The relay optical system was widely used in optical systems such as infrared spectral imaging system, light field imaging system, optical microscopy system, polarization interference imaging system, compound eye imaging system, ring-belt panoramic optical system, multi-scale imaging system and head-mounted enhanced display system, etc, which can link up, match pupil and deflect optical systems. The structures of the existing relay optical system were studied. The design method of telecentric off-axis three-mirror optical system with front aperture and the description method of free-form surface were introduced. According to the design parameters, the telecentric relay optical system with broadband and real entrance pupil was completed. The system was an off-axis three-mirror optical system structure, and each mirror was a free-form surface described by XY polynomial. The results of CODEV software simulation show that the MTF of the system is close to the diffraction limit, the distortion is less than 1%, and the imaging quality is good with the working spectral range of 0.4-5.0 μm, f'=400 mm, F/3 and 2ω=8°.

A bi-axial half-butterfly flexure hinge for an fast steering mirror （FSM） was presented to adapt high stability accuracy of beam-pointing control performance in laser weapon systems. According to the requirements of reciprocating movements and high bandwidth provided for the FSM, the solid model of the bi-axial half-butterfly flexure hinge was designed. By applying Castigliano’s displacement theorem, the numerical model was simplified and deduced. Furthermore, to quantify the numerical model, natural frequencies of the finite-element analysis and experiments were carried out, of which the results were compared with the analytic solutions. The experiment results show that the in-plane natural frequency is 165.29 Hz. The comparison shows that the error between numerical analytic and experimentation is 1.3%, and the error between FEA and experimentation is 3.2%. It is proven that the bi-axial half-butterfly flexure hinge is an appropriate structure as a guide mechanism for an FSM system.

With the development of metal powder 3D printing technology, how to accurately extract the particle size and spheroidization rate information of powder particles from microscopic images has gained much more importance. In this paper, a particle auto-statistics and measurement system on microscopic imaging of the spherical powder was presented, based on one deep learning framework—Mask R-CNN. The proposed model can efficiently detect more than 1 000 particles in a microscopy image, even under the existence of many occlusion particles, and provide statistical results of particle size distribution, degree of sphericity and spheroidization ratio, simultaneously. Compared with traditional image segmentation method, the particle recognition accuracy was improved by 23.6%. Moreover, smaller particles that stuck on big particles can be recognized, according to the comparison in particle size distribution between proposed method and the laser diffraction technique.

A simulation model of the ultraviolet (UV) radiance of the aircraft plume was established based on the improved spherical harmonic discrete coordinate method (SHDOM). Based on the thermal radiation of the plume, the modified model took into account the chemiluminescence of gas molecules and multiple scattering source caused by the average optical properties of cluster particles in the plume. On account of the basic radiation transfer equation, the UV radiation of the plume at different observation angles and medium distribution was numerically calculated in the spherical harmonic discrete coordinate system, and the space distribution images simulation of high resolution three-dimensional (3-D) plume radiance was presented. At the same time, the gray-scale co-occurrence matrix (GLCM) algorithm was used to obtain the gray-scale map of the plume radiation distribution. Combined with two-dimensional super resolution rotational invariant subspace algorithm, α-shape method and the rubber band algorithm, the central position and peak contour curve of the plume radiation was calculated, and the accurate extraction of the plume radiance feature was obtained, which provide an important reference and approach for the high resolution radiation image feature recognition of aircraft target and plume.

With the rapid development of three-dimensional point cloud acquisition tools such as Light Detection And Ranging (LiDAR), the semantic information of point clouds has become more and more important in computer vision, intelligent driving, remote sensing mapping and smart cities. Aiming at the limitations of the point cloud semantic segmentation method based on segmented block feature matching, such as cannot handle the over-segmentation and under-segmentation, the semantic segmentation accuracy of street trees and rods is low, a point cloud semantic segmentation method of street trees and rods based on the segmented blocks merging strategy was proposed, which could merge interested segmented blocks after density-based spatial clustering of applications with noise (DBSCAN) clustering segmentation, and the merged objects were classified by calculating their multi-dimensional geometric features, then the semantic segmentation results were optimized by the interpolation optimization algorithm, and finally the semantic segmentation of street trees and rods in the urban road environment was realized. The experimental results show that the method proposed can improve the recall rate and semantic segmentation accuracy of point cloud data such as street trees and rods in an urban road environment to more than 89.9%. The semantic segmentation method based on segmentation merging can well solve the problem of low accuracy of semantic segmentation of street trees and rods under urban roads. This method is of great significance for the research of three-dimensional scene perception and other problems.

The conventional Iterative Closest Point(ICP) matching algorithm for laser point clouds had problems of slow convergence and poor robustness, therefore, a point clouds registration method combining multiple optimization methods was proposed. Firstly, point clouds were de-sampled using voxel grid filtering and key points were extracted by ISS operator, then feature extraction algorithm was performed to obtain Fast Point Feature Histograms(FPFH) features of key points, and the multi-core and multi-thread OpenMP parallel processing mode was operated to improve the speed of feature extraction. Then, based on the extracted FPFH features, the Sample Consistency Initial Alignment(SAC-IA) algorithm was used for coarse registration of similar feature points to obtain initial transformation matrix between point clouds sets. Finally, the ICP algorithm was used for fine registration, and the K-D tree nearest neighbor search method optimized by Best Bin First(BBF) was used to accelerate the search speed of corresponding point pairs, and dynamic threshold was set to eliminate the wrong corresponding point pairs, so as to improve the speed and accuracy of point clouds registration. Experimental research on two sets of point clouds shows that the optimized registration algorithm has obvious speed advantages and improves the registration accuracy.

In view of the sparsity and spatial discrete distribution of lidar point cloud, a graph convolution feature extraction module was designed by combining voxel partition and graph representation, and a 3D lidar point cloud object detection algorithm in view of voxel based graph convolution neural network was proposed. By eliminating the computational redundancy of the traditional 3D convolution neural network, this method not only improved the object detection ability of the network, but also improved the analysis ability of the point cloud topology information. Compared with the baseline network, the detection performance of vehicle, pedestrian and cyclist 3D object detection and bird’s eye view object detection tasks in KITTI public dataset were improved greatly, especially improved with 13.75% precision in 3D object detection task of vehicle at maximal. Experimental results show that the proposed method improves the detection performance of the network and the learning ability of data topological relationship via graph convolution feature extraction module, which provides a new method for 3D point cloud object detection task.

In order to improve the accuracy and adaptability of the LiDAR point cloud filtering algorithm, an improved moving surface filtering algorithm had proposed. The boundary points of the grid were used to construct the surface constraint conditions to test whether all the building points in the grid. The area fitting was used to solve the terrain fluctuations. The Gaussian Mixture Model (GMM) in machine learning was introduced to filter and classify the terrain undulations, and the seed points in the moving surface were used as the target points in the clustering algorithm to participate in the classification learning. The experimental data was the self-test area of radar flight. The filtering effect of the self-test area was tested and judged with random sampling. At the same time, the Kappa coefficient was added as the test method to test the accuracy of the GMM algorithm on the basis of the three types of error test methods. Compared with the pedigree clustering classification algorithm, it is proved that the proposed algorithm can achieve better filtering effect.

A synthetic aperture radar (SAR) target recognition method based on image blocking and matching was proposed. The tested SAR image was blocked into four patches, which described the local regions of the target, respectively. For each SAR image patch, the monogenic signal was employed to construct a feature vector, which described its time-frequency distribution and local details. The monogenic signal decomposed the input image from amplitude, phase, and local orientation. Therefore, it could reflect the local variations in the image so providing more reference information for the analysis of target changes under the extended operating conditions. For the 4 feature vectors, the sparse representation-based classification (SRC) was used for classification and produce the corresponding reconstruction error vectors. Accordingly, based on the linear weighting fusion, the random weight matrix was constructed for analysis. For the results from different weight vectors, an effective decision variable was defined based on statistical analysis. By comparison of the decision values of different classes, the target label of the test sample could be decided. The proposed method made sufficient analysis of the uncertainties about the operating conditions during SAR image measurement, an optimal decision was made based on statistical analysis. Experiments were set up and conducted on the MSTAR dataset including one standard operating condition and three extended operating conditions. Compared with several present methods, the results confirmed the validity of the proposed method.

The method of active image quality reconstruction adaptive camouflage based on flexible display technology is to use flexible display devices to combine spectrum transfer technology and active image quality reconstruction technology to achieve the change, transfer and selective distribution of the target surface spectral radiation characteristics. In this design, the optical characteristics of the target could be changed by modulating the optical characteristic parameters of the target surface, and the background image could be captured in real time during the activity of the target and displayed on the flexible display. The use of active image quality reconstruction, flexible display and emissivity control layer achieves the purpose of modulating the infrared radiation intensity of the target and effectively segmenting the target heat map, so as to achieve a high degree of integration with the surrounding natural environment throughout the weather and the whole process, and the spectrum of the target surface. The distribution characteristics do not change with the change of the detection direction, achieve the effect of target camouflage. Compared with the method of changing the physical structure characteristics of the target surface, this technique makes the target environment more adaptable and more survivability, and easy to implement.

Infrared thermal imagers are widely used in security, night vision and infrared temperature measurement with the advantages of target recognition under all-weather conditions. However, the imaging quality is restricted by the quality of ZnS infrared optical lens. The annealed ZnS bulk was fabricated through Three-Temperature zone Gradient Chemical Vapor Deposition (TTG-CVD) furnace. The ZnS bulk with cubic sphalerite structure was characterized by X-ray diffraction. No hexagonal wurtzite structure was detected. It indicates that ZnS bulk possessing optically isotropic property, which can meet the design requirements in lens. The average transmittance of ZnS bulk was measured as 71.6% in the long-wavelength infrared band of 8-12 μm. The refractive index uniformity of ZnS bulk was measured as 1.94 × 10-5 at 1.06 μm. As encouraged by the above optical parameters, ZnS infrared optical lens was further produced by adopting optically cold-mechanical process and single-point diamond turning techniques. When the spatial frequency of ZnS infrared optical lens was 20 lp/mm, the modulation transfer function (MTF) of the half or 0.707 field of view was close to the diffraction limit. The root mean square value (RMS) of the diffuse speckle caused by aberration in the central field and 0.707 field of view was less than 20 μm in the pixel size. Meanwhile, the system distortion was less than 1% in the infrared imaging system. It shows TTG-CVD based ZnS crystal is promising for infrared applications.

The infrared thermal imager can monitor the target temperature, which plays the role of accident warning and location confirmation, large-scale human temperature screening and so on. Due to the temperature drift caused by the change of ambient temperature and infrared radiation absorption, most of the infrared thermal imagers for temperature measurement need blackbody for real-time calibration, but the blackbody-based infrared thermal imagers are limited by the fixed scene and poor portability. To solve this problem, a temperature calibration and compensation method without blackbody was proposed. By deducing the principle of infrared temperature measurement, the prior relation between target temperature and radiation quantity was obtained with multiple blackbodies calibration, and for addressing temperature drift caused by the internal structure of the detector, the temperature compensation was realized by non-linear modeling based on Newton's cooling law. The experimental results show that the proposed methods have the long-term stability to keep the relative error of temperature measurement within 0.9%, reduce the average relative error by 64%, and realize the portable, real-time, stable and high-precision temperature measurement of miniaturized infrared thermal imager.

CFRP/Al honeycomb panel components have been used in harsh environment for a long time, which is easy to cause debonding and water accumulation and other defects related to operation safety. Taking CFRP/Al honeycomb composite plate with debonding and water accumulation defects as the test sample, the halogen lamp driven by Barker coded modulation pulse compression signal was used as the external excitation source, and the infrared thermal image sequence of the sample surface was collected by infrared camera. Three dimensional matched filter was designed and used to process the acquired image sequence. With 4 different quantization methods, 8 thermal wave response results of debonding and water accumulation defects were obtained, and these results were compared and analyzed from multiple aspects and the signal-to-noise ratio (SNR) was evaluated. The results show that Barker coded excitation has the advantages of simple modulation and easy implementation, and it can effectively suppress the background noise of infrared thermal image and improve the SNR by combining with three dimensional matched filter, and can effectively detect the internal debonding and water accumulation defects of composite materials under the condition of low-power thermal excitation.

In order to meet the requirements of high-precision Noise Equivalent Spectral Radiance calibration of infrared remote sensors, the development and manufacturing process of the infrared integrating sphere radiation source were improved on the basis of the original design to meet the vacuum cryogenic use requirements. The integrating sphere radiation source used 8 sets of carbon fiber quartz electric heating tubes as the infrared radiation medium to achieve a working band coverage of 3-15 μm, and the adjustable radiation dynamic range was doubled. The radiation calibration reflective light path was designed, and the standard cavity blackbody radiation source was compared and measured to realize the radiation calibration of the infrared integrating sphere radiation source under vacuum and low temperature conditions. The calibration results show that the uniformity of the surface within the normal Ф200 mm of the infrared integrating sphere radiation source is 99.75%, the angular uniformity within ±10° is 99.81%, and the instability is 0.05%. It is the realization of the infrared integrating sphere radiation source spectral radiance output and color temperature near-linear adjustable function, 5 μm and 10 μm radiance adjustable range are up to 12.8 and 1.6 μW/(cm2·sr·nm).

Theoretical and experimental research on annular drilling of injector micro-hole by using the femtosecond laser micromachining system. The 06Cr19Ni10 stainless steel was used as the target material, and the orthogonal experiments with 5 factor and 5 level were designed based on L25(55) orthogonal table. The significance level of the influence of laser power, repetition frequency, defocus, scanning speed and scanning times on micro-hole processing was analyzed and the formation and evolution rule of micro-hole were explored. Then, the influence of various parameters on the micro-hole accuracy and topography was explored. Finally, the relatively optimal laser processing parameters were as follows: laser power was 1.0 W, repetition frequency was 9.0 kHz, defocus distance was 200 μm, scanning speed was 1.0 mm/s, scanning times was 40 times. In addition, based on BP neural network, a mapping model with the above five parameters as input and micro-hole entrance and exit aperture as output was set up. The results show that the prediction error of the relationship model is less than 7.6% by iterative training and verification of orthogonal experimental data.

In order to realize the engineering application of the branch-cut method in the phase diagram unwrapping of laser speckle pattern interferometry, and to solve the problems of dense branch lines and slow calculation speed caused by the interference of external light, the degradation of laser performance and the undersampling of local points taken by cameras, an optimization and improvement scheme based on Goldstein branch-cut method was proposed. The residual points were regarded as "electrons" with positive and negative unit electric quantities, guided by electromagnetic force, and eliminated by phase smoothing or increasing phase jump, thus reducing the number of branch lines. At the same time, GPU parallel computing technology was adopted to improve the processing speed of high-resolution and large-size phase diagram. The simulation experiment and actual measurement data show that the unwrapping diagram quality is better by the improved branch-cut method. For the 5 million pixel speckle phase diagram, more than 98% of the residual points can be eliminated by electromagnetic force, and the branch lines can be reduced by more than 90%. The processing time can be reduced from the previous 15 seconds to 1.5 seconds, which meets the engineering application requirements of high quality and quick unwrapping by branch-cut method.

The experiment result of gain performance of JG2 Nd:glass and the output of whole furniture was introduced which would be applied in 400 mm aperture slab amplifiers from high power inertial confinement fusion laser equipment. A 4×2 combined slab amplifier system of three-pieces-length and 400 mm-aperture was used to study the gain function. The gain and output experiment results of JG2 Nd: glass show that the small signal gain coefficient reaches 5.37%/cm, and the gain multiple of small signal is 1.284 times/pass when the system operating voltage is 31 kV. The thermal recovery time is about 2 h after cooling with clean and dry gas at a velocity of 0.3 m/s. Besides, experimental results acquired by Integration-Test-Bed (ITB) laser facility show that the optimal combination of JG2 and N41 Nd: glass reaches the maximum output energy of 21.3 kJ/1053 nm, and has been operated stably for more than 500 shots, without any faults such as abnormal cladding layer and material body damage.

A satellite-ground laser time transfer (LTT) with high repetition rate will be carried out in the China’s space station. The onboard laser detector intends to adopt a fixed gated opening mode, which puts forward the control requirements of high repetition rate, high precision and high real time for laser emission epoch on the ground station. Based on the principle of range gate in satellite laser ranging (SLR), a precise control method of laser firing signal in satellite-ground LTT with high repetition rate was proposed, so that the uplink laser pulse could reach the detector within a short time after the onboard gate signal, which greatly reduced the background noise interference. This method could be implemented in a field programmable gate array (FPGA), which had the advantages of repetition rate greater than 10 kHz, control accuracy of 5 ns (200 MHz clock) and simple software interaction, etc. Combined its theoretical calculation accuracy and nanosecond jitter of the diode-pumped picosecond laser, the final realization accuracy of laser emission epoch was expected to be within 10 ns. It could meet control requirements of the laser emission epoch of the China’s space station LTT project and provide technical support for other LTT projects.

The laser emission two-axis turntable is a device that takes the turntable as the main body, carrying the laser transmission and emission functions. The turntable is required to have good performance in the temperature range of -40-55 ℃. To better understand the impact of ambient temperature on the performance of the turntable, the structure of the laser emitting two-axis turntable was introduced, and the finite element method was used to analyze and study the static and dynamic characteristics of the turntable under different ambient temperatures. The effect of the ambient temperature on the accuracy of the shafting was obtained from the result of the deformation of the turntable after the static analysis. According to the analysis, under the condition of 22 ℃ turntable performance parameters as the benchmark, within the temperature index range, at room temperature 22 ℃, the static characteristics of the turntable and the shafting accuracy were the best. The farther the ambient temperature deviated from 22 ℃, the worse the static characteristics and shafting accuracy. When the temperature changed by 10 ℃, the maximum deformation of the turntable changed by about 0.529 mm, and the maximum stress value changed by about 18.418 MPa; the verticality error between the azimuth axis and the pitch axis changed by about 4.715″, and the verticality error between the azimuth axis and the base support surface changed about 4.649″; Azimuth shaft system wobble error changed about 0.22″, pitch shaft system wobble error: (1) The temperature was below 22 ℃, the change was about 0.33″; (2) The temperature was above 22 ℃, the change was about 0.569″. When changing with temperature, the first 6 modal frequencies of the turntable had different laws, but the frequency change rate did not exceed 2%. The analysis results show that the ambient temperature has a greater impact on the static characteristics and shafting accuracy of this type of turntable, but has a small impact on the dynamic characteristics. The conclusion is consistent with engineering experience.

The theoretical simulation of the extension structure of high-power GaN-based laser diodes is of great significance to improve the photoelectric performance of GaN-based laser diodes. A green laser diode extension structure with an n-side dual-wave conductor structure was designed. The effect of indium parts in the n-InxGa1-xN waveguide layer on its photoelectronic performance in laser extension structure was discussed. And the mechanism of the n-InxGa1-xN waveguide layer on the photoelectronic performance of laser diode was clarified. The results showed that when the indium part of the n-side InxGa1-xN waveguide layer was 0.07, the photon loss was minimal, and the threshold current was the lowest. When the indium part of the n-side waveguide layer was high or low, photon loss and operating voltage were increased, and meanwhile, the output power of the laser diode was reduced. Therefore, by regulating indium parts in the n-InxGa1-xN waveguide layer and controlling the optical field distribution of the outer layer, the photon loss was reduced by 0.2 cm-1, and the threshold current was reduced by 193.49 mA to 115.98 mA, and the operating voltage was reduced, which increased the output power and electro-optical conversion efficiency of the laser diode, increased the laser output power to 234.95 mW at 6 kA/cm2. The n-side dual-waveguide structure design provides theoretical guidance and data support for the preparation of high-power green laser diodes.

Safety is essential to deep mining operations. To monitor and detect the safety condition of surrounding rock of roadway and deep coal mining site, many methods obtain competitive results by monitoring targets situation based on laser scanning spacial information. The Hyperspectral LiDAR (HSL) technology can acquire spacial and spectral data for deep mine safety detection and further fine structure analysis. The exact coal/rock classification is the basis of detection and analysis. While, in on-site operation, HSL signals are susceptible to instrument attributes and environmental factors, and need calibration for further classification application. However, due to serious dust pollution in deep coal mines, conventional calibrations are hard to achieve the desired results. To address this issue, a new method was proposed to classify coal/rock without calibration. First, the new feature values, waveform entropy (WE) and joint skewness-kurtosis figure (JSKF), were extracted from coal/rock samples based on HSL measurements. Then, the coal/rock classification tests were conducted with random forest (RF) and support vector machine (SVM) classifiers. Additionally, the spectral properties of different wavebands were evaluated by spectral segmentation test and the classification performances were optimized further by selecting specific channels. The results show that the proposed method can achieve excellent classification accuracy for coal/rock without calibration.

In order to study the influence of the thermal accumulation effect of the multi-pulse nanosecond laser on the cladding process of the boron(B) doped silicon(Si) nano-film, the single-temperature model and the three-dimensional finite element method were used to numerically analyze the distribution of temperature field during the interaction process between the laser and the Si film, then the law of temperature field change under multi-pulse coupling was obtained. Compared with a single pulse, the simulation results of the multi-pulse laser action shows that the peak temperature has increased 3.2%, the size of the molten pool has enlarged 18.75%, and the range of the heat-affected zone has also significantly increased; after the laser irradiation, the surface temperature of the cladding layer drops, while the substrate temperature will continue to rise. The multi-pulse heat accumulation effect provides favorable conditions for the B diffusion in the Si nano-film. Finally, through single-pulse and multi-pulse laser cladding experiments, the different conditions of the cladding layers were analyzed, and the general law of the B diffusion assisted by laser cladding was obtained. The technology of laser-assisted B doped Si nano-film will provide the foundation for the applications in semiconductor devices.

In order to obtain microwave signal with high spectral purity, low phase noise and flexible tunability, a novel approach to achieving a tunable optoelectronic oscillator (OEO) which was based on optically injected semiconductor laser and subharmonic microwave modulation for microwave signal generation was proposed and experimentally demonstrated. The fundamental concepts for realizing the OEO were based on the wavelength-selective amplification effect and the period-one(P1) oscillation state of optically injected semiconductor laser. The frequency stability, side-mode-suppression ratio and spectral purity of the generated microwave signal could be optimized by introducing subharmonic microwave modulation via a phase modulator in the OEO loop. The experimental results show that the central frequency of the microwave signal generated by the proposed OEO could be tuned from 12 GHz to 18 GHz, and output power of the generated signal was more than 5 dBm. At the same time, the generated signal had a side-mode-suppression ratio of 51 dB and a 3 dB bandwidth of 100 kHz. Finally, the phase noise of the measured microwave signal could be optimized to -78 dBc/Hz and -109 dBc/Hz at 100 Hz and 10 kHz frequency offset by introducing subharmonic microwave modulation in the system, respectively. Furthermore, the tunable frequency range of the generated signal was restricted by the operating bandwidths of the optic-electronic devices which were utilized in the system. A higher frequency of the generated microwave signal could be achieved by using the devices with larger bandwidths in the OEO loop.

Aiming at the problem that the quality of the data stream transmission path in optical fiber network communication affected the utilization of network resources, an improved data transmission path optimization machine learning algorithm was proposed. Firstly, the machine learning was used to complete the preprocessing of the initial data, the data feature information was obtained, and the data stream classification was completed. Based on the analysis of the data flow within the optical fiber span, a cluster group was constructed to complete the adjustment of the data path and realize the full use of network resources. Secondly, the optimization of the cluster analysis was completed by taking the similarity matrix containing the characteristic parameters as the constraint condition. The similarity matrix was ​​established according to the data characteristic parameters, and the function mapping relationship was established between the characteristic parameters and the data flow type of the communication path. Finally, the kernel function was used to optimize the transmission path to realize the optimization of the network transmission path. The experiment optimized the path for a network containing multiple fiber spans, and compared it with the traditional K-means clustering algorithm. The ratio of the 6 different data streams in the test can fully reflect the data communication status under different conditions. The experimental results show that the classification accuracy of the algorithm is 94.6%, the average execution time is 12.8 s, and the average cluster change degree is 31.3%. The classification accuracy of the traditional K-means clustering algorithm is 84.6%, the average execution time is 20.8 s, and the average clustering change is 46.2%. The convergence time of this algorithm is also better than that of traditional algorithms, and it has higher accuracy and real-time performance in network data transmission.

As a precise optical component, microlens array has applications in fields as optical information processing, optical sensing, optical computing, optical communications and high sensitivity imaging. Researchers have developed many advanced fabrication techniques, some of which already realized the preparation of the microlens array with required geometries, profile and optical properties. However, it would be extremely difficult to achieve a compact packing as such 3D micro-manufacturing techniques are hard to control. A novel rapid and low-cost microfluidic-manipulation based technique was proposed to fabricate high-filling-factor microlens array. A brief demonstration of the fabrication was given, which had excellence of suited to volume production and significant productivity boost. Meanwhile, the microlens arrays of three different properties were produced, which were realized by adjusting the size of the array of micro-posts whose sizes were 300, 500, 700 μm in diameter, respectively. The imaging system was set up to demonstrate the imaging performance of each of the microlens array, evaluating the precision of each microlens array and imaging uniformity of the microlens array. The results show that the fabricated microlens arrays have good imaging performance and have a promising prospect in the use of 3D imaging and optical uniformity.

In order to realize the output of single longitudinal mode dual-wavelength laser signal at lower pump power and obtain the high frequency microwave signal with narrow line width, the narrow linewidth high frequency microwave signal generator based on multiple filter compound structure was proposed and demonstrated. The dual-wavelength Stokes optical signal with four times Brillouin frequency shift interval was realized through the eight shaped Brillouin cavity structure and the wavelength selective filter composed by reflective fiber grating. The 200 m length single mode fiber was used as gain medium, and it forms a cascaded fiber ring structure with 50 m long single-mode fiber. A three-port coupler and 2 m long unpumped polarization-maintaining erbium-doped fiber was used to form a Sagnac ring structure. The cascaded fiber ring configuration and Sagnac ring configuration were designed to select mode for single longitudinal mode Stokes optical signal. The experiment proves that the high-frequency microwave signal of 42.85 GHz can be generated by the beat frequency detection of the output single-longitudinal-mode dual-wavelength Stokes optical signal, and the line width is 38 kHz; Changing the output wavelength of the tunable pump laser, frequency tuning in the range of 42.25-43.51 GHz can be achieved; Through the stability test, the frequency change of the 42.85 GHz high-frequency microwave signal is in the range of 0.83 MHz, and the peak power change is in the range of ±0.8 dB. It has good stability and meets the actual application requirements.

Based on the design of the 220 GHz sub-harmonic mixer, the vertical conversion structure of the IF transmission waveguide was proposed, and the four-channel mixer integration module was realized, the transverse size of the single channel of the mixer was shorten effectively. It provided a feasible scheme for the multi-channel linear array integration of the terahertz receiver system. In order to further optimize the accuracy of the system model, the three-dimensional semiconductor device modeling calculation was carried out for the Schottky-barrier diode based on TCAD, and the high-frequency electromagnetic wave simulation of the mixer was carried out according to the extracted key characteristic parameters. Through the test of the design scheme, the test results show that when the local frequency is 110 GHz, and the power is 7 dBm, the conversion loss of the mixer is 8.6-13 dB as the RF input is 200-240 GHz, and the conversion loss is 8.6-11.3 dB at 204-238 GHz. When the local frequency is 108 GHz, the driving power only needs 3 dBm. In addition, the 220 GHz receiver system based on the mixer module has a temperature sensitivity of 1.3 K as the integration time is 700 μs.

In order to reduce the impact of the detector out-of-focus problem on the laser rangefinder system, an active rangefinder detector adjustment system based on split-image mirror was designed. Firstly, the imaging principle of the split-image mirror was analyzed, and its the detection accuracy was theoretically analyzed, and a mathematical model of split-image defocus was established; Secondly, in view of the error factors that affected the detection, the error curve under different values was given, and the change trend of the detection accuracy was obtained; Finally, the theoretical prototype was designed and verified by experiments. The experimental results show that when the defocus is 0-6 mm, the detection accuracy can reach 0.07 mm, close to the ideal value, and the smaller the defocus, the higher the defocus detection accuracy, which is in line with the installation and measurement habits. The defocus detection system provides a new method for the adjustment of non imaging photoelectric detector, which can further improve the measurement accuracy of laser rangefinder system.

The shaft guide plays an essential role in the production of the coal mine. Therefore, it is necessary to carry out a safety inspection on the shaft guide to ensure that its deformation range is within the normal range to avoid affecting the coal mine's safety production. To solve the problems of low precision and cumbersome process in the current detection methods of coal mines shaft guide in our country, a shaft guide detection algorithm based on two-dimensional laser scanning technology was studied, which could calculate the gap, misalignment between two shaft guides, and the wear of a single shaft guide, and write this algorithm into the upper computer software. In the laboratory environment, the algorithm had good repeatability. Simultaneously, in the actual coal mine experiment, the algorithm’s recognition accuracy rate reaches 62.5%; after the algorithm is improved, the recognition rate increases to 87.5%, proving that the algorithm has reasonable practicability and reliability. Compared with the traditional method, the method is simple to operate and has high measurement accuracy. This method's application in the detection of coal mine cage will effectively improve the production safety of coal mine and has great popularization and application value.

To solve the problem that the low defect detection accuracy on fiber laminated surface in Automated Fiber Placement(AFP) process based on visible images influenced by poor light source, low texture contrast of prepreg and other factors, a method for defect detection of laminated surface infrared images in AFP process based on improved CenterNet was proposed to improve defect detection performance on laminated surface in AFP process. First of all, due to the limit on hardware configuration of IPC, and large amounts of parameters on CenterNet model, lightweight MobileNetV3 network based on ASFF was utilized as the backbone network of AFP-CenterNet model to construct an anchor-free and lightweight detection model and reduce the number of network parameters and the occupancy of storage resources. Then, as for solving the bandwidth parameters of Gaussian kernel function, a method of adaptive adjustment of bandwidth parameters according to the aspect ratio of ground-truth bounding box was proposed to reduce the number of negative samples and loss error. Experimental results reveal that the improved AFP-CenterNet owns 90.2% AP in defect detection accuracy on the AFP infrared data set, 12.9 MB in model memory capacity, and 52 ms in detection time of a single sheet. Compared with the original backbone of CenterNet, detection accuracy of AFP-CenterNet is slightly worse than that of DLA-34, almost same with that of ResNet-101 and 7.7% higher than that of ResNet-18. Moreover, compared with DLA-34, ResNet-101 and ResNet-18, the model capacity of AFP-CenterNet decreased by 83.2%, 93.6% and 78.6% respectively. As for comparison with typical anchor-based network such as SSD and YOLOv3, AFP-CenterNet owns higher detection accuracy with specific 9.6% and 8.3%, and lower model capacity respectively reduced by 85.1% and 94.5%. Time spent on defect detection of AFP-CenterNet is nearly half that of CenterNet, SSD and YOLOv3 without using GPU to accelerate.

The separate optical probes have been widely used in scramjet laser absorption spectroscopy measured system. High resolution flow field cannot be obtained because of the limitation of the probe size. A high-resolution optical measuring ring based on free-curved lens and cylindrical lens was designed for combustion flow field. The beam distribution was determined by numerical simulation. The measuring ring adopted double-layer structure, the transmitters were located at the most edge of each edge. The fan-beam was formed by collimating lens and free-form lens. And then after passing the measuring flow field, the laser beam was defected by the wedge lens and focused by the focusing lens. The measuring ring receiving unit minimum interval was 5 mm, achieved dense array of 88 beams in 5 cm×7 cm space. The beam distribution of the optical system was discussed. The structural design scheme suitable for engine measurement was given. The measurement result shows that the light efficiency is greater than 50%, the total transmission efficiency is greater than 55%. The high-resolution optical measuring ring can be directly connected with the engine body to avoid the interference of environmental factors. The two-dimensional distribution of temperature and concentration in the isolation and outlet of the combustion can be measured.

Aiming at the problem of insufficient disturbance suppression capability in the current photoelectric stabilized platform servo system, a disturbance separation active disturbance rejection control (DSADRC) algorithm was proposed. The active disturbance rejection control method based on disturbance separation took full use of part of the known photoelectric stabilized platform model information and controller information in classical control that could be obtained in engineering practice, and added it to the design of active disturbance rejection control. This algorithm increased the disturbance observation accuracy and disturbance suppression of the system ability by reducing the total disturbance in the photoelectric stabilized platform system. At the same time, through proposed algorithm design, the reuse of classic controllers was realized and the design workload was reduced. The simulation results show that under the same controller and the same disturbance conditions, the step response settling time of the DSADRC is reduced by 58.8%, the rise time is reduced by 26.5%, and there is no overshoot; under the 1V2Hz equivalent disturbance, the system steady accuracy is increased by 51.5%, and the system performance is improved obviously. In the physical verification experiment, for equivalent disturbances of different frequencies, compared with PID control, the steady accuracy of the DSADRC is improved by more than 50%, which effectively improves the steady accuracy of the photoelectric stabilized platform.

Variable intelligent trusses have great advantages in optical equipment and on-orbit assembly. The relevant applications and the development of variable intelligent trusses on large-aperture optical equipment were summarized. Firstly, for serial configuration with less degrees of freedom, the design and application of variable intelligent trusses applied on telescopes were discussed; then, the application of the variable intelligent trusses in the on-orbit assembly, service and adjustment of the telescope were introduced. In view of its high degree of freedom and high positioning accuracy, the feedback methods that could be used in the adjustment of the trusses were summarized, the structure and control algorithm of the trusses were discussed as well. Finally, the technologies currently applied to smart trusses were summarized and the development trend was summarized.

In order to ensure the imaging quality of large-aperture survey telescope under dramatic changes of external environment and realize fast alignment, the wavefront sensing system was required to cover a large dynamic range while maintaining the aberration detection accuracy. First, a set of large dynamic range alignment technology was established based on the light intensity distribution of the single-sided defocus image and curvature sensing. Both analytical formulas and machine learning methods were used to figure out the defocusing and other low-order aberrations. Then, theoretical analysis was made on the resolution accuracy of different types of aberrations. Finally, experimental verification was carried out. The results show that the defocusing detection error (take wavefront RMS variation as the criterion) is less than 5%, while the misalignment detection error is less than 15%, which meets the alignment and adjustment requirements.

Laser guidance is one of the most commonly used guidance methods nowadays, and the performance of laser semi-active optical system directly affects its guidance accuracy. The method for the aberration optimization design of laser semi-active optical system was discussed, the initial structure of the optical system was realized by giving different spherical aberration and defocus values, and the spot uniformity design was realized by controlling the asymmetric aberration. A refractive laser semi-active lens was designed and fabricated, in 1064 nm working wavelength, with ± 9.2° field of view, 5 mm spot size, uniform energy distribution. In order to solve the problem that laser semi-active lens can not be detected separately, the principle of a low-cost visual testing of lens was proposed, which based on the chromatic aberration characteristics, and the visual testing system for laser semi-active optical lens was built. Test results of the lens show that the spot size satisfies the design requirements, and the low-cost visual testing system significantly improves the efficiency of lens testing and is easy to engineering and mass-produce.

To solve the problem of the large amount of noise in the event stream output by the event camera, an event stream denoising algorithm based on the probability undirected graph model was introduced. Due to the imaging principle of the camera, the change of the target had certain regularity and correlation in time and space. By mapping the event to the polar coordinate space-time neighborhood, the local correlation of the event was established to build a complete probability graph model. In addition, the improved conditional iterative mode algorithm was used to optimize the iterative solution of model. The experimental results of simulated data generated by the event camera simulator and the real data recorded by DAVIS346 show that the proposed algorithm can effectively remove noise events. Finally, the comparison with the filtering algorithm proves that the algorithm is superior to the filtering algorithm.

An prototype of imaging spectrometer for the remote sensing of solar induced chlorphyll fluorescence of vegetation was researched to satisfy the scientific requirements of high SNR and high spectral resolution for the weak fluorescence observation. Based on the principle of relationship between the fluorescence and solar Fraunhofer lines, the optimization of the optical system with high performance was completed. It covered the visible-near infrared working waveband of 670-780 nm, which also covered the special wavelengths of the solar induced fluorescence. The numerical aperture was 0.25 which ensured the enough SNR of the system. The imaging spectrometer also owned spectral resolution better than 0.3 nm and excellent imaging quality. The performance tests of the prototype were studied. The research will supply an excellent engineering application in the further solar induced chlorophyll fluorescence high spectral imaging observation.

High-resolution space optical camera defocused in orbit due to the change of space operational environment, which would affect the image quality. Therefore, refocusing were required by the camera in orbit. In order to meet the requirements of high resolution and lightweight space optical camera, a main support structure with both support and refocusing function was designed. The position of the secondary mirror assembly was adjusted along the optical axis by the precise temperature control of the supporting structure, then the camera can be refocused by the thermal control of the supporting structure. Firstly, the refocusing accuracy was analyzed according to the optical system parameters to determine the design requirements of the support structure; Then, the global optimization of the supporting structure was performed based on the variable density of continuous topology optimization (SIMP); Finally, the thermal optical test was performed to verify the thermal refocusing function and measure the thermal refocusing parameters of the apace camera. The experimental results show that the thermal refocusing parameter of the supporting structure is 0.071 mm/℃, the refocusing accuracy and range are 0.014 mm and ±0.385 mm respectively. The proposed method had been used in the design of "Jilin-1"gf03 satellite which had been tested in orbit, and the focusing accuracy and focusing range meet the design expectations.

The design, assembly and experiment of the optomechanical system of a compact spaceborne video camera developed for the 20 kg micro-nano optical remote sensing satellite were introduced, and the integrated optimization method was also proposed. The camera was a Cassegrain optical system including two mirrors and one corrector assembly. In order to obtain the best thermal stability, the mirrors were made of silicon carbide. Firstly, based on the task and overall design of 20 kg micro-nano video satellite, the requirements of video camera were proposed; Then, the optical and optomechanical structure system of the video camera were introduced respectively; In order to further improve the lightweight rate, while meeting the requirements of optical performance, the optomechanical integration optimization method was used to the lightweight design. After optimization, the mass of the optomechanical system was only 3.03 kg, the weight of the whole camera was only 4.76 kg, and the 1st mode was larger than 120 Hz; Finally, the assembly and ground mechanical experiments of the camera were summarized. The results showed that the camera had very dynamic properties and stability.

To solve the problem of high-precision testing of large-diameter plane mirrors, a mathematical model of subaperture stitching testing based on global optimization was established, and a stitching factor was proposed for overlapping area values. Based on the above method, combined with engineering examples, the stitching testing of a plane mirror was completed with a diameter of 120 mm, and four subapertures to be tested were planned. In order to compare the stitching performance of the algorithm described in this paper with the traditional least-squared fitting stitching algorithm, two algorithms were used to complete the surface reconstruction of the plane mirror to be measured. The experimental results show that the stitching results obtained by the two algorithms are smooth, continuous, no "stitch marks". At the same time, the results of the two algorithms are also compared with the full-aperture testing results. In this paper, obvious "stitch marks" can be seen in the residual map of the traditional splicing algorithm, and the stitching results obtained by the algorithm method in this paper are smooth and continuous, while the PV and RMS values of the residual graph are 0.012λ and 0.002λ, respectively, which are less than the PV and RMS values of the traditional algorithm residuals chart, which verifies the reliability and accuracy of the algorithm.

With the development of advanced optical system design and manufacturing, large aperture optical system has been widely used. However, the lack of high precision surface shape detection means limits the manufacture and application of large aperture plane mirrors. In order to detect the surface shape of large aperture planar mirror with high precision, a Shack-Hartmann scanning and stiching detection method was proposed. The scanning and stiching principle and wavefront reconstruction algorithm were studied, and the mathematical model of microlens array imaging was established to verify the feasibility of Shack-Hartmann scanning and stiching detection principle. A scanning and stiching test experiment was carried out for a 150 mm aperture plane mirror, the full aperture surface shape 0.019λRMS(λ=635 nm) was obtained. Compared with the results of interference detection, the detection accuracy was 0.008λRMS. The results show that the method can realize the high precision detection of large aperture planar mirror.