The “cat-eye” effect of optical window can be effectively weakened by circular holes array for its modulation, meanwhile the imaging quality of “cat-eye” target can be affected to some degree. Considering that Modulation Transfer Function (MTF) could comprehensively evaluate the imaging quality of photoelectric equipment, this paper put forward the testing models of MTF for telescope and Low-light-level night vision. The digital test of MTF of these two typical “cat-eye” targets was realized by a digital test system of MTF for optical instrument. The test results show that the circular holes array can lead to fall of “cat-eye” targets’ MTF at full spatial frequency, and if the aperture is smaller and the holesarray is more compact, the effect on MTF of “cat-eye” target may be less. This conclusion can provide reliable experimental data to reduce the effect on imaging quality when designing structure of circular holes array.

To improve the simulation precision of Line of Sight of Laser Target Seeker (LSLTS), the source of the errors affecting the precision is analyzed according to the process of the simulation. The precision model of angel of LSLTS is constituted based on errors composition theory, and the evaluation criteria of the whole screen is established. On the basis of the criteria, the equipments layout of the laboratory is designed by applying the Genetic Algorithm (GA). The distribution of angel errors in the screen is achieved, which gives support to the design of the LSLTS simulation.

Multi-target tracking methods based on random finite set theory can avoid problems appeared in the progress of data association. It can also deal with multi-target tracking problems in the complex environment with target number unknown and target number varying with time. In this work, multi-target tracking methods based on data association and random finite set are analyzed. Characters and advantages of tracking methods based on random finite set are described. Some key problems, like target state extraction, track-to-track association, more accurate filtering algorithms, and PHDF algorithms under complex situations, are summarized and reviewed, and then some research focuses in the following years are given.

To improve the post positional accuracy of aircraft, a processing method based on the tracking data fusion of multi radar photoelectric theodolites was presented. The math model of the tracking data of radar photoelectric theodolite was put forward, the spline restraint method was applied to the fusion processing of angle, range and velocity measurement data, and the joint model of positional parameters of aircraft and systematic errors of devices was established, whose redundancy was improved by using radar measurement data. The corresponding simulation results show that, compared with the fusion method of only angle measurement data, the estimation accuracy of positional parameters and system errors are improved significantly by the fusion method of angle, range and velocity measurement data.

In order to analyze the accuracy of three-dimensional coordinates transformation and guide the distribution of common points in measurement on site, a mathematical model of the relationship between the accuracy of three- dimensional coordinates transformation and common points was established using least square estimation of unequal accuracy theory. Analyzing the model, we could get the transformation accuracy at any point of the space and draw a conclusion that the mean square error and distribution of the common point are the main factors of coordinate transformation accuracy. Finally, the factors of coordinate transformation accuracy were analyzed using three-dimensional coordinate measurement data of common points. The experiment shows that transformation accuracy can be effectively improved by improving the measurement accuracy of the laser tracker and changing the distribution of the common point.

As for the problems that question that color’s effectiveness had been drastically influenced by the variety of illumination and pose and interference of background in image tracking which will lead to tracking failure, a face tracking algorithm based on particle filter is proposed, which uses color histogram and contour distributions as its clues. There are three characteristics in this algorithm. Firstly, based on the basic framework of particle filter, a new method for describing the outline of face by histogram was proposed. It could reduce the influence when illumination, face rotation and partial face occlusions happened during tracking. Meanwhile, it could capture the lost target caused by large parts of occlusions quickly. At the same time, the feature point number in our method was adaptable which had improved the tracking efficiency. Secondly, for reducing the influence of the background, a method for suppressing the interference of similar background was proposed. Thirdly, the template was updated to improve the robustness of tracker during tracking. Experimental results show that our proposed algorithm could perform better than conventional tracker.

Measurement precision and imaging quality of photoelectric measurement system were affected by kinematic image resolution of photoelectric measurement system. Kinematic resolution of imaging system of photoelectric measurement system is often limited in performance by photoelectric measurement system tracking precision. Analysis and experiment prove that working photoelectric measurement system has lower image resolution than static photoelectric measurement system. Objects with high movement frequency and low movement frequency opposed to exposures have different effect model on kinematic resolution of photoelectric measurement system. Static resolution and control system needs optimal design for some kinematic resolution.

Super-resolution reconstruction is an effective method to solve the problem that the face image resolution is too low to recognition in video, but the non-rigid change of face distortion and expression changes greatly affect the accuracy of registration and reconstruction. To solve these problems, a method of multi-level model Free Form Deformation (FFD) elastic registration algorithm based on B spline is proposed. It first uses low-resolution FFD grid for global registration, then divides the global registration image into several sub-images pairs and calculates the correlation coefficient of each pairs. At the same time, local registration with high-resolution FFD grid is done when the correlation coefficient of sub-image is small. In the process of optimization, adaptive step length gradient descent algorithm based on chaotic variables is used to improve optimization efficiency. After registration, the algorithm of POCS is used to reconstruct high resolution image through several low resolution images to realize recognition. Experimental results from standard video database and self-built video database show that this method can register and reconstruct accurately in the condition of great facial distortion and expression variation, and the face recognition accuracy is also improved.

A pattern recognition named phase-only matched filtering is presented to realize high precision alignment in projecting lithography. The method achieves the alignment with the characteristic direction of templates and silicon substrate using the coherent peak rotating sensitive feature in the filter, meanwhile the relative translation distance of templates and silicon substrate is obtained in recognition algorithms. Such an alignment system is assembled in our mature projection lithography machine, and we got the experimental data to prove the higher accuracy and efficiency compared with conventional algorithms.

A palmprint recognition method is presented based on sparse representation, which takes advantage of two-dimensional principal component analysis (2D-PCA) of its better data compression property and faster feature extraction speed to generate the palmprint feature image. The 2D-PCA method not only overcomes the shortage of complex calculation by PCA method due to its higher data dimension, but also retains the data structure of original image to obtain better feature. In order to facilitate sparse representation, we take the PCA method to extract features of palmprint feature image to obtain the training samples. In this case, the training samples still retain the data structure of original image and improve the recognition rate compared to simple PCA method. We take the training samples to construct a redundant dictionary, and express the testing samples as a linear combination of dictionary atoms by sparse representation theory. Then, the classification is achieved according to the sparsity of representation coefficient and sparse concentration. Due to the sparsity of representation coefficient, this method reduces the time and space complexity. Andthe experiments show that the recognition rate of this method is obviously higher than traditional method for the Hong Kong Polytechnic University MSpalmprints Database.

In order to make Enhanced Fisher Disriminant Criterion (EFDC) avoid of impairing the discriminating information caused by PCA dimension reduction, based on EFDC, a method of two -dimensional Intra-class Diversity Preserving (2D-IDP) for face recognition is proposed. In this method, we built a more robust discriminate criterion which can make the data points of different class as distant as possible and simultaneously preserve the intra-class compactness and variation, and thus avoid the over-fitting problem. At the same time, we redefined parameter t in the neighbor graph of EFDC which made it change adaptively according to different samples. Thus it avoided the problem of performance degradation caused by inappropriate choice of t. Experiments on YALE and AR face database verify the effectiveness of the proposed method.

In an optical fiber perimeter security system, the longer distance and background environment complex border detecting, the system based on optical fiber cable line implements the event recognition. In order to distinguish the various induced cable vibration excitations, according to the intrusion signal and environment caused by vibration signal in wavelet scale above difference amplitude distribution characteristics, by using wavelet multi-resolution analysis theory, the scale variance component of feature vectors is constructed. According to the variance feature vector, scale- variance signal method for identifying different kinds of vibration signals is proposed. In the experiment system, total cable length is 56 km, the power of the light source of 300 μW, working at a wavelength of 1 550 nm. The experimental results show that the method can effectively distinguish the intrusion signal noise, environment and human activities caused by non -invasion, improve the detection probability and reduce the false alarm rate.

It is essential to monitor the resin injection process of Resin Transfer Molding (RTM) for manufacturing large integrated co-cured structures. This paper discusses the efforts in developing fiber Bragg sensors for monitoring the flow of the resin during the infusion process. Three Fiber Bragg Gratings (FBG) sensors are put into the composite layers at different location, one is used to monitor the temperature and the two others are used to detect the arrival of the resin flow front. The experimental results show that the arrival time of the resin flow front to the No.1 and No.2 sensor are 50 min and 115 min, respectively. The injection end time is 166 min at the maximum temperature of 24.59 ℃. The network of FBG sensors would be very useful to provide information of resin injection of RTM procession.

Polarization and phase modulation noises caused by the lead fiber are harmful to a long-haul fiber hydrophone system. In this paper we used a pressure and acceleration insensitive reference hydrophone to obtain the lead fiber induced common noise, and then used adaptive noise cancellation to remove them from a primary hydrophone. Experimental results show that by the use of Normalized Least-mean-square (NLMS) algorithm, a broadband polarization noise caused by disturbing the lead fiber is suppressed by maximal 20 dB and a narrowband phase modulation noise is suppressed by 20 dB. Thus the lead fiber induced noise is eliminated and no reduction in the amplitude of the simulated signal is found. It is shown that the results of adaptive noise cancellation are much better than that of the noise subtract method.

Current applications for photoelectric detection based on PSD were mostly with the incidence of laser, whose effective signal was highly stronger than that of background. A method of light-spot space position measurement under strong background based on PSD technology was proposed to enlarge the range of measurement. The pre-processing circuit was designed to convert the weak current signal from PSD to a voltage signal and finally to a digital signal from high-speed AD transformation. Influence on the results of PSD measurement form the background signal was analyzed. A method for extracting effective signal and accurate calculation for it was proposed by subtracting background and dark current signal. Related experiments with LED light-spot under different outdoor backgrounds show that this method can achieve 200 kHz high-speed measurement. Because of eliminating the influence from different background and dark current signal, linearity error of the measurement under strong background can reach the range of ±1.2%.

The motion estimation in stereo visual localization is usually carried out in 3D Euclidean space. 3D points are corrupted with heteroscedastic and anisotropic noise, and the depth direction of 3D reconstruction is less accurate than the two other direction. And therefore, the accuracy of motion estimation in 3D Euclidean space is low. We propose a new high-precision stereo visual localization algorithm based on motion estimation in disparity space. Initial robust motion parameters and matching inliers are obtained by four points closed-form linear motion estimation in disparity space combined with RANSAC algorithm. Then, we propose nonlinear motion optimization method based on L-M algorithm in the new disparity space re-projection error function. The mothod is appied to optimise the initial motion parameters. The noise in disparity space is homogeneous and isotropic, and our initial linear motion estimation and nonlinear optimization method all are carried out in disparity space and can achive globally minimum. Simulated and real experimental result show that our method can achive higher localization accuracy than traditional motion estimation method in 3D Euclidean space.

The real-time measuring of particle diameter and velocity of raindrop is the key for the statistical of raindrop spectrum parameters. In order to obtain real-time, continuous, reliable and accurate statistical raindrop sample data, the light array method based on high speed line array CCD is adopted for non-contact measurement of the speed and diameter of raindrop. Raindrop images cache unit is structured by FPGA, which stores two rows of continuous dynamic raindrop images. With using FPGA parallel computing ability, raindrop dynamic image connected domain detection and mark algorithm are achieved rapidly by FPGA, and then realizing the particle diameter and velocity detection. The experiments show that this method can automatically and continuously measure and record the number, size and terminal velocity of raindrop, and provide a reliable sample data for the raindrop spectrum parameters statistical.

laser forms a smaller diameter spot in the diffuse reflection screen. By using the light meter to measure the spot illumination, artificial multiple operations is likely to cause errors. This article uses digital photography, calculating the image intensity of the spot camera, establishing the spot illumination and gray of the mathematical mapping model and deriving the mathematical formula based on the Tsai model. The method is simple and accurate. Through experiments and data analysis, the experimental results show that illumination mapping of the location of points and the image corresponding to the gray spot on the CCD response linear relationship provide a new reference for the mapping of the diffuse reflection screen illumination and CCD image gray value.

A method of eliminating texture in defects inspection of LC cell of small-size Liquid Crystal Display (LCD) is proposed, and the domestic first LC cell defects Automatic Optical Inspection (AOI) for small-size LCD was made. The basic principle of this method is to eliminate texture in preprocessing of detecting LCD image by compensation matrix which is established from the image of standard LCD. The method of eliminating texture presented solves the problem that the texture which appears in LCD image will makes it more difficult to detect defects. Theoretical analysis and experimental results show that, by using this method, the texture is eliminated and the defects are kept completely, and the detection accuracy of AOI is improved greatly.

A two-dimensional device of gap detection in tool-setting is designed, which is applied to a small lathe. A way of orthogonal layout is used to construct an optical measurement system, and a gap detection system of tool-setting is established based on image processing. The optical system is calibrated precisely by using standard gauge blocks and applied to measure the micro gap of tool setting. The inspection system is installed on the small lathe developed independently. The system verification is done through motion stage of lathe applying micro displacement. The results show that: the detection algorithm can recognize and detect the gap of tool-setting in small lathe fast and effectively. Moreover, the detection precision can reach 15 μm. The correctness of detection algorithm for small lathe is verified experimentally for high-precision tool setting of small lathe.

In order to overcome the impact of external disturbance on airborne photoelectric gyro stabilized platform, the control strategy based on Pseudo-derivative Feedback (PDF) was applied to achieve high stability and isolation. First, the basic structure of the PDF control was introduced before it was simplified. The advantages of the simplified PDF control in the control structure and control performance compared with classic PI was analyzed theoretically. Then parameters setting strategy of trial and error method for PDF controller was proposed. Finally, a series of simulation comparison experiments were carried out between PDF controller and PI controller in the anti-torque disturbance, isolation of the angular rate disturbance of the vehicle and the plant variation on the airborne photoelectric gyro stabilized platform. The results showed that the servo system of airborne photoelectric gyro stabilized platform with PDF controller was superior to the system with PI controller in the same closed-loop bandwidth. It has smaller overshoot and shorter rise time and regulating time. It is not sensitive to the model variation, and better than PI on the suppression of torque and the airborne attitude disturbances within gyro noise.

A vision-controlled model for two Degrees of Freedom (2-DOF) PZT is established, and through the analysis of the tracked object's location on the image planes of adjacent frames taken by the cameras, we realize a high efficiency and time-saved aim surveillance. Alternative functions are taken to depict the input angle that adjusts the PZT to track the moving object. It is found that the deviant brought in by the approximate model can be neglected compared to the accurate model. The location-feedback method improves the efficiency of aim-searching. The results demonstrate that with 25 frames per second, the camera is able to track the object with lower location deviant, compared with the original. Even when the object’s speed increases to 2 m/s, it could also stably track the object. A binocular vision model is proposed for the controlling system, and it can avoid the object-loss area during the tracking.

In order to overcome the problem that image processing systems based on PC is weak in real-time processing, and systems based on DSP and FPGA cost a lot and inadequate use of resources, a design of real-time image processing system using FPGA with hardware software collaborative processing is proposed. The system uses an FPGA chip as the core of the system, a CCD camera to do image acquisition, and SSRAM to cache the image. SOPC is the control of the core, coordinating hardware and software to processing image. Image processing modules which are easy to use hardware implementation (such as filtering, morphological algorithm, image correction, edge detection, etc.) use the Verilog HDL language to implement. The SOPC is used to control these image processing modules. The image processing functions whose hardware is difficult to achieve (such as process control, complex branching to determine) use the SOPC. The experiment shows the high speed and real-time of the image processing system, and the ability on computation of complex image algorithms. Moreover, it is simple in design, flexible in application and low in cost.