image quality. Analysis of application of Stokes vector, Muller matrix in fiber-based PS-OCT system shows that Muller matrix is not affected by polarization of input light. Compared with Stokes vector, Muller matrix is more suitable for studying fiber-based PS-OCT.

Carrier phase delay resulted in output signal fading in the Phase Generated Carrier (PGC) demodulation, and in critical conditions, the output signal vanished. A carrier phase advance technique was developed to solve this problem. The digital carrier waveform was generated using cosine waveform Look-up-table (LUT) in the digital PGC demodulation system. An address offset was introduced at the LUT input port and a phase-advanced carrier was generated. The phase advance completely compensated the constant delay in the demodulation system. The address offset could be obtained from both analysis and measurement of the demodulation system. Experiment result verifies that the carrier phase advance technique is effective.

This paper points out the obvious wrong segmentation in the existing two-dimensional histogram vertical segmentation method. A new method of two-dimensional histogram oblique segmentation is proposed, in which the histogram is divided into inner, edge and noise parts by four oblique lines paralleled with the main diagonal. The image is segmented according to the sum of point grayscale and neighborhood average grayscale. The method could be used in almost all the two-dimensional histogram threshold algorithms. The formula of the Tsallis–Havrda–Charvát entropy thresholding based on the two-dimensional histogram oblique segmentation and its fast recurring algorithm are deduced. The segmented images and processing time are given. Compared with the original two-dimensional Tsallis–Havrda–Charvát entropy algorithm based on two-dimensional histogram vertical segmentation, the proposed algorithm makes the inner part uniform and the edge accurate in the segmented image, and has better tolerance capability to noise. The processing time is reduced by 5 orders of magnitude.

In image measurement system, accurate focusing is critical for high precision measurement. For image focusing, a focusing evaluation function is introduced to verify whether the image is in focus or not. When the edge of a work does not lie in a plane perpendicular to the vertical direction, it is difficult to focus correctly with conventional focusing evaluation functions. So a focusing evaluation method is proposed based on the edge tracking. Firstly, the edge of interest is extracted with edge tracking. For each edge point, sums of gradient square within a certain area along the edge direction are calculated and the focusing evaluation function is the sum of the sums of gradient square. When the focusing evaluation function reaches its maximum, the edge is in focus. Experimental results show that, when the edge of a work does not lie in a plane perpendicular to the vertical direction, it can be focused correctly with the proposed method.

To satisfy the requirement of preserving details in equipment geometric model simplification, a novel 3D geometric model simplification method based on preserved vision feature was put forward. In the algorithm, a new error metric rule based on the geometric importance (including the length of the edge and the vertical curvature) and change error of a half-edge was proposed to determine the order of edge collapse in 3D geometric model simplification. And then, a vertex micro-adjusting algorithm was adopted to modify the local details of the 3D geometric model. Experimental results show that the proposed simplification method is practical and convenient, and can effectively preserve model details and obviously improve model distortion in 3D geometric model simplification.

Moving cast shadow usually is misclassified as foreground. In order to solve the problem, a novel method is proposed to detect moving cast shadow based on energy minimization. Firstly, the stationary background and moving objects could be obtained by the traditional pixel-wise method. Then, the color and texture invariance of cast shadow is employed with the temporal-spatial coherence to construct the energy function through Gibbs energy. Finally, moving cast shadow and foreground are segmented accurately through minimizing the energy function. The method is tested with some videos including moving shadow and the results show that the proposed method excels previous methods both in indoor and outdoor scenes.

To meet the potential demand of large scale mapping, a multiview photogrammetry approach was presented for low altitude digital images. A multiview matching was carried out under a restricted condition derived from multiview geometry in computer vision. Multiple image spatial forward intersection method was adopted to deal with multi-baseline stereo from low altitude digital images with large overlapping. Experiments on spatial forward intersection with stereo pairs of different overlapping and multiple images indicate that multiview photogrammetry is able to improve the photogrammetric accuracy of low altitude digital images.

The nonsubsampled Contourlet transform is built upon nonsubsampled pyramids and nonsubsampled directional filter banks and provides a shift-invariant directional multi-resolution image representation. The nonsubsampled contourlet transform and its coefficients characteristics are introdued. A method for image fusion based on nonsubsampled contourlet transform is presented, which could extract the edge and detail coefficients and fuse the information in high and low resolutions efficiently. The multi-direction and multi-resolution method fully extracting the fusion information, improves the image definition and information quantity by using different fusion methods in different frequency and directions, which overcomes the directional limitation of Wavelet. By simulation experiment and comparison with other arithmetic, the validity and superiority of the method are proved.

A profilometer was developed for micro-nano structure profile testing on the basis of interferometry and micro vision system. A Linik interferometer was adopted in this profilometer, in which the scanner and the phase shifter were integrated through scanning the reference mirror. The five-step phase-shift algorithm was selected for Phase Shift Interferometry (PSI) mode and the Squared-Envelope function estimation by Sampling Theory (SEST) algorithm was selected for Vertical Scanning Interferometry (VSI) mode. The experiments that used a standard multi-indents structure and a standard step structure to test PSI mode and VSI mode respectively verified that the profilometer could measure the profile of the micro-nano structure quickly and accurately, and be applied in the measurement of micro-electronics and micro electronic mechanic system.

A method of feature extraction which is composed of invariable moment functions and Principal Component Analysis (PCA) is presented in order to recognize and classify the surface defects of strips. First, a 22-dimensional eigenvector which was invariable was extracted from images when the image was translated, scaled and rotated. And then, in order to improve the efficiency of classification, PCA was applied to reduce the dimension of the eigenvector. As a result, the 4-dimensional eigenvector was obtained. Finally, using these eigenvectors as input, weights and thresholds of the BP neural network were trained for the purpose of defect classification. Experimental results show that the average efficiency of the correct identification can reach 85%, and it’s fit for the application for detection of surface defects of strips.

A dual-camera vision 3D coordinate measuring system using a probe is proposed to solve the problem of the low measuring repeatability about similar system with one camera. It mainly consists of a specially designed optical bar, two orthogonally placed CCD cameras and a computer. In the dissertation, the dual-camera model is established based on the principle of redundancy and the single camera system model. And the analytical equations of the measured point’s coordinates are derived. A test device is built using the easiest three-point model light bar. The experimental result shows that the system with linear points and dual cameras has achieved a measuring stability precision of 0.2 mm and measuring precision of ±0.15 mm in the direction of every axis for a distance of 1 500 mm away from dual-camera system.

The parabolic mirrors with high Numerical Aperture (NA) are used as efficient light collectors and its focus with the slight deviations of the incident beam from the optical axis or from parallelism give rise to huge aberrations. A device is designed to measure and collimate the angle error between the optical axis and the structure axis of high NA parabolic mirrors. The imaging principle of the parabolic mirrors is used to analyze the angle error of the parabolic mirrors. The experimental result shows that we can get the angle distinguishing precision 0.06′ and the maximum measurable region 3.65′. The result can satisfy the application of the parabolic mirrors used as the light collector.

The traditional pulsed laser ranging with direct counting time measurement has low precision. In order to improve time resolution and precision of pulsed laser ranging, the interpolation method of Time-to-amplitude Conversion (TAC) is introduced. Several key points of TAC circuit design are provided here. On the pulsed laser ranging experiment platform, the performances of the designed TAC circuits are tested. The experiment results show that TAC interpolation method improves time resolution of pulsed laser ranging from 10 ns to 20 ps and ranging precision from 1.5 m to 5 cm. At last, the error factors influencing the precision of pulsed laser ranging are analyzed.

Through theoretical analysis and simulation, the quantization errors of optical wavelet filters constructed by different wavelet bases were studied, and the effects of the exchange of the decomposition wavelets with reconstruction wavelets on quantization errors were analyzed. The experimental results show that the quantization errors of different wavelet bases are usually different and the wavelet base of the minimum quantization error in space and frequency domain is 5/3 wavelet after exchanging the decomposition and reconstruction filters. According to the least quantization error criterion of optical wavelet filter in space domain and frequency domain, an optimal wavelet was constructed based on lifting algorithm.

For the need of high-precision spectral radiance measurement, a holographic plane grating double monochromator with the double decked Ebert-Fastie structure is designed. This system consists of two spherical collimating mirrors, a plane mirror, a roof mirror, two holographic plane gratings, entrance slit, exit slit and mid slit. Wavelength region from 160nm to 400nm can be scanned. This double decked design has a compact structure, high spectral resolution and low stray level. Its spectral resolution is less than 0.15nm, and stray level is 10-6, which satisfies the need of high-precision spectral radiance measurement in ultraviolet wavelength region in space.

A novel method for ternary polarization-hologram storage was presented, and the model of ternary optical memory based on the photoinduced anisotropy of indolylfulgide/PMMA film was constructed to store 3-state line-polarized beam for ternary optical computer. Using He-Ne laser as recording and readout beam and coder/decoder of ternary optical computer as data input/output element, ternary digital optical signal was recorded on indolylfulgide/PMMA film by Fourier transform holographic method. The memory can store directly ternary information denoted by cross-polarized and non-light states, and may realize page addressing and parallel read-write. The experiment result of one-bit ternary polarization-hologram digital storage with indolylfulgide demonstrates the method is feasible.

The sol-gel silica films on K9 glass substrates were prepared with the dip method, and some of them were treated in saturated ammonia gas. Microcosmic surface morphology, chemical structure, refractive index and weak absorption of the films before and after ammonia treatment were respectively measured by Atomic Force Microscopy (AFM), infrared spectrophotometer, ellipsometer and transmission photo-thermal lens. The experimental results indicated that the porous ratio of the films decreased from 0.73 to 0.63 after ammonia treatment, and the weak absorption increased from 67.88×10-6 to 74.58×10-6. The Laser-induced Damage Threshold (LIDT) of silica films decreased from 18.0 J/cm2 to 16.9 J/cm2 after ammonia treatment. Considering the improved mechanical property of films with ammonia treatment, a trade-off is always needed between mechanical property and LIDT according to the requirement in practice.

To describe and analyze the system level noises of the thermal-mechanical optical readout uncooled infrared imaging system, a photoelectricity coupling model is presented. The photoelectricity coupling device in the model shows the noise isolation characteristic of the optical readout IR imaging system. Based on the model, the noises are partitioned into two kinds: internal noises which are from the Focal Plane Array (FPA), and external noises which are from the optical readout system. The Noise-equivalent Temperature Difference (NETD) caused by internal noises is calculated to be 5.94 mK theoretically. Experiment shows that the NETD caused by external noises is 98 mK, which is close to the NETD caused by the whole noises of the system. The results indicate that the calculation of NETD caused by the internal noises is reasonable and the primary noise problem of the IR imaging system is from the optical readout system

The performance of an infrared imaging system is affected by spatial fixed-pattern noise badly due to the different photoresponse of each detector in a focal plane array for the same irradiance. A novel nonuniformity correction algorithm based on Wiener filtering is developed, which first determines an output delay for being able to use the later information of an uncorrected image, and then by depending on the front and back information of a current frame, Wiener filter is adopted to estimate the current frame for several times. At last, take the average value of the estimations as the final correction result. The performance of the proposed algorithm is demonstrated with a real infrared image, and its advantage is that it can employ sufficient information to eliminate the fixed-pattern noise.

Three kinds of CMOS compatible photosensors, P+/Nwell photodetector, Nwell/Psub photodetector and N+/Psub photodetector, were designed by using SMIC 0.18μm standard CMOS technology. Some critical parameters, such as responsivity, dark current and maximal response wavelength were analyzed based on the mathematic model established. The influences of some technology parameters, such as doping concentration, junction depth, were pointed out as well. The experiment results indicate that P+/Nwell photodetector can reach a maximal sensitivity of 0.08 A/W at 460nm with 55nA/cm2 dark current, Nwell/Psub photodetector has a maximal sensitivity of 0.35A/W at 580nm with 64nA/cm2 dark current and N+/Psub photodetector attains a maximal sensitivity of 0.29 A/W at 580nm with 600nA/cm2 dark current. The test results show that the photodetectors designed agree with theoretical analysis basically and have prominent performance on sensitivity and response wavelength.