The main function of optoelectronic payload is remote sensing and measuring by optoelectronic sensor in order to search and detect the interested target and locate it with a determinate precision,which implements remote sensing and measuring integrity.One optoelectronic payload for Unmanned Aerial Vehicle (UAV) is made up of 20× optical zoom visible-light TV,wide/narrow Field of View (FOV) movable long wave Forward Looking Infrared (FLIR) and 5Hz repetition frequency laser range finder.They are fixed compactly on a two-axis four-frame stable platform whose stability precision is 25μrad.Based on this configuration,the three axes alignment accuracy of visibility,IR and laser is up to 20″,and the measuring angle precision error is less than 1mrad.The flight-testing indicates that the optoelectronic payload download image is clear and the target orientation data is reliable.Increasing laser repetition frequency and platform isolation can improve the efficiency and quality of searching and detecting.

To avoid failing in visual tracking situation when employing single feature,a nonlinear target tracking method based on multi-feature fusion is proposed.Grey histogram is used to describe the overall distribution characteristics of the target and edge feature is employed to extract the high frequency details.The two algorithms are fused in the probabilistic model of particle filter.Feature reliability estimation based on half-band width and contribution is proposed,which provides more reliable features with more particles.In this way,the particle numbers of the features are adjusted dynamically.Compared with single-feature tracking method,the tracking result shows that the algorithm has the strong ability of tracking under local obstruction.The average tracking error of the new algorithm decreases by 0.5 pixels.

When there are many fuzzy rules during Fuzzy-neural Networks (FNN) operation,the networks usually have slow learning speed and long running time.To solve this problem,a rough set theory was introduced to improve FNN model.Rough set data analysis method was used to obtain the reductive rules which were used as the fuzzy rules of the FNN.The input to the model was mapped into the output subspace by using these rules and the output of the system was approximated by improved BP algorithm training in this subspace.The results show that rules acquired by rough set data mining technology not only can get minimum rules but also are incomplete rules.Dimension of input and nerve cell numbers of network are decreased and learning speed is improved,which can meet time limitation of system.

A new method for measuring dynamic attitude of missile is proposed by adopting the binocular vision measurement system,in which two CCDs are used and their optic axes are parallel.Firstly,the centroid of missile is obtained by the weighted average method.Secondly,the axis vector of the measured missile is fitted by the 3-D coordinate of missile’s surface which is obtained with image matching method.Because the axis of missile is a straight line which gets through the centroid of missile,the interval between the measured missile’s surface points and its axes is equal.Lastly,the attitude of measured missile is obtained by the mathematics relationship between the axis vector and the attitude.The simulation experiment shows that the measurement method has higher precision,and can meet the measurement requirements on the dynamic parameter of moving object in long distance.

Background subtraction based on Gaussian Mixture Model (GMM) is a common method for real-time video segmentation of moving objects.An effective adaptive background updating method based on GMM is presented.The number of mixture components of GMM is estimated according to the frequency of pixel value changes,and the performance of GMM can be effectively improved with the modified background learning and update new distribution generation rule and shadow removal based on morphological reconstruction.The detection of sudden illumination change and background updating are also proposed.Compared with existing approaches,experimental results with different real scenes demonstrate the robustness of the proposed method.

Due to outlier disturbance in kernel window,the traditional Mean-Shift method fails to track enlarging objects.Based on assumption that initial object model can decrease outlier disturbance on tracking results effectively,we put forward new object candidate model and similarity measure which make use of initial target model.In order to make the center and size of kernel window similar to object,a new method which can adjust dynamically to kernel window‘s size and location according to the distribution of pixel in sub-band around kernel window is also presented.The proposed method is applied to track enlarging cars,which verify the effectivity of the method.Experimental results show that the new method decreases computing complexity greatly.

Traditional Hough transform for circle detection has several limitations,such as high computation and memory complexity.To solve these problems,we presented a modified Hough transform method based on gradient information and fuzzy theory.First,gradient information was utilized to reduce the computational complexity.Furthermore,edge points were mapped based on fuzzy theory to avoid peak diffusion and false peak.Finally,a post-processing algorithm was designed to remove the false alarms,since Hough transform would neglect the connectivity between points.Experimental results show that the proposed method has many advantages,such as low computational complexity,high detection accuracy,high detection rate,and low false target rate.

The Infrared Focal Plane Arrays (IRFPA) imaging system is the main technique tendency of infrared imaging,but the inherent nonuniformity of IRFPA crucially limits its imaging quality.Because the operands and data volume are so vast that it is difficult to realize real-time processing in the Non-uniformity Correction (NUC) on the infrared focal plane.In the drop-in handware system,TMS320C6000 DSP was applied to resolve this problem,which combined Linearity and Second-degree Polynomial segment algorithms.The hardware design and implementation flow of the system was described and the experiment results were given.The results show the real-time non-uniformity correction system meets the real-time and high-precision requirements completely.

The paper solved the problems of less valid data in thermal infrared imager,impurity being on surface of specimen and uneasy dipartite corrosion area in image which happened in detection of structure parts’ hidden corrosion by thermal wave imaging.Firstly,the contrast of one unheated image was enhanced by self-adaptive mathematical morphology sharpening,and then location of the impurity was confirmed by change of the image grads.Secondly,the impurity in post-heated image sequences was filtered based on minimum of (5×5) template at the same impurity location.Thirdly,the sequence of images was transformed based on wavelet Daubechies9/7.Low frequency’s energy was cumulated and high frequency’s energy was reduced through analyzing distribution of the active thermal image’s energy.And then the thermal image was reconstructed.Finally,quantitative evaluation was presented between the reconstructed image and the original image.The experimental result indicates that the method can improve the quality of IR thermal image,reduce noise of IR thermal image,and obtain identified IR thermal image easily.

Both the theoretical analysis and computing simulation were done for the photon correlation spectroscopy (PCS) of an actual particle system,and the results showed that the relative error of the reversed particle size could be up to 100% if the mixed PCS was reversed only by the self-beating or heterodyne arithmetic.So we presented a method of reversing the particle size from a mixed PCS.A simulated experimental arrangement which can give rise to the mixed PCS was built.Two particle systems,composed of standard polystyrene latex spheres with nominal diameters of 60nm and 200nm,were examined with the arrangement.The experiment result shows that compared with their nominal values,the relative errors of the reversed particle sizes are all within 5% when the mixed PCS is reversed by the proposed method.

The electro-optical effect of jamming smoke screen depends on the composite of the smoke agent.The paper introduced the influence factors of IR extinction property with the investigations of optical and microphysical characteristics of aerosol.Take NaCl as the filler,and compare the attenuation performance of NaCl water vapor on infrared radiation under different concentration,dosage and target temperature conditions by experiment.Experimental results show that increasing the concentration and dosage of NaCl aerosol is good for enhancing extinction ability of water fog.IR extinction performance was calculated and analyzed according to Mie theory for water vapor whose granularity distributing keeps up logarithm normal distribution.The transmissivities are 38% at 3.5μm and 55% at10.591μm,respectively.At last,the approach to choose interference filler with good electromagnetism and suspending characteristics is put forward.

A novel virtual sensor measurement system with one high-speed camera was proposed for measuring motion parameters of free flying insect.An optical image system was designed and the virtual four-ocular stereo vision measurement system was developed based on analyzing the principle of virtual four-ocular measuring.To measure the motion parameters of free flying insect,a free flying inducement system aimed to the small view field of virtual stereo system was designed.The measurement system and inducement system were verified by measuring the honeybee’s double wings motion parameters.The experimental results show that the measurement system is feasible for measuring free flying insect.

In some case,the axial distance from the sample to the lens is long in laser particle size measurement,so vignette appears because the lens aperture is limited,which reduces the measure precision and accuracy.In order to solve the problem,the gray system theory is presented to check the optical signal data to find the vignette point.And then,the signals after the vignette point are predicted by grey-prediction model.By this way,the vignette data are amended.Experiments on a particle plate were carried out.The particle plate was placed at several positions in front of the lens.When the data were modified,a no-model algorithm was used to invert the particle size distribution.The results show the gray-prediction method is valid.The purposed means can be used to enhance the precision and accuracy in some measurement,where the axial distance from the sample to the lens is long or the covered area of the sample is large.

Since a straightness inspection system for long shaft needs line and angle datum,a method of straightness inspection using cross laser for long shaft is proposed in this paper.Based on the principle that the distance from the axle centre of inspected long shaft to the two laser planes is unique,the image processing is used in the method to get the coordinate of the cross point and the angle of the horizontal line of the laser image and calculate the straightness parameters with distance information.The structure of inspection system is discussed.The calculation method for parameters of straightness is explained in detail.With a camera (640pixel×480pixel and 0.097 3mm/pixel as resolution of target),50 samples are taken at each distance of 1m,9m and 18m.The standard deviation is 0.009 37mm,0.034 2mm and 0.086 0mm for polar radius,0.843′,1.26′ and 1.57′ for angle,respectively.Experimental results show that the straightness inspection system for long shaft using the method is simple in structure,easy in operation and precise in inspection.

In order to break through the limitation of Electronic analog-to-digital converter in speed and time-delay precision,a wideband digital frequency measurement method based on optical sub-sampling is proposed.Taking advantage of ultrashort optical pulse from passive mode-locked fiber laser,multi-wavelength optical pulse cluster with controllable time-delay is produced.Traveling through electric-optic modulator,pulse cluster samples the signal and is detected by photodiode array,and then we could calculate the frequency spectrum in electric domain.Combined with Multiple Signal Classification algorithm (MUSIC),the method has advantage of high sampling bandwidth,accurate time-delay and Wavelength Division Multiplex (WDM) technology.According to theoretical research and numerical simulation result,this method is able to get high-resolution multiple frequency measurement with a bandwidth of 20GHz.

The structure and the incorrect discharge status of three-axis laser gyro are introduced briefly.The necessity of status monitoring circuit is proposed.Based on the designed high voltage and current stability circuit,the sequence between current stability and status monitoring is controlled by adding an analog switch to each current stability branch of the high voltage power supply.The status monitoring is also realized by monitoring the voltages of the sampling resistances.Moreover,the realization of software and hardware of status monitoring circuit for three-axis laser gyro are introduced in detail.In the end,the circuit is tested,the ignition ratio of three-axis laser gyro is up to 100% in 5 seconds and long-term current stability exceeds1×10-4.The feasibility and reliability of the circuit is validated.

Based on coupled-wave theory,a new approach of deriving diffraction efficiency equation on reflecting volume Bragg gratings was proposed,On the basis of the derived equation,the detailed theoretical modeling for diffraction of plane and Gaussian beams on reflecting volume Bragg gratings were obtained,and the diffraction efficiencies including plane monochromatic,divergent monochromatic,and polychromatic beams were analyzed.Numerical results show that the angular selectivity could be varied from more than 100 mrad to less than 0.01 mrad,and spectral selectivity could also be varied from more than 100 nm to less than 0.1 nm.When angular divergence or spectral width equals the angular or spectral selectivity,the grating can provide maximal diffraction efficiency of more than 88%.When angular divergence or spectral width is far less than the angular or spectral selectivity,the decrease of diffraction efficiency is insignificant,which is less than 1% of plane monochromatic wave.

TiO2-Ta2O5 thin films are new-style optical thin films which are made of mixed compounds as starting material.The TiO2-Ta2O5 compound thin films were fabricated by ion beam assisted deposition on K9 glasses with Ta2O5 and TiO2 of different proportioning as starting materials,and the optical properties were investigated.The testing results show that TiO2-Ta2O5 thin films have the higher transmittance and the lower extinction coefficient at visible light wavelength,and the refractive index varies from 1.80 to 2.07(550nm).TiO2-Ta2O5 thin films are of great significance in optical fields.The optical band gap of TiO2-Ta2O5 films increase from 3.266eV to 3.417eV while the content of Ta2O5 enhances from 0 to 20%,and the shift of absorption edges in transmittance spectrum is explained by Kayanuma model.

According to the relation between intensity Autocorrelation Function (ACF) and normalized light field ACF,the effect of different baseline values on the results measured by using Photon Correlation Spectroscopy (PCS) was presented.The cumulant method which was used to inverse the size and polydispersity of nano-particle in PCS was analyzed.Three types of materials with five different sizes were measured by using PCS and the results showed that when the baseline methods were different,the value of baselines and the results of the effective diameter and polydispersity inversed by cumulant method were different,too.The reason was analyzed theoretically.It is pointed out that the autoslope method of acquiring baseline is the appropriate measured method.

Combined wavelet with directional filter banks,a non-redundant image transform called as Wavelet-Directional Filter Banks (WDFB) was realized.It fulfilled the anisotropy scaling law and could represent the geometrical features such as edges and texture more sparsely.By taking the advantages of the WDFB,an efficient image compression algorithm based on morphology operator was proposed.The new algorithm used the significant tree to express the correlation of inter-band while utilizing morphological dilation operator to cluster the significant coefficients within sub-band.Experimental results show that the performance of the new algorithm is superior to the wavelet-based algorithms in terms of Peak Signal to Noise Ratio (PSNR),especially for the images including a large portion of texture.For example,for Barbara image of 512×512,at 0.25bpp,PSNR of the new algorithm outperforms that of SPIHT and MRWD by 1.05dB and 0.77dB,respectively.

A new adaptive contrast enhancement algorithm for infrared images is presented based on histogram equalization.By analyzing the histogram of infrared image,a weight function is gotten adaptively by which the original image’s histogram is weighed.Then,conventional histogram equalization is performed on the original image based on the weighed histogram.With the proposed method,threshold is not needed.Moreover,disadvantages that histogram equalization enhances background are overcome.The proposed algorithm is used to enhance the contrast of infrared images generated from an infrared focal plane array system.Results illustrate that contrasts of processed image are improved significantly.Theoretical analysis and experimental results also show that it is an effective enhancement algorithm for most infrared images.

To adapt the diversification of network and terminal devices better,an optimization scheme of scalable video coding was proposed with Motional Region of Interest (MROI) based on H.264.The motion information and coding mode in base layer were used to determine the MROI,and then the MROI was enhanced as one slice by scalable coding.Experimental results show that the proposed scheme can greatly decrease coding complexity and make the MROI information have a higher probability to be included in the truncated bit stream.Therefore,the scheme improves the reconstructed quality of the potential MROI and subjective quality of the whole image.

In order to solve the matching problems of feature points in two images,we introduced a new point feature matching method based on a new shape content descriptor.Firstly,the corners and the curves relative to each corner were obtained using the Curvature Scale Space (CSS) technology.Secondly,the shape content descriptor of each corner was computed,and the matching degree was gained by χ2 distance between the two descriptors.The matching was successful if the degree was bigger than a certain threshold,so the initial set of matches was obtained.Finally,more accurate matches were eventually found by using the semilocal constraints to the initial set of matches.The experimental results show that a very high correction matching rate can be obtained by the algorithm.

A new algorithm for medical image edge detection is proposed based on multi-structure element morphology.Four different edge detection results are gotten by using the multi-structure elements with the anti-noise morphology gradient respectively,which is derived from the open and close operations.And then we adopt the quad-decomposition characterized by mean and standard deviation to enhance the thin edges.The performance of the new algorithm is analyzed and proved both on theoretical and stimulation results.Experimental results show that the new algorithm are more efficient for image denoising and it can preserve more edge details in the dark region than those traditional edge detection algorithm.

Frequency shift effect in photonic crystals modulated by shock wave was studied.The physical model of one-dimension photonic crystals modulated by shock wave was built.Quasi-static bandgap structure was put forward,and its bandgap was calculated using Finite-difference Time-domain (FDTD) and Pulse Width Modulation (PMW).By performing numerical computation and perturbation analysis,the incidence frequency increased by integral multiple of a certain amount in the process of frequency conversion was found,and photon-phonon interaction and surface decaying mode interpretation for the mechanism of frequency shift was put forward.The characteristic of frequency shift effect was also analyzed.

Aiming at current difficulty of the real-time and dynamic reconstruction of digital holograms with large viewing angle,a novel method was proposed according to the principle of time-sharing reconstruction based on the Liquid Crystal Spatial Light Modulator.The computing principle of large-viewing-angle holograms and technologic principle of time-sharing reconstruction utilizing spatial light modulator were analyzed,and a method of multi-screen jointing in space was proposed.Furthermore,we developed the experimental system by reshaping the optic engine of rear-projector.The computed holograms were processed to be with RGB format so as to accord with the transmission of RGB signals in the driving circuit.The processed holograms were output to the optic engine sequentially with a frequency of 60Hz.The dynamicly reconstructed image with large viewing angle could be obtained.

The experimental study of gain flattening dispersion compensation wide-band Fiber Raman Amplifier (RFA) was presented by using 1445nm high power Fiber Raman Laser (FRA) as pump and two different source combinations as signals.One used broadband ASE light source and WDM-emulator as Dense Wavelength Division Multiplexing (DWDM) signal source,and the other used four-channel External Cavity Laser (ECL) and filter type Wavelength Division Multiplexer (FWDM) as signal source.Moreover,Chirped Fiber Bragg Grating (CFBG) as gain flattening filter was designed to flatten actually tested gain spectra.Wide-band FRA with bandwidth from 1519nm to 1574nm (55 nm) was successfully obtained.Its average gain and gain ripple could reach 15.2dB and ±0.8dB respectively.It provides a new solution for wideband FRA design within 60nm.

According to skin color distribution in YCbCr color space,a novel face recognition method based on radial basis probabilistic neural network (RBPNN) in Eigenface Space of Mahalanobis Distance Map to Skin Color Information was proposed.Mahalanobis distance map of image was set up with skin color information,and Eigenface space was constructed depending on K-L transformation.In Eigenface space,statistical characteristics of image were extracted,which were served as inputs of neural network.Radial basis probabilistic neural network was constructed in terms of its nonlinear computation and mapping,and used to recognize and classify face image.Experimental result shows the approach is effective to identify.

Two different architectures of Independent Component Analysis (ICA) to palmprint recognition which were architectureⅠand architectureⅡwere discussed.Region of Interest (ROI) in the palmprint images were extracted automatically by preprocessing before feature extraction and match so as to increase the recognition accuracy and reliability.In order to reduce computational complexity,Principal Component Analysis (PCA) was used to eliminate second-order dependencies in the palmprint images.The remaining higher-order dependencies were separated by ICA.The Square Project Error (SPE) of ICA model was smaller than that of PCA,and the reconstruction of the original palmprint images was superior to that gotten by PCA in PolyU Palmprint Database.To compare the recognition performance of two ICA architectures with PCA,we applied them to extract the palmprint feature subspace inside ROI.Then the images to be recognized were projected on small dimension subspace.Finally,we used a classifier to palmprint match based on cosine distance.Experimental results show that two ICA architectures perform better than PCA and ICA architectureⅡis the best in performance.

Due to the differences of various infrared cameras parameters and the surroundings influence,the robustness of the threshold auto-selection algorithm in infrared images segmentation has not been well resolved.Based on the mechanism of infrared imaging,we presented a new solving scheme and put it into practice.Firstly,the image histogram was clustered by K-means clustering method,and then the distribution characteristic of the cluster centers was analyzed in detail and the threshold for image segmentation was determined.This algorithm doesn’t need to equalize the image before segmentation and assume the distribution of the background.Experimental result shows good robustness.