For large optical telescopes, the relative poses between the primary and secondary mirrors have very strict requirements. Due to the large quality of the primary mirror, the secondary mirror system is often designed as an adjustable mechanism with multiple degrees of freedom. It has an important impact on telescope imaging. As the diameter of the telescope continues to increase and the application scenarios continue to develop, the secondary mirror adjustment mechanism must not only ensure high precision, but also have a high load, and its design is becoming more and more challenging. In order to find a feasible solution for the adjustment mechanism of the secondary mirror of the large-aperture telescope, according to the development needs and different applications of the adjustment mechanism of the secondary mirror of the large-scale optical telescope, the different adjustment mechanisms of the secondary mirror are sorted, classified and compared. The advantages and disadvantages of the secondary mirror adjustment mechanism are summarized, and the future development of large-aperture telescopes is prospected.

Based on fractional Talbot effect, a kind of two-dimensional chessboard-like Talbot array illuminator was designed by using a pure phase grating. With Fresnel diffraction theory, the distribution of light field of a pure phase grating illuminated by a plane light wave was analyzed. The results of the analysis shows that two-dimensional chessboard-like Talbot array illuminator with a compression ratio of 2 can be obtained by using two step phase grating with phase modulation of(0,π/2). By binary optics technology, the phase grating was fabricated on an optical glass substrate. Under illumination of an expanded He–Ne laser, the chessboard-like high contrast bright and dark spots was obtained at the fractional Talbot distance and had efficiency of 83.2%. The experiment results agree well with the theoretical analysis. The results reported have high theoretical and practical value for the generation and application of array microbeam in the field of optical interconnection, optical communication and photoelectric hybrid processing.

A compact multi-spectral imaging system was designed by using the model of aperture segmentation in order to solve the problem of volume and complex optical path, and this system consists of the front optical element, array aperture, micro lens array, filter and image sensor, with the overall size less than the12.4×8×8(mm3). Some problems, including large vertical axis chromatic aberration, small field of view and, low utilization of light energy, were all overcame by using the array aperture. The system has a field of view of 20°, F/#0.4, focal length of 2.5mm and total length of 12.4mm in the wavelength range of 480～650nm. The single-channel MTF is greater than 0.35 at 153lp/mm, and the distortion is less than 0.23%. Compared with the existing technology, the system takes into account the pixel, spectral and time resolution, achieving a pixel resolution of 500×400 and a spectral resolution of 6.8nm for a single channel. Compared with the computational reconstruction imaging system, the direct imaging ensures the high temporal resolution of the system, and shows the characteristics of miniaturization, light weight and low cost.

In order to improve the utilization rate of high-end detector and enhance the applicability of Raman spectrometer to detect different substances at the same time, the integrated design of multiple wavelength excitation Raman spectrometer into a spectrometer will have important application value.A design method of dual-channel Raman spectrometer based on CT structure dual wavelength excitation (532nm and 785nm) is proposed.Based on the working characteristics of the detector, the design method adopted creates conditions for the simultaneous operation of the dual optical path structure, thus maximally increasing the spectrum measurement range. At the same time, the method also realizes the full utilization of the effective photosensitive surface pixels of the detector and the maximum realization of the hyperspectral resolution. The optimized results show that the resolution is 5cm-1 and the spectral range is 80~4200cm-1 for 532nm excitation wavelength.For 785nm excitation wavelength, the resolution is 3cm-1 and the spectral range is 200~2300cm-1.

In order to realize the space application of strontium atomic optical clocks, the use of permanent magnet Zeeman slower can effectively avoid the problems of high power consumption and large volume ratio caused by tower coils, and is conducive to the spatial development of optical clocks. The annular and cylindrical Zeeman slowers of strontium optical clocks based on permanent magnets have their own advantages and disadvantages in the deceleration principle, magnetic field construction and sample development. The deceleration efficiency of the two slowers can be measured by Doppler velocimetry, and the deceleration distribution curves of the two slowers are cumulatively distributed and compared. The experimental results show that both the two permanent magnet Zeeman slowers achieve a certain deceleration effect, but the volume and deceleration effect of the annular permanent magnet Zeeman slower are compared with the cylindrical permanent magnet Zeeman slower with a single x direction. It has been reduced by 60%, the weight has been reduced by 80%, and the efficiency of some areas of the deceleration effect is doubled, so the advantages are even more significant. The ring-shaped permanent magnet Zeeman slower was further used to capture three isotopes of strontium atoms, and the first-order capture of a miniaturized strontium atomic optical clock was completed, meeting the design requirements of a compact optical clock with zero power consumption and small size.

In order to solve the problem of primary imaging of visible light and infrared light, a wide-spectral aberration-corrected telescope system based on wedge-shaped beam splitter for simultaneous imaging of visible-infrared light was developed. The device uses a Cassegrain main system with a common aperture channel, and the beam splitter adopts a wedge plate design. By optimizing the design of the wedge angle of the beam splitter, the generation of reflected visible light ghost images can be suppressed, and the transmission is corrected to a certain extent. The aberration of infrared band imaging improves the resolution of the system. According to the design index, the Zemax software was used to simulate, and the results showed that the three imaging channels of the infrared light path, the visible light path and the infrared spectrum switching light path were better than λ/10 for the RMS wave aberration. In addition, the whole device system was analyzed and discussed separately, and the influence of manufacturing errors caused by the wedge beam splitter and system installation errors on the imaging quality was discussed, which served as a reference for the subsequent processing and installation of the device. Finally, the as-built telescope system was tested and analyzed with interferometer. The test results showed that the RMS of the main system was 0.093λ, the RMS of the visible light path was 0.120λ, and the imaging quality met the design requirements. The system adopts a Cassegrain common optical path, and uses a wedge-shaped plate to achieve synchronous imaging of visible light to mid-infrared light, which reduces the volume of the system and expands the function of the device.

In order to study the laser induction damage threshold (LIDT) of metal Al-protected, single-period and double-period enhanced reflective films under 532nm laser with 5ns pulse width, Comsol software is used for simulation and analysis. The tensile strength (film breaking stress) of SiO2 and Ta2O5 are introduced as the threshold condition in the simulation process, and the damage threshold values of three film systems are 0.318J/cm2, 1.325J/cm2 and 3.382J/cm2 respectively. By building an experimental platform to test the laser damage threshold, the 1-on-1 mode is adopted, and the laser damage point with a probability of 50% is selected as the damage threshold. The results of measured laser power are 0.288J/cm2, 1.232J/cm2 and 3.152J/cm2 respectively, which are close to the simulation results, indicating that the model based on tensile strength analysis is more in consistent with the actual damage, which provides a theoretical basis for the prediction of the threshold value of this kind of film system in practical application.

Laser inter-satellite link is used for high-precision measurement and high-speed data transmission between satellites. It is a crucial technology to build a space-based information network. Presently, foreign satellite systems such as communication, remote sensing, navigation, and relay have planned to deploy laser communication terminals. Laser inter-satellite link technology has developed from on-orbit demonstration and verification to large-scale networking applications. This paper briefly combs the development of laser inter-satellite link terminal technology of existing networking satellite systems in the world, and summarizes the technical strategy and terminal classification of laser inter-satellite link. The development status of laser inter-satellite link terminal technology in China and the main problems faced by the large-scale application are emphatically analyzed, and solutions to key and difficult technologies are considered; Finally, the future development trend of laser inter-satellite link terminals has prospected, especially the development trend of lightweight design and implementation, broadband, one to many optical antennas and networking technologies.

The non-linear characteristic of the light emitting diode can give rise to non-linear distortion of the light signal, aiming at this problem, the artificial neural network is introduced to the receiver end to realize the non-linear distortion suppression for the signals, and then the bit error rate of the visible light communication is reduced. The input electrical signal of the light emitting diode and the transformed electrical signal of the receiver are combined paired-wise data, then the paired-wise data is collected to train the artificial neural network, and learn non-linear distortion characteristic of electric-optical conversion, channel transmission and optical-electric conversion. As a result, the non-linear distortion of signals is estimated and suppressed with the neural network. Besides, EDA is adopted to search the super parameters of the neural network during the training phase, to reduce the training difficulty. Experimental results show that, the proposed method can effectively improve the performance of the visible light communication with different channel conditions.

Through binocular observation and matching psychophysical experiments, the visual evaluations of metallic paints under two geometric conditions were carried out, and the visual average data of metallic paints were linearly fitted by the least squares method with the prediction of color appearance model based on the bidirectional reflectance distribution function (BRDF) data and the PR715 spectrophotometer measurements. The experimental results show that BRDF data can be used to predict the color appearance of metallic paints under different geometric conditions. For prediction of the three-attribute, the prediction accuracy of the color appearance model based on BRDF data is not as good as that based on PR715 spectroradiometer data in terms of the brightness and chroma, while the prediction accuracy of the color appearance model based on BRDF data is higher than that based on PR715 spectroradiometer data for hue angles.

A differential measurement configuration with single cylindrical focusing lens for thermal lens technique is proposed and a theoretical model based on the optical phase shift theory and the ABCD rule for Gaussian beam propagation is developed to describe the thermal lens amplitude. The optimal configuration condition and the theoretical limit of measurement sensitivity are derived. The theoretical results show that this measurement configuration can obtain the same measurement sensitivity as the measurement configuration with double cylindrical focusing lens. The influences of configuration parameters on the light intensity distribution on the detection plane and the thermal lens amplitude are analyzed by the theoretical model, as well as by the finite element method. The both reach a good agreement, which proves the validity of differential measurement configuration with single cylindrical focusing lens and the correctness of the theoretical model.

Pose estimation is one of the most popular research directions in the field of intelligent and automatic control at present, and it has many applications in the fields of unmanned vehicle, intelligent robotic arm, and household robot. However, most of the traditional methods are complicated in calculation and difficult in real-time computing. A pose calculation scheme is proposed, which uses convolution neural network to compress the size and extract the information of the image input of binocular camera, and carries out regression learning on the feature vectors through the bidirectional long-short time neural network and the standard results calculated by lidar. The trained deep learning solution has a certain improvement in accuracy and speed compared with the traditional solution.

Based on the difference in spatial coherence between target echo and multi-scattering clutter in underwater lidar system, a spiral phase plate and an additional ring mask plate were proposed to form a spatial filter to separate the target reflected signal from the multi-scattered clutters. Due to the good spatial coherence of the target reflected signal, vortex light will be formed after the target reflected signal passing through the spiral phase plate, while the spatial coherence of scattering clutter will be destroyed by multiple scattering and presented a Gaussian-like distribution according to the scattering law. The use of ring-shape mask plate only allows vortex light to pass through and reach the detector, thereby reducing the effect of scattering on ranging. The above-mentioned spatial filtering device is applied to radar-lidar and pulse ranging underwater lidar systems, which can greatly improve the ranging accuracy of targets in turbid water.

In a fuel element product, fuel particles are randomly distributed in non-metallic materials. In order to measure the distribution uniformity of fuel particles, a heat map method based on X-ray Micro-CT three-dimensional image region statistics is proposed to characterize the uniformity. Firstly, the Micro-CT method is used to conduct three-dimensional scanning and imaging for such products, then the fuel particles are extracted through the three-dimensional image segmentation method, and then the three-dimensional image is divided into three-dimensional regions, and then the volume proportion of fuel particles in each region is counted, and finally characterized by the heat map method. The Micro-CT is used to detect and process the fuel element, and the statistical heat map is used to characterize the distribution uniformity. The distribution uniformity of fuel particles in the fuel element is visually characterized.

An optical lens performance evaluation method based on multi-band regional mutual information (MRMI) is proposed. Object plane and its image plane is decomposed into multiple frequency band by wavelet transform. The retina is sensitive to different signals at different frequencies in human visual system（HVS）. So the wavelet transform simulates this process. Regional mutual information (RMI) is calculated for each sub-band according to information theory. The weighted coefficients of RMI in each sub-band are calculated by integral ratio of wavelet filter and MTF in spatial frequency domain. Then MRMI is done by weighted synthesis of RMI. Six lenses with different Abbe number and refractive index on the market were selected for measurement experiments, and the imaging MRMI value of each lens was obtained, which quantitatively reflected the imaging performance of the lens and was consistent with the visual perception of human eyes.

The dynamic resolution detection system of aerial camera is used to simulate the motion state of ground object relative to aircraft flight in the laboratory, realizing the resolution detection of aerial camera in dynamic photography. In this paper, the virtual instrument technology and motion control technology of dynamic resolution detection system based on LABVIEW are mainly studied, the movement curve of dynamic target simulation system is designed according to the image motion velocity equation of aerial camera, the friction fluctuation compensation algorithm is used to improve the conventional PID control algorithm to complete the linear motor control with high precision at uniform velocity section of the motion curve. The system realizes the speed range of uniform linear motion of dynamic graphics from 10~1000mm/s, the instantaneous velocity error of uniform velocity section is less than 3mm/s and the average velocity fluctuation error is less than 0.1%.

The inter-frame differential method is the classical method for motion target detection, but it is difficult to detect a motion target by using this method when the motion target speed is as low as a certain value. Against at the limitation of its detectable target speed, a dynamic frame difference method is proposed based on the inter-frame difference method. Combining the characteristics of FPGA processing, the FPGA design and implementation of dynamic frame difference algorithm is carried out, and a new real-time motion target detection system is constructed. The system is capable of adaptively adjusting the frame spacing used for differential images according to the target motion, which is applicable to the detection of both faster and lower-speed moving targets, expanding the speed range of moving targets that can be detected by the system. Regional division of DDR3 ports solves the problem of multi-frame storage of images on FPGA with limited resources. Several sets of experimental results confirm the effectiveness and real-time performance of the dynamic frame difference method and its FPGA design module. Compared with other systems, this system has a clear advantage in detecting low-speed moving targets and can handle color video streams of 1080p@30Hz.

In order to remove the Gaussian noise in the infrared image while preserving and restoring the textures, edges and detailed features of the image, an infrared image denoising method based on dual density complex wavelet and coefficient correlation is proposed. This method makes the most of the advantages of dual density dual tree complex wavelet in image processing: translation invariance of image information, multi-directional selectivity of image texture and detail, etc. Based on the assumption of the distribution of image wavelet coefficients, according to the correlation between the current wavelet coefficient and its parent and child wavelet coefficients, Bayesian estimation is made for the noiseless wavelet coefficients to restore the noiseless infrared image. Finally, the denoised image is subjected to guided filter, so as to remove the ripple effect. Experimental data show that this method is superior to some existing algorithms in EPI, FSIM and visual effect of image, which proves that this method has better performance in noise removal, texture edge preservation and restoration.

A major critical challenge for hill climbing algorithm is the problem of poor accuracy and slow speed in the automatic focusing. Here, a partition search algorithm combining self-adaptive step and two-step search are therefore proposed to solve this challenge. This method divides the focus evaluation curve into gentle and steep region according to the value of focus evaluation function and curve shape. In gentle region, self-adaptive step is utilized to increase the auto-focus search speed. Then two-step search method is explored in steep areas to avoid focusing errors caused by false peak interference and improve auto-focus accuracy. The experimental results show that the focus speed of the partition search algorithm is increased by 11.25% compared with the hill climbing algorithm, and it has better accuracy and anti-interference ability.

The multispectral dimension reduction method based on principal component and weighted principal component realizes the mutual conversion between multispectral data and low dimensional spatial data. However, the low dimensional spatial data contains a large number of negative values and cannot be connected with chromaticity space such as CIELAB, which brings difficulties to the follow-up research of spectral color replication; Establish the conversion from XYZ tristimulus to multispectral data, and retain more color information in the process of reducing the dimension of multispectral data to XYZ tristimulus value; The corresponding relationship between tristimulus values and three-dimensional space obtained by dimensionality reduction through weighted principal components is established through second-order polynomial regression, and the mutual conversion relationship between multispectral data and tristimulus values is realized; In different training samples and different test samples, compared with the principal component and weighted principal component, the proposed method improves the accuracy of colorimetric reconstruction under various lighting conditions, and can be better applied to the high fidelity dimensionality reduction and compression of multispectral images.