The simulation of the size of the target laser facular is a key process of the Hardware-in-the-loop (HWIL) simulation of laser guided weapons. Through the analysis of the principle of laser guidance seeker, the model for position and size of laser facular in laboratory is built. According to the principle of comparability, the telescope system is used to control the movement of the size of laser facular, and the real target laser facular sought by the guidance seeker is simulated. Based on the mathematical model, the simulation experiments were carried out. The experimental results prove that the mathematic models are right and the simulation way of the size of the target laser facular is realizable, which resolves the key problem of HWIL simulation of laser guided weapon.

Referring to the problem of true-fake warhead recognition in missile early warning satellite system,a new algorithm of warhead target recognition for decision-making system is established. By use of some infrared attributes as characteristic parameter of warhead target recognition, such as temperature change rate, thermal capacity difference, radiant intensity, gray characteristic and infrared movement characteristic, information fusion based on the Analytic Hierarchy Process (AHP) algorithm of multiple attribute decision making is applied. Furthermore, fuzzy dynamic method is applied to construct the attribute importance estimation matrix, and the problem of the attribute weights being changed with time is efficiently solved. The true warhead is recognized by the result of fusion and ranking. The simulation results show that this target recognition algorithm is trusty and practical.

In optoelectronic images of fixed large visual field, the angular velocity of space target in long distance is slow and its image is small, which make it difficult to be detected among so many stellar images. In order to solve this problem, a detecting method for slow small target is put forward based on time-space difference fusion after stellar image registration. In this method, the morphological Top-hat operator is adopted firstly in space domain to suppress the background and enhance the target. Then, the stellar registration and difference fusion are done in time domain to enhance the moving target. Finally, the target can be confirmed through the adaptive segmentation and track correlation. The test result shows that by considering the target’s characteristic in time and space domain, the method mentioned above can effectively detect the slow small target with low Signal-to-noise Ratio (SNR) in complicated background and own good real-time performance.

In order to analyze the error of the measuring co-ordinates of shots with sky screen in ballistic trajectory and evaluate the accuracy of survey parameter, the creating factor and transitive relationship of parameter error were studied according to the sky screen structure and principle of measuring co-ordinates of shots with sky screen. Moreover, the parameter error of measuring co-ordinates of shots with sky screen was analyzed based on the parameter of experiment, and the error scope of the five measuring parameters was calculated in theory. Through the fire experiment, the target pitch angle, azimuth, and coordinates of measuring system are basically the same to the paper target, the velocity is the same to the other testing instrument, and the error is in the scope of analysis. The experiment proves that, the measuring co-ordinates of shots with sky screen in ballistic trajectory are high precision and stable reliable measuring instruments.

To improve the stability and accuracy of small target tracking in cluttering background, a new small target tracking method based on gradient Laplacian of Gaussian (LOG) operator was proposed. First, imaging model of small target in gradient image was presented. Second, the theory of scale space was used to design a new operator for target detection and tracking. Then, the property of scale invariant was analyzed and proved, and a tracking scheme based on the gradient LOG operator was proposed. Experimental results show that the gradient LOG operator is appropriate to describe small target in complex background, and the detection precision is in the level of sub-pixel. Furthermore, it is efficient and robust to noise.

As one of the background estimation algorithms for Infrared (IR) point target detection, Least Squares (LS) method has a poor performance to the complex nonlinear background. A nonlinear version of the least squares algorithm, called Kernel Least Squares (KLS) is deduced by using Kernel Methods (KMs). Furthermore, the exponential weighted form of KLS, called KEWLS, is deduced. KEWLS is more adaptive to dynamic nonlinear system’s time-series prediction. A kernel-based IR target detection algorithm is proposed, image background is estimated by KEWLS nonlinear regression, and then target is detected by self-adaptive threshold detection in the difference image. It is shown by nonlinear function regression and sequence IR images detection experiments that the kernel methods improve the performance of nonlinear function regression and IR background estimation.

A novel approach for evaluating the tracking ability of photoelectric theodolite is proposed. Equivalent model of theodolite tracking error based on the BP neural network structure is identified. The Levenberg-Marquardt (LM) algorithm is adopted in the training method of BP neural network for the sake of speeding up training process. The equivalent sine signal is inputted to the model, and the output is gotten. The evaluation of tracking performance is obtained based on the statistical calculation of output. The estimate errors of equivalent model including average error, maximum error and standard error are 2.5872e-006°≈0°, 3.6″and 1.6″, respectively. The result shows that the equivalent identification model based on BP neural network meets the needs of evaluating the tracking performance of theodolite. The accurate evaluation of tracking performance is achieved.

The meaning of Micromachine Membrane Deformable Mirror (MMDM) figure influence function is discussed, and the figure influence function matrix of oko company 37-channel MMDM is measured. The linear relationship between the deformation of mirror surface and single actuator voltages square is experimentally verified. MMDM correction capability is analyzed and a closed-loop correction model of adaptive optical system is established. An adaptive optical experimental system based on MMDM is built. Moreover, wavefront from human eyes is used as the incident wavefront of adaptive optics system. Experimental results show that the model can correct dynamic distorted wavefront aberration rapidly and effectively, is very suitable for adaptive optical system based on MMDM.

To meet the request of real-time control and the flexility of the Stochastic Parallel Gradient Descent (SPGD) algorithm for Adaptive Optics (AO) system, an implementation of the SPGD algorithm in FPGA was proposed. Firstly, the diagram of the SPGD algorithm controlling system was designed. Then, the key modules were processed for real-time control, and a deformable-mirror controlling module that can be updated for future application was designed according to the pipeline and RAM techniques. Based on this implementation, a laser beam focusing system with a 61-element deformable mirror and high frame-frequency CCD was built. The experimental results demonstrate that the AO system can correct static distorted wave-front successfully with different system performance metrics and can also control deformable mirror and tilt mirror simultaneously.

The method of traditional cumulant is a standard technique used to analyze dynamic light-scattering data. However, the different baseline values influence the results measured by using dynamic light-scattering. A method of moment-cumulant, which fits the baseline automatically, is proposed. The correlation function is used as the model function to inverse the particle diameter and distribution factor pi directly. An experiment is done with latex particle of 90 nanometers in laboratory. Compared with the two cumulant models, it is shown that the particle diameter and particle distribution factor pi are unstable in width fit range and lead to deviation with traditional cumulant, while measurement results is good in stability and not influenced by correlation time with the cumulant method of baseline auto-fitting.

A portable detecting device based on image measurement is developed to study wheel-rail geometrical contact state and make it visible. Two high speed CCD cameras are installed on both side of the rail to capture wheel-rail contact images synchronously while wheels passing through the inspecting area in high speed. Wheel-rail profiles are drawn out by specially designed image processing arithmetic, abnormal geometrical contact states of wheel-rail are recognized, and wheel-rail contacting track along inspecting area are worked out. This paper introduces inspection principle and emphasizes on parameters designing of high speed cameras. Installation feasibility and inspection capability are validated by many experiments, such as static test, dynamic tests in low speed and high speed.

For effectively acquiring object 3D surface data to reconstruct its 3D digital model, the structured-light measurement technology is used to establish a new 3D measurement system. Based on the Gray code and phase shifting coding algorithms, a new coding algorithm called Gray＋Phase shifting code is developed. Firstly, the map is divided into certain areas. Secondly, structured-light is projected separately in every area. Finally, every phase-shifting picture moves to right a certain pixel. The coding period is shortened due to the simplified plane projection, and the image processing is decreased. Moreover, the scanning rate is improved so that it is suitable to fast 3D model reconstruction. Some experiment results in our 3D surface information collection system are given to demonstrate the effectiveness of the presented scheme.

Long-playing large-scale monitoring and small-target tracking can be implemented by the photoelectric detection system based on unmanned airship, which is particularly suited to momentous task, such as earthquake relief or flood succor. An enhanced photoelectric detection system based on unmanned airship is designed by using the two-axis, double-frame structure and three closed-loop control methods combined with dual GPS information. The whole system design is described in detail, including the system structure, function, control method as well as hardware and software. Moreover, the correctness and reliability of the system are verified by the system test and flight experiments.

In order to evaluate the detection distance of laser detecting system, a simple method for estimating the detection distance of laser detecting system is presented based on the simulation of atmospheric transmission. The specific property of atmospheric transmission and the method for simulating the process were both analyzed in detail. Pointing to the imitation of laser detecting system, a test was carried out both inside and outdoor. Some valuable data was acquired. In the light of fluctuation of outdoor test data, influence factors were also theoretically analyzed. The testing result confirms that the method is feasible and practical to evaluate the detection distance of laser detecting system simply, to provide the distance reference for the outdoor test, and to heighten the test efficiency, which will be useful for the test of laser detecting system.

The slow imaging speed of traditional Optical Coherence Tomography (OCT) system limits the practical researches and applications of OCT , so a kind of fast OCT system without scanning frequency-domain was presented. The imaging resolution ratio, the depth of investigation and a variety of noise were analyzed, and a new method for eliminating background noise was brought forward. Based on the characters of the frequency-domain OCT and imaging spectrometers, the OCT system collected data rapidly with the line-focus mode and a new interferential optical path. Experimental results indicates that the two-dimensional tomographic image can be obtained without any scanning drive unit, the imaging speed for a glass slide which has 8mm width is two orders of magnitude quicker than the traditional point scanning mode and the background noise is eliminated successfully with the data processing algorithm.

Due to the Depth of Focus (DOF) which is tens of times more than particle diameter in digital in-line Fresnel holography system, resolution of particle axial position is low and uncertain. In order to reduce the impact of DOF,an auto-focusing algorithm for accurate location of particle was proposed based on the most gradient. In the focus window containing particle, the gradients in all possible directions were calculated. Then, all of the possible gradients were compared, and the largest gradient was chosen as the result. At the same time, in order to improve the precision and stability of auto-focusing algorithm, another threshold parameter was introduced to eliminate the interference caused by background information and noise. Through experiments, we find out that the auto-focusing algorithm has the strong single peak feature and good stability.

Atmospheric attenuation has strong effect on optical signal, and scattering atmospheric optical communication system usually adopts receiving antenna to strengthen the signal collecting ability and increase its communication distance. The receiving angle and bottom radius are the determinative parameters of Compound Parabolic Concentrator (CPC). It is showed that CPC’s height will be taller and its light collecting ability will be stronger when CPC’s field of view is smaller at the same bottom radius. The ray tracing analysis shows that CPC’s field of view is wide and its gain is high. The light propagation of scattering optical communication based on multiple scattering was simulated by Monte Carlo method. The results show that using CPC as receiving antenna could strengthen the signal collecting ability of communication system and increase its communication distance. CPC is the suitable optical antenna for scattering optical communication.

A cost-effective approach for Cable Television (CATV) signal transmission over a Wavelength Division Multiplexing Passive Optical Network (WDM-PON) is proposed and CATV signal transmission performance is analyzed in detail. The analytic models for a fundamental optical link,consisting of optical amplifier spontaneous emission (ASE) source, Mach-Zehnder Modulator, optical filter, transmission fiber and photodiode,are presented. Considering a variety of system noises’ influence on the transmission performance of CATV signal, the analysis concentrates on the impact of optical modulation depth, average received optical power and ASE spectrum-sliced bandwidth on Signal-to-noise Ratio (SNR). The numerical analysis shows that SNR can reach above 43 dB at the bandwidth of 750 MHz, thus meets the need of CATV transmission over fiber to the home. From the experiment result, it is seen that EDFA can effectively suppress the relative intensity noise, improving SNR by 3-4 dB.

Aiming at the problem of object-based image retrieval, a new Multi-instance Learning (MIL) algorithm based on Fuzzy Support Vector Machine (FSVM), called FSVM-MIL algorithm is presented. The standard MIL problem assumes that a bag is labeled positive if at least one of its instances is positive, otherwise, the bag is negative. FSVM-MIL algorithm treats the whole image as a bag, the segmented regions as instances, and then each image with the desired object is labeled as a positive bag, while the other is labeled as negative bags. In order to address the ambiguity of instance labels in positive bags, the Diverse Density (DD) method was used to search multiple hypotheses, and according to noisy-or probabilistic model, a fuzzy membership function was defined which can give different fuzzy factors to these instances in the positive bags. Thus, the multi-instance learning problem was converted to a FSVM learning problem. Experimental results based on the SIVAL data set show that this algorithm is feasible, and the FSVM-MIL is competitive with other state-of-the-art multi-instance learning algorithms.

To design more effective algorithms for Independent Component Analysis (ICA), a general framework of Gradient Descent Algorithms (GDAs) for ICA was proposed, which covers many popular algorithms such as Infomax, Minimization of Mutual Information (MMI), Maximum Likelihood Estimation (MLE) and so on. This framework was derived from a new theory of the contrast functions for ICA based on the superadditive (or subadditive) function of class II. For better performances, the Equivariant Adaptive Separation via Independence (EASI) form was generalized and used as the updating rule. An example of using the framework was also shown based on the quadratic entropy. Furthermore, a fast method of computing the gradient in the example was proposed and the simulation proved its validity. The results demonstrate that this framework is a useful tool to discover more effective algorithms for ICA.

The characteristics of Human Visual System (HVS) on noticeable brightness difference and correlation properties in natural color images are brought into the image enhancement algorithm. Based on the HSV color space, a novel Unsharp Mask (UM) image enhancement method guided by Optimum Noticeable Difference (OND), called OND-UM, is proposed for the V component to bring about the edge and detailed information. The key procedures include that the OND curve is obtained by the numerical fitting technique and an adaptive gain function is designed with local variance. Especially, for the low-contrast regions, the proposed Adaptive Stretching Function (ASF) improves the global contrast. And the S component is histogram-equalized and H component remains the same. At last, the color restoration is conducted in RGB space on statistic data of color correlation properties. Experimental results show good performance of the OND-UM to enhance the edge and reduce noise sensitivity, and the enhanced images after the color restoration look more natural and vivid and are suitable for the human vision.

Based on the analysis of the current kinds of Region of Interest (ROI) coding, a mask embedded SPIHT algorithm was proposed, which supported multiple ROI in arbitrary shape. The innovation of proposed algorithm was that the coding of ROI mask was embedded into the SPIHT coding process. As the image and the mask information were coded synchronously, the bits stream could be truncated arbitrarily. Image quality evaluation method for ROI coding was also discussed. An image quality evaluation value called importance & area Weighted PSNR (WPSNR) was proposed by fully considering importance differences and area ratio between ROI and background. Experimental results show that proposed algorithm supports the compressing of multiple arbitrary shape ROI images from loss to lossless compression with configurable ROI priority. The generated bit stream could be truncated at any place, and the decoder can still decode image and ROI mask information well. Proposed algorithm was proved that it had a better performance than BbB-shift method. As the computational complexity of proposed coder was close to SPIHT, it can be used for ROI coding in low-rate applications or ROI preferential progressive transmission.

Retina recognition has the advantages of high stability and security. Because low resolution retina image obtained by traditional fundus camera is difficult to be coded and matched, retina identification recognition is not widely applied. High resolution capillary vessel image is acquired by adaptive optics high resolution retina imaging technique, and the texture of image is extracted by Gabor filters after being normalized and denoised. The character matching is completed by Hamming distance. Experimental results show that the proposed method can accurately identify different samples and is feasible. Retina capillary vessel identification recognition extends the application of adaptive optics high resolution retina imaging technique and has a potential application in some important situation.

A new model of doublet Accommodating Intraocular Lens (AIOL) is proposed by using a reasonable simulation method which is more applicable for the practical case of human eyes. The induced accommodation of the singlet AIOL and doublet AIOL including the combination of two positive lenses, positive-negative lenses and negative-positive lenses are compared with nearly the same vision performance for the eyes implanted with these AIOLs. These results show that the induced accommodation of doublet AIOL constituted by two positive lenses or an anterior negative lens and a posterior positive lens is lower than that of the singlet AIOL, the induced accommodation of doublet AIOL comprising an anterior positive lens and a posterior negative lens is obviously higher than that of the singlet AIOL, and the accommodation amount increases with the power of anterior lens. When the spherical aberration Z40 varies from 0.096 to 0.263, the tuning range of the effective accommodation is only 0.03D. The above results may be helpful for understanding the current designs and developing doublet AIOL.

With the technology of computer aided engineering, supporting structure of optical reflector of the remote sensor was designed and analyzed. Because supporting structure of the reflector can meet the demands of imaging quality under environment of gravity, but out of tolerance under thermal environment, the supporting structure was improved by weakening stiffness and strengthening flexibility. By modeling, simulating, analyzing the improved structure, and adjusting parameters of flexible structure, the reflector groupware can achieve demand of imaging (PV≤63.2 nm) under environment of gravity. At the same time, the stability of thermal dimension was well. Through dynamic analysis, the structure can not be shaken, fatigued and broken under environment of sin scan and random vibration. Simulating results indicates that surface accuracy of the reflector meets the needs of the imaging quality under mechanics and thermal loads through regulation of flexible structure, and dimension’s stability of structure is better than ever, which shows that the flexible regulating structure is reasonable and feasible.

3-RPS symmetrical parallel manipulator is the key part of the parallel supporting structure. By means of the complete differential-coefficient matrix theory, the mechanical position-stance error model expressed by Rodrigues Parameters including the main errors of the parallel supporting structure was established. The sensitivity percentage was proposed for describing the error sources of the parallel supporting structure, and the corresponding sensitivity model with statistical basis was achieved. According to the results of numerical simulation, the actuator errors and the errors of the moving platform along Z axis should be controlled strictly in the course of the design, manufacture and assembly of the manipulator. Although the structure scale was changed, the errors of high sensitivity should be considered fully. Based on the sensitivity analysis of the parallel supporting structure, the calibration is effective to reduce the errors of the kinematic parameters with high sensitivity.