Filters for soft-x-ray are mostly self-supported metal thin films. Under experimental circumstance, while long exposure to the air of the self-supported metal thin film, impurity atoms will enter into self-supported thin film, which will result in decline of self-supported thin film′s optics performance. Zr filters have high transmittance at band from 5 nm to 20 nm because of its low mass absorption coefficient and density. After vapor depositing NaCl film on cleaned float glass as a release agent, Zr film was sputtered on NaCl film. Self-supported Zr film wass prepared after NaCl film being dissolved in deionized water. To prevent surface oxidizing and impurity atoms from the air entering Zr film in experimental environment, C or Si film with 10 nanometer thickness was sputtered on both Zr film sides for separating Zr from the air. Measurement result of C/Zr/C and Si/Zr/Si film shows that C or Si film has little influence on self-supported Zr film performance.

A visible/infrared broadband dichroic beamsplitter which can meet the light energy boundary of polarization remote sensing optical system was designed with the induced-transmission method. The design was based on polarization characteristics of the visible/infrared broadband dichroic beamsplitter, and actual characteristics of the dichroic beamsplitter were analyzed. When the polarization characteristics at the transmission spectrum from 400 nm to 900 nm was calculated, the coating series was designed on the back of substrate with the equivalent layer theory to control the polarization character of the dichroic beamsplitter. A needle optimal algorithm was adopted to set various optimization objectives, and came out with three different designs which have respective polarization ranges of 0.013~0.0376, 0.068~0.108 and 0.098~0.147. The design result shows that the polarization characteristics of dichroic beamsplitter at transmission spectrum are controllable to some extent, and the design can meet the needs of different polarization characteristics in optical system.

Polycrystalline hexagonal AlN thin films were prepared by DC magnetron reactive sputtering on Si(111) substrates.The influences of sputtering pressure on structure and surface roughness of the AlN thin films were investigated.The results show that:when the sputtering pressure is less than 0.6 Pa,the AlN thin films are amorphous and there is no obvious absorption in the Fourier transformation infrared spectra; when the sputtering pressure is no less than 0.6 Pa,the thin films are polycrystalline structure with the hexagonal phase (100),(110) and weak (002) peaks and an intense absorption exits at 677 cm-1 of wavenumber in the Fourier transformation infrared spectra; when the sputtering pressure increases,the surface roughness of the thin films decreases firstly then increases,while the deposition rate increases firstly then decreases,and the higher deposition rate can help to reduce the surface roughness of thin films; it has the minimum surface roughness and the maximum deposition rate when the sputtering pressure is 0.6 Pa.

Aiming at overcoming the limitations in extracting impervious surface by traditional methods, two non-linear spectral mixture models, Mixture Tuned Matched Filtering (MTMF) and Multi-Layer Perceptron(MLP) neural network, are used to decompose all pixels to the four fraction images representing the abundance of four endmembers. In these models, MTMF performs a “partial” unmixing by only finding the abundance of a single, user-defined endmember, by maximizing the response of the endmember of interest and minimizing the response of the composite unknown background. The MLP is a hierarchical structure of several perceptrons, and capable of learning a rich variety of nonlinear decision surfaces. The Maximum Noise Fraction(MNF) is used to transform the six bands of TM image into a new feature space and the first three components accounting for the majority (more than 90%) of total information content are used to endmember extraction. After that, the Pure Pixel Index(PPI) is used to select pure pixels. The N-dimensional visualizer is used for assisting selection of four endmembers: vegetation, high-albedo objects, low-albedo objects and soil. The fraction images are derived to represent the abundance of the above four endmember. Impervious surface is estimated by analyzing high-albedo and low-albedo fraction images. QuickBird multi-spectral image is used to evaluate the accuracy of impervious surface extraction by different methods. Experimental results indicate that the accuracy of artificial neural network is higher than others, which means non-linear spectral mixture models is also effective to impervious area extraction, even better than linear models.

In order to increase the second harmonic generation conversion efficiency, the periodic optical superlattice was sandwiched with nonpoled and nonlinear media with finite and variable width, which can bring about the significant modulation for the fundamental wave with changing the thickness of the additional media. It is found that: the conversion efficiency increases or reduces when the energy of fundamental wave increases or reduces; it can be enhanced significantly when transmission resonance (the rate of transmission obtains 1) takes place; compared with the conventional sample (the optical superlattice embedded in infinite and nonlinear media), the second harmonic generation conversion efficiency is enhanced 1.5 times at most; for the practical sample (the optical superlattice embedded in air), it can be enhanced 3.3 times at most. So the second harmonic generation conversion efficiency can be enhanced apparently by the transmission resonance.

Full Maxwell-Bloch equations were solved by Predictor-Corrector method and finite- difference time-domain method to investigate spectrum of a sub-cycle laser pulse propagating in a dense Λ-type three-level atomic medium.It is found that both in a dilute medium and a dense medium,the spectrum is broaden due to the self-phase modulation(SPM); however,the width is much larger in the dense medium than that in a dilute medium because of NDD interaction.In the dense medium,the supercontinuum spectrum occurs due to SPM and NDD interaction,and the maximum frequency is 10 ωp.

A new modified form of the rapid numerical difference recurrence formula was proposed to calculate the nonlinear Shrdinger equation, with ites of the pulse self-steepening and the intrapulse Raman scattering.Compared with the traditional analytical results of solitons and the solutions of Split-step Fourier method, the results show that the modified rapid numerical difference recurrence algorithm is a fast and precise numerical method, which takes into account simultaneously both the effects of the chromatic dispersion and the nonlinear Kerr effect together along each fiber segment, and the effects of all the higher-order nonlinear items of the optical medium. By using this method, it was validated successfully to simulate the supercontinuum generated as a result of the optical fiber nonlinear effects and the group velocity dispersion effects together. The formula was proved that it can provide some new ideas of research for the simulation of the other nonlinear phenomena in the optical medium.

The asymmetrically clipped orthogonal frequency division multiplexing technique was applied to the atmospheric communication system,and the performance of atmospheric communication system was analyzed using asymmetrically clipped orthogonal frequency division multiplexing technique through atmospheric turbulence channels.On this basis,asymmetrically clipped optical orthogonal frequency division multiplexing modulation was compared with on/off keying modulation and DC optical orthogonal frequency division multiplexing modulation,which was simulated in the atmospheric turbulence channel.The simulation results show that the proposed scheme can satisfy the requirements of atmospheric laser communication system.

In the semiconductor optical amplifer,based on the theory of four wave mixing effect and refractive modulation-index generated by equivalent grating,dynamic changes of the signal light time delay with the outside controllable parameters,such as the injection current into semiconductor optical amplifier(SOA),the pump light power,the detuning frequency between pump light and signal light,were numerically simulated by the method of subsections model of SOA.Compared with the former simulation of slow light from the slow down factor view,this method has the advantage of associating the signal time delay with outside controllable parameters and providing the intuitional guidce for the experiment.Meanwhile,the method by subsections model of SOA has higher accuracy.Compared with the existing experimental results,it indicates that the numerically simulated results agree well with the measurements.

To generate millimeter-wave signal using optical method, a optical frequency multiplication (OFM) system was constructed with a phase modulator and a fiber periodic filter. The periodic filter was made of two polarizers and a segment of birefringent fiber. The theoretical analysis shows that in this system the unexpected odd/even harmonic components can be suppressed effectively by adjusting the birefrigent fiber delay. Furthermore, the desired electrical harmonics can be given prominence to the others by optimizing the phase modulation index. The simulation results demonstrate that the 60 GHz millimeter-wave signal can be effectively generated, and 2.5 Gbps NRZ signals can be correctly transmitted in this OFM system. The theoretical and simulation results also indicte that the OFM system based on the sinusoidal-signal phase modulation should be analyzed using the harmonious optical sidebands, instead of the sweeping frequency theory.

A linewidth broadening and high efficiency erbium-doped superfluorescent fiber source (SFS) by using the cascaded dual-backward pumped configuration is proposed. The effects of fiber length arrangement and pump power ratio on the output characteristics of the L-band SFS are investigated by simulation. The simulational results show that the linewidth broadening L-band SFS can be achieved by the proposed design.The spectral linewidth is over 15 nm broader compared with the conventional L-band SFS, up to 60 nm. Further more, the design gives the highest pump efficiency at pump power ratio of 0.5. An L-band SFS of 62 mW output power, 30.1% pump efficiency and 56.6 nm linewidth is experimentally achieved with 206 mW pump power of a 980 nm LD.

A hyper dimensional multiplexed chaos communication system was proposed based on nonlinear delayed feedbacks of electrooptical modulator.Different from traditional chaos communication system，the chaotic waveform is not produced by the semiconductor but by the electro-optical modulator.The scheme is advantageous of high nonlinear dimension，easily reproduced and manipulated accurately.The encryption method of signals modulating feedback delays and cross-correlation decryption method were illustrated.High rates of multiplexing and demultiplexing were numerically simulated for three channels.The main factor for error bit rate was analyzed.The influence of optical fiber on the multiplexed chaos communication system was further investigated.The research results indicate that the influence of optical fiber on chaos synchronization can be reduced significantly and the original signals can be recovered successfully as long as corresponding gain compensation schemes are adopted.

The ultraviolet upconversion luminescence properties of the Tm3+ ion sensitized by the Yb3+ ion were studied, in fluoride glass as excited by a 980 nm continuous wave diode laser. Red, blue, and even ultraviolet emissions were observed with ZrF4-BaF2-LaF3-A1F3-NaF glass, and it was also found that the peaks at 291, 347, 363, 454, 475, 643, 687 and 804 nm correspond to the transitions of 1I6→3H6, 1I6→3F4, 1D2→3H6, 1D2→3F4,1G4→3H16, G4→3F4, 3F3→3H6, and 3H4→3H6, respectively. According to the relationship between the pump power and the upconversion peaks intensity, 347 nm and 363 nm ultraviolet emissions are five photon and four photon processes respectively. The research results show that the ZrF4-BaF2-LaF3-A1F3-NaF glass of the Tm3+ ion sensitized by the Yb3+ ion can effectively transform the near-infrared waveband laser to the ultraviolet.

The 8 mol% Er3+-doped Yb2Ti2O7 powder, which was composed of single phase of Yb2Ti2O7, was prepared by sol-gel method. The green and red upconversion emissions of Er3+-doped Yb2Ti2O7 centered at about 524, 548 and 660 nm were detected by a 980 nm semiconductor laser diode excitation. The intensities of green and red upconversion emissions (Igreen, Ired) increased with increasing the pump power of laser diode. The two-photon absorption upconversion process was involved in the green and red upconversion emissions. With increasing the temperature of Er3+-doped Yb2Ti2O7 powder, Igreen and Ired decreased with the ratio of Igreen/Ired increasing, while the intensity ratio of green upconversion emissions centered at about 524 and 548 nm (I524/I548) increased. The result implies that the Er3+-doped Yb2Ti2O7 powder with the change of green upconversion emissions depending on the temperature of powder has a potential application for optical high temperature sensor.

The sensitivity expression of spatial-resolved diffuse reflectance to first-order scattering parameter was derived and numerically analyzed based on the P3 approximation of transport theory,and the influence of scattering parameter on the P3 approximation and diffusion approximation reflectance were also compared.The results show that:the sensitivity is distinct with that of the diffusion approximation in the region of about two transport mean free paths;the sensitivity to scattering parameter equals zero at the distance about four transport mean free paths which differs with the diffusion approximation;the optimum source-detector distance decreases with the increase of the scattering coefficient,and this change becomes more obvious with lower absorption.The research has theoretical significance to obtain absorption information of tissue with diffuse reflectance at the optimum source-detector distance.

In order to study the influence of the analyzer angle of polarization detection system on the backscatter characteristic of optical pulse in the backward detection device of optical pulse underwater, firstly, the depolarization by scattering was analyzed qualitatively when polarized light travelled in the water using the reflection and refraction theory on the scattering interface. Secondly, with linear polarized laser pulse as incident light, backscatter signals of optial pulse were detected experimentally at different analyzer angles. Finally, after comparing statistical average of leading edge, trailing edge, width, peakvalue, shape and area of optical pulse′s backscatter signal at different analyzer angles, the influence of analyze angle on the backscatter characteristic of optical pulse was analyzed. The result shows that, with the increase of analyzer angle, the backsctter singal′s width becomes narrow, the shape becomes gentle, the peakvalue and area are decreased. The leading edge and trailing edge move backward and the moving distance of leading edge is much larger than that of trailing edge. And the backscatter signal from nearby water area is restrained more effectively by the polarization detection system.

Based on the features of LD end-pumped Yb∶KGW laser crystal, the thermal model was established. Considering the convective heat transfer on the side faces, the temperature distribution, pumped-face distortion and thermal focal length under different pump power were obtained by solving anisotropy heat conductive equation through finite difference method. Under the pump power of 14 W, the highest temperature located at the center of the pumped face is 272.8 ℃,the highest distortion is 0.48 μm, and the thermal lens focal length at the pump spot is 4.8 cm.

To further investigate the red band of Nd∶YAG ceramic laser, an electro-optical Q-switched ceramic laser at the wavelength of 670 nm was studied. The Nd∶YAG ceramic crystal with 1.1at% Nd doping and dimensions of Φ3×50 mm2 was side-pumped by three laser diodes positioned with 120° respectively. The folding cavity parameters based on the ceramic thermal focal length were designed. The influences of thermal focal length in ceramic crystal and frequency crystal on spot radius were discussed. The LN crystal to electro-optical Q-switched and KTP crystal typeⅡphase matching for intracavity frequency-doubling were used to realize the laser. Under repetition rate of 1 000 Hz and single pulse pumped energy of 144 mJ, the output energy of 1.48 mJ and pulse width of 35 ns were obtained at 670 nm, and the optic-optic efficiency was 1.03%.

A supporting structure of segmented mirror for Fourier telescope field experimental system is proposed, and the fabrication and assemble sequence of each part are introduced. The segmented mirror, 6 meters in height and 5.5 meters in width, which is composed of 61 hexagon spherical sub-mirrors, is the biggest segmented mirror telescope used for energy collecting in China till now. The sub-mirror cell is designed in modular block so that it is interchangeable and its three FOD precision adjust can be realized. The supporting truss adopts fission structure for its assemble and transport to become easy. The ground base uses pre-embedded bar and beton. Each component joins with the ground base by bolts. In this way, the integer rigidity is ensured and the temperature adaptability is also satisfied. The experiment indicates that the stability of support structure is better than 0.075 mrad and the angle adjust accuracy is better than 0.05 mrad. The co-focus experiment of 8 segments which are already assembled indicates that superposition accuracy of spots center of mass is less than 20 mm. All the results above show that the proposed supporting structure satisfies the technical requirements of the field experimental system.

In order to test large and off-axis aspheric surfaces without the aid of other null optics,a new method subaperture stitching interferometry(SSI) is introduced.The basic principle of the technique is analysed,and the synthetical optimization stitching model and effective stitching algorithm are established based on homogeneous coordinates transformation and simultaneous least-squares method.The stitching software and prototype for testing of large aspheres by SSI are devised and developed.An off-axis asphere with the aperture of 376×188 mm2 is tested by this method.For the comparison and validation,the asphere is also tested by null compensation,the synthesized surface map is consistent to the entire surface map from the null test,and the difference of PV and RMS error between them is 0.047λ and 0.006λ,respectively.So it provides another quantitive measurement for testing large aspheric surfaces and off-axis aspheres besides null-compensation.

The measurement of the optical properties of highly scattering media was carried out based on optical coherence tomography(OCT).The impact of scattering coefficient and anisotropic factor on heterodyne efficiency factor were analyzed by the theoretical simulation and analog calculation.An algorithm was developed to extract the optical properties using Extended Huygens-Fresnel model under the case of the specular reflection.The reliable optical scattering properties experimental results were obtained by measuring the IntralipidTM solution of the different concentration.And the relationship between scattering coefficient and the correlation of IntralipidTM were studied.The fitting results show that the ralation of scattering coefficient and the concentration of IntralipidTM are approximately linearity.

In order to overcome the limitation of diffraction effect on the resolution of photoacoustic imaging system,the deconvolution method needs to be used for image retrieval.The acoustic lens imaging theory was introduced firstly,and the effect of the acoustic lens point spread function was simulated based on the resolution of acoustic lens-based imaging system and photoacoustic imaging restoration process of Wiener filter deconvolution.Using self-erected acoustic lens-based photoacoustic imaging system,the direct photoacoustic imaging pictures and the restored pictures from Wiener filter deconvolution of two dots of black adhesive tape with a distance of 4 mm and 3 mm are obtained and compared.The simulation and experimental results indicate that Wiener filter deconvolution can improve the resolution of photoacoustic imaging system.

Aiming at the regional homogeneity and processing of the edges in course of image fusion, a novel image fusion algorithm for visible and infrared images based on polyharmonic local sine transform is proposed. The polynomial u of the source images are fused with average in order to extract the global feature, and the residual v are fused in the field of Fourier sine transform to keep region homogeneity and extract details of the visible image. The polyharmonic local sine transform avoids the disadvantages of edge effect. Combing this advantage with the quickly decaying coefficients of the residuals, polyharmonic local sine transform is effective for image fusion. Experimental results show that the proposed algorithm improves the visual effect, and enhances the contrast and information entropy.

For infrared detection target signal uncertainty problem, a robust matched filter algorithm was presented. Based on Parceval theorem,the uncertain spectral model of infrared target signal was established by qualitative analyzing the target signal spectrum distortion source, and signal spectrum distortion associated factors were analyzed. The saddle-point solution for Signal-to-Noise of robust matched filtering was constructed using the minimax method, and the most negative signal spectrum was obtained with Genetic Algorithm of minimax bilevel optimization. Then transfer function of the robust matched filter was designed based on the classical matched filter. The simulation results show that Signal-to-Noise of robust matched filtering under low background noise can significantly eliminate the Signal-to-Noise over-sensitivity to the abnormal change of infrared input signal spectrum ,and spectral Signal-to-Noise compensatio in signal spectral is the highest to nenergy distortion, followed by correlation peak distortion, spectral related distortion to at least.

An effective method based on Hough transform, which is used to detect the fuzzy line fringes in image, was put forward. Taking advantages of the insensitivity feature of Hough transform to random noise and the properties of Fourier transform, the method can detect the fringes with the same period. The method can recognize the general feature of the fringes with the same period and do not recognize the single fringe. Using the proposed method, the influence of incomplete fringes to the measurement result can be efficiently avoided. In the application of the automatic interpretation of the holographic interferogram, the holographic fringes were automatically recognized and exactly counted. The experimental results show that the proposed method can avoid the influence of the uneven gray and the serious blur of fringes, and increase the test precision and efficiency of the digital holographic interferometry.

In digital holography,the reconstructed optical field phase generally has high spatial variation frequency,however,the space-bandwidth product is so limited that it is easy to introduce undersampling and to cause the existence of massive unreliable data points in the wrapped phase,which makes the least- squares phase unwrapping algorithm difficult.The principle of shearing interfermetry was introduced to refactor the digital holographic reconstructed optical field.By changing the phase gradient calculation of the reconstructed optical field, the impact of undersampling was reduced.The feasibility and validity of the new method were proved by the simulated calculation and experimental verification.And the results show that the phase obtained by the new method is closer to the experimental value and it has stronger resistance ability to undersampling than the traditional method.

By analyzing the existing assessment of image fusion quality, according to the characteristic of human visual system, a novel remote sensing image fusion quality assessment was proposed, which was based on the feature and structure similarity. Firstly, the spatial feature and spectral feature were extracted from the fused image and the original panchromatic (PAN) and multispectrum (MS) images, then the structure similarity of the spatial feature and spectral feature were involved to calculate the assessment index, so as to achieve the evaluation of image quality. The experimental results and objective evaluations indicate that, the evaluation of the proposed assessment is consistent with the subjective evaluation, which shows that the proposed method is an objective and effective assessment for the remote sensing image fusion quality metrics.

In consideration of the characteristic of polarization image which can enhance the target identification,a polarization system in visible light and the dual-tree complex wavelet transform(DTCWT) fusion algorithm of polarization image were proposed. In the low-pass sub-band of wavelet coefficients,the fuse rule based window energy was adopted,and in the high-pass sub-band,the rule of spatial frequency self-adaption weighted at window was selected. The setting of autocorrelation and the application of simulated annealing algorithm improved the robustness and adaptability of fusion. The results show that the fused image has good subjective quality and objective evaluation of performance which will enhance the target and background contrast.

Besides the adoption of range-gated to eliminate the backscatter, another way to advance the performance of underwater laser imaging is denoising of the images. Soft-morphological was introduced based on the analysis of the characteristic of noise in the underwater laser range-gated images. The multidirectional structures and polarized soft transformation were designed to build a OC filter. Discrete DE algorithm was used to search the best value of repletion degree aiming at the speckle index and edge energy. The experiments show that the proposed algorithm is robust because of the ability to filter the noise and protect the edge at the same time. When compared to Lee filter and nonlinear weighted averaging multidirectional generalized morphological filter, the performance of the proposed algorithm which is universal at a certain extent is better.

Because the existing robust watermarking algorithms cannot involve the research of robustness against dual print-and-scan process, a robust zero-watermarking algorithm against dual print-and-scan process was proposed. The algorithm was based on the stability of the numerical relationship between each block′s direct coefficient and the average of direct coefficients form all blocks in discrete cosine transformation domain. The image was divided into non-overlapping blocks, and each block was transformed with discrete cosine transformation. Zero-watermark sequence was produced by judging the numerical relationship between each block′s direct coefficient and the average of direct coefficients form all blocks. Some tests were carried out including its robustness against single print-and-scan process, conventional signal processing after single print-and-scan process, dual print-and-scan process, and conventional signal processing after dual print-and-scan process. The results show that the proposed algorithm has excellent robustness towards those above attacks.

To overcome shortcomings of current polarization image fusion methods, a novel false-color imagery fusion approach based on color transfer and clustering-based segmentation was proposed. Firstly, the images by polarization information analysis were mapped into HSI color space, thus the initial false-color fused image was obtained. Then, the fused image was transformed into YIQ color space, and the chromatic characteristics each channel was modified using those of the reference image. After that, the modified image was transformed into HSI color space, and the segmentation result of the image of degree of linear polarization using fuzzy C-means clustering algorithm was used to adjust its hue component and saturation component. Finally, the resultant image was transformed back into RGB color space, so the final fused image was obtained. Experimental results show that, the proposed method has the merit that the fused result is consistent with the habit of human vision perception, and makes the contrast of artificial targets within the cluttered background more outstanding.

Aiming at shortages of the mean shift tracking algorithm in dealing with the cases that the displacements of target between two successive frames are relatively large and the scales of target change quickly, and the poor adaptability of single feature to the changeable circumstance, an adaptive feature selection method is presented, to determine the most effective feature by analyzing the discriminative value of target and background. By representing the target model and the target candidate in terms of background weighted histogram and kernel weighted histogram respectively, and using the pyramid analysis technique, the pyramid mean shift tracking method is proposed to localize target via a coarse-to-fine way. Furthermore, a scale update mechanism is presented. Experimental results on various videos show that the proposed method can successfully cope with the cases such as high-speed moving target, scale variations, camera motion, partial occlusions, etc.