A random phase screen model of ocean turbulence is proposed. Based on the fluctuating spectrum of ocean refractive index, the influence of the refractive index fluctuation of ocean medium on the transmitted beam caused by oceanic turbulence is equivalent to the influence of random phase screens on the beam with the power spectrum inversion method. The validity of the proposed model is testified by the phase structure function and the transmission characteristic of orbital angular momentum beam in ocean turbulence. Results show that the phase structure function values calculated by the random phase screen model are coincident with the theoretical results when the offset of beam refractive index is small. At the same time, the transmission characteristics of the orbital angular momentum beam obtained by the random phase screen model are also consistent with the theoretical analysis results.

The retrieval precision of atmospheric turbulence intensity profile is affected directly by the signal noise from differential light column image motion lidar. A valid denoising method can improve the detection performance of lidar. To reduce the dependence of the combined denoising method with wavelet-empirical mode decomposition (EMD) on wavelet, an adaptive modulation strategy with the signal trend term is proposed for wavelet denoising signal, and then the modulated signal is denoised by EMD, which is called wavelet-trend-EMD method. The trend term is still extracted by EMD. To ensure the validity of modulation, a decision criteria of modulation suitable for coherent length (r0) profile is presented, and the detrended fluctuation analysis is carried out to identify the EMD denoising threshold adaptively. To illustrate the validity of the proposed method, four other methods of wavelet, EMD, ensemble empirical mode decomposition (EEMD) and wavelet-EMD are used for comparison. The numerical simulation and experimental results indicate that all the five methods can improve the signal-to-noise ratio of r0 profile and the retrieval precision. The two joint methods are better than the single method and the wavelet method is superior to EMD and EEMD methods. More importantly, wavelet-trend-EMD further improves the denoising ability of wavelet-EMD, which provides a new improvement consideration for the joint method of Wavelet-EMD.

Based on the CCD side scattering lidar (Clidar) and Mie scattering theory, a model of the relationship between the side scattering light intensity distribution and the PM2.5 concentration is established, and the measurement error is analyzed. In the study, the system setup of Clidar PM2.5 concentration measurement, which uses a laser at 532 nm as a light source and a CCD as a receiver, is constructed. The obtained side scattering echo signal is recorded to compare with the PM2.5 concentration measured by the BAM-1020 particle monitor, and hierarchical model expressions under high and low gains are built. When we compare the predicted values of hierarchical model PM2.5 concentration with the measured values, the average error, residual variance and comprehensive deviation rate statistics of each model are obtained. The results show that the inversion accuracy of the high gain model is better than that of the low gain model when the PM2.5 concentration is in the range of 0-70 μg·m -3. In the high gain model, the inversion accuracy of S(20) model is the highest.

An electrostatic trapping scheme using three charged spherical electrodes with opened optical access for cold polar molecules in the weak-seeking states is proposed, and the schematic drawing is given. The analytical expressions for the space electrical field distribution are derived with the image charge method. The numerical solution of electrical field distribution is calculated by the finite element software. Dynamic process of the loading and trapping cold polar molecules in an electrostatic trap is simulated with the classical Monte Carlo method. The influences of velocity of the incident molecular beam and loading time on the loading efficiency are studied, and the temperature of the trapped cold molecules is given. Some potential applications of the scheme in the electrostatic surface trapping especially the electrostatic lattices for cold polar molecules are discussed. Results show that the loading efficiency can reach 47.4%, and the temperature of trapped cold molecules is 25.4 mK.

With the HTV-2-like hypersonic gliding aircrafts as the research objects, their infrared radiation characteristics are simulated. With the overall consideration of the direction and spectral characteristics of the target, the background and the transmission process, the infrared detection capabilities of the ground-based platform, the floating machine and the space-based satellite to the HTV-2-like hypersonic aircrafts are systematically analyzed, and the maximum detection distances for different detection platforms and detection wave bands are obtained. The study results show that the observation direction has a great influence on the infrared radiation intensity of targets, and the maximum detection distance increases with the decrease of the detector sensitivity. The detection distance in the medium-wave bands (3.7-4.8, 3.0-5.0 μm) is longer than that in long-wave bands (7.7-9.5 μm, 8.0-12.0 μm).

The traditional optical lens detection is mainly based on echo intensity (cat-eye effect). However, there are obvious diffraction phenomena in reflection echoes for micro-lens. Thus, it is important to investigate the diaphragm diffraction effect of micro-lens. Based on the theory of angular spectral diffraction, we study the spatial modulation characteristics of micro-lens for irradiated laser, and deduce the equation for optical field distribution of echo signal. The influences of the detection distance, incident angle, and diaphragm diameter on the diffraction optical field distribution are simulated, as well as the maximum detectable incident angle under the condition of different diaphragm diameters. Moreover, the diffraction phenomena are detected using the constructed active laser detection system. The results indicate that the diffraction patterns observed from the experiment match well with the simulated patterns. The theory, developed in this article, is proved to be qualified to predict diffraction profiles of micro-lens' echo signal to an extraordinary degree.The theory can accurately predict diffraction phenomena in echoes for micro-lens.

It is necessary to measure and compensate the phase fluctuation introduced by fiber optic links in high-precision fiber frequency transfer. Frequency crosstalk is one of the influence factors, which is easy to be introduced and has a certain impact on the frequency transfer performance. In order to evaluate the impact of frequency crosstalk on fiber frequency transfer performance, the influence model of frequency crosstalk on frequency transfer stability is established. The stability loss caused by frequency crosstalk due to temperature change of fiber link is analyzed by simulation and experiment. The results show that the stability loss of the frequency transfer is related to signal crosstalk factor and the frequency of frequency standard. The larger signal crosstalk factor, the larger the overall up-shift amplitude of Allen deviation curve, and the crosstalk factor is approximately linear with the stability loss peak value. When the frequency of frequency standard changes, the long-term stability of the frequency is not only affected by the slow delay drift caused by the temperature variations, but also related to frequency points and delay fluctuation. If the frequency rises, the peak of the frequency stability loss will move to the short term.

The effect of dispersion compensating photonic crystal fiber (DC-PCF) on dual-channel photonic time-stretched analog-to-digital converter (PTS-ADC) is studied. The principle that DC-PCF can suppress third-order harmonic in PTS-ADC is theoretically verified. The dual-channel PTS-ADC based on DC-PCF is designed and simulated with the Optisystem software. The effective number of bits (ENOB) of PTS-ADC is simulated and compared between DC-PCF and dispersion compensation fiber (DCF) as dispersive mediums under the condition of five different input radio frequencies (RF) signals (16.25, 20.25, 24.25, 28.25, 32.25 GHz). The simulation results show that DC-PCF can effectively suppress the generation of third-order harmonics and improve the quantization resolution of dual channel PTS-ADC system.

We use highly coherent and very low-noise laser as the light source to design fiber Michelson interferometers with an optical path difference of 60 m used as the front-end transducers. Binary rectangular pulse is used to modulate the phases of the light source, and the phase signals are restored with an orthogonal demodulation algorithm. A highly efficient interrogator controller based on field programmable gate array(FPGA) is used to calculate phase signals in real time. Real time and accuracy of invasion alarm signals are improved, based on the time-domain level crossing (LC) algorithm. Finally, a four-zone security system is set up, which has advantages of simple interrogator controller, low cost and real-time handling capability. The experimental results show that the system can alert all kinds of intrusions in real time and suppress the false alarm caused by the environmental noises, such as winds and rains, under the chosen proper threshold. The system has worked for 1 month without alarm failure and the false alarm rate is less than 1%.

In mode division multiplexing systems based on few-mode fibers, it is inevitable to carry out the fusion splice between the few-mode fibers. Precise measurement of mode coupling at each splice point can provide reliable bases for evaluating welding-quality and positioning faults of the system. We analyze the mode coupling characteristics at a splice point between few mode fibers based on Rayleigh backscattering theory, then build a measurement system of mode coupling at a splice point based on the structure of a photonic lantern and an optical circulator, and successfully measure the mode coupling at a splice ponit between two pieces of three modes fibers(the lengths are 0.9 km and 9.8 km, respectively). The experimental results show that the mode couplings at the splice point are respectively -14.9 dB and -13.9 dB with the offset distances of 1.5 μm and 2 μm.

In the study of coherent optical imaging, the coherent transfer function is the fundamental basis for investigating the imaging quality of coherent light. At present, the coherent transfer function is defined under a particular approximation condition. The imaging formula of single lens system has been created based on coherent transfer function, which can only approximately calculate the amplitude distribution of the image field. Nevertheless, amplitude and phase of the image field are of equal importance. In this paper, the single lens imaging system has been studied. An accurate formula for the complex amplitude distribution of the ideal image and the image field near the image plane is established. The calculation method for the amplitude and phase distribution of the defocussing image field is given. Finally, the relative experimental verifications are performed.

Based on the nodal aberration theory, an aberration analysis method for pupil off-axis freeform surface optical system with Zernike polynomial is proposed. The analytical expressions of third-order astigmatism and coma of the pupil off-axis freeform surface system are deduced. The impact of freeform surface on the distribution of aberration nodes of pupil off-axis system is analyzed. According to the aberration distribution properties, with pointed optimization of the Zernike terms, a pupil off-axis freeform surface space telescope optical system is designed. The effective focal length is 25 m, the pupil diameter is 2 m and the field of view is 1.2°×1.2°. By introducing the freeform surfaces, the third-order astigmatism and coma nodes are both moved back to the field of view. Therefore, the asymmetric aberrations induced by pupil off-axis are balanced, the point spread functions (PSF) ellipticity are controlled effectively, and the measurement errors of star observation are decreased. The final imaging quality is closed to the diffraction limit, which meets the requirements.

The structure and characteristics of asymmetrical compound parabolic concentrators with plane absorber (PACPC) is researched. The surface structure model of compound parabolic concentrators (CPC) with plane absorber is established, and the structural equation of PACPC is further established on the basis of CPC, and the analytic solution is obtained. The optics characteristic of the established PACPC surface structure model is simulated by the optical design software TracePro, and the optical characteristics are investigated. It is found that, with the change of incident angle of light, the radiant energy at the absorber surface is distributed unevenly. According to the solar direct and dispersion radiation theory, the acquisition of direct and scattered radiation in the course of the year is calculated by numerical calculation. The results show that the average annual lighting quantity obtained by the system of PACPC is obviously more than that of the plane absorber system with same area, and the increment is 15.28%.

The robustness of an optical encryption system directly determines the anti-jamming capability of the ciphertext in the process of storage and transmission, and is always evaluated in the form of the ability against noise and occlusion attacks. The parallel encryption system for multi-channel images based on a joint transform correlator can encrypt the multi-channel images simultaneously into a single ciphertext, and also recover the arbitrary original images with corresponding keys. The robustness of this system is simulated and analyzed. The results show that the decryption effect of each channel is reduced obviously when the ciphertext area is cut by 50%. The correlation coefficient (CC) values between some original and decryption images get the level of about 0.2. The reason is analyzed and the arrangement of the joint power spectrum is optimized. The simulation results show that not only the robustness of the encryption system is significantly improved, especially when the ciphertext area is cut by 60%, the CC values of all channels are greater than 0.49, but also the advantage on resisting noise attacks can be maintained. Finally, an optical experiment system is built to verify the effectiveness of the proposed method for improving the robustness of the encryption system.

A super-resolution reconstruction algorithm based on convolutional neural network (CNN) is proposed to solve the problem that the traditional depth map super-resolution reconstruction algorithm needs to extract the feature manually and the computational complexity is higher, and it is not easy to get the proper representation feature. CNN does not need to extract the specific features from the image in advance for the specific task, but the simulated human vision system can extract the feature independently by hierarchical abstraction process on the original depth map. This algorithm can achieve mapping learning directly from the low resolution depth map to high resolution depth map. The mapping is implemented by seven convolution layers and one deconvolution layer. The convolution operation is used to learn the rich image features,and the deconvolution realizes that the upsampling is used to reconstruct the high resolution depth map. The experimental results of the Middlebury RGBD dataset show that the average peak signal-to-noise ratio (PSNR) and root-mean-square error (RMSE) obtained from the model can increase by 2.7235 dB and decrease by 0.098 compared with the traditional bicubic interpolation algorithm, respectively. Compared with the classical image super-resolution reconstruction using deep convolutional neural networks, the performance is also improved with 1.5244 dB of PSNR increment and 0.043 of RMSE decrement.

Aiming at the transmission and storage requirements of three-dimensional model in the large data environment, a three-dimensional point cloud lossy compression algorithm based on the octree is presented. The stop condition of the octree segmentation is improved, so the segmentation can be stopped at an appropriate depth, and the proper size of voxel is ensured. At the same time,the K neighborhood is established based on the segmentation and the outliers of original point cloud are removed by simple and effective statistical method. In the data structure, each node is assigned to a bit mask. The data query and manipulation are traversed by manipulating the bit mask. Then the subsequent point position coding are optimized. The proposed algorithm effectively removes the outliers and miscellaneous points on the surface, and improves the efficiency of point cloud compression in range encoding. The experimental results show that this algorithm can preserve the key information of three-dimensional point cloud data more completely, obtain a good compression rate and shorten compression time.

In view of existing problems of image super-resolution method based on sample-learning, which is difficult to operate rapidly and generate high quality image at the same time, a rapid image super-resolution method based on deconvolution is proposed. A new type of network model is designed and low resolution images are taken as input images directly, and then convolution layer is used to extract and represent features. Deconvolution layer is used to enlarge image feature maps, and the following pooling layer is used to concentrate the feature maps and extract features which are more sensitive to the results. Moreover, sub-pixel convolution layer is applied to features mapping and images fusion simultaneously and the super-resolution image could be obtained. The proposed method is tested on images of test datasets, and compared with other methods. The test results of the proposed method have higher peak signal to noise ratio (PSNR) and can process more than 24 images in size of 320 pixel×240 pixel per second, which shows that the proposed rapid image super-resolution method based on deconvolution can not only generate images with higher quality, but also satisfy the requirement of real-time video processing.

In order to achieve the high precision mapping between digital micromirror device (DMD) micromirrors and CCD pixels in high dynamic range imaging system based on DMD, three different pixel-level mapping algorithms are put forward. The high dynamic range imaging system based on DMD is introduced, and the mapping principle and necessity are analyzed. The proposed algorithms are used to map, including the direct linear transformation algorithm, nonlinear distortion polynomial fitting algorithm and back propagation neural network algorithm. And the theoretical methods are combined with the imaging system. The corresponding parameters are obtained and the error data are calculated by image testing. The error of fitting polynomial distortion mapping algorithm is minimum, and the root mean square error is 1.02 pixel, which can reach the pixel-level error. This algorithm fits the partial distortion error mapping at pixel-level and meets the needs of the imaging system.

In order to evaluate the influence of the tracking and imaging platform on the modulation transfer function of the image accurately and provide a theoretical basis for the determination of the platform’s technical specifications, an analytical and numerical model is established, which is used to calculate the movement optical transfer function by the image motion function. The truncation error of the finite term is given, and the truncation error is less than 10% when the 7 order approximation is taken. The proposed model includes analytic expressions of low frequency sinusoidal motion and high frequency sinusoidal motion. The tracking angular velocity errors of a tracking imaging platform are analyzed by frequency spectrum. The fundamental components of frequency spectrum are composed by multiple low frequency sinusoidal vibration components. Image motion caused by the tracking angular velocity error can be approximated to uniform linear motion at the exposure time scale. The modulation transfer function is calculated by the statistical moment method, and is almost equal to that calculated by the root mean square of the tracking angular velocity error. The calculated deviation is less than 0.01 at 200 mrad-1 frequency. Therefore, the root mean square error of the tracking angular velocity is a reasonable parameter describing the tracking stability of the tracking imaging platform.

The imaging process of the underwater image is similar to that of the haze image. However, the traditional dehazing methods used to underwater image processing can not obtain fruitful results. Aimed at problems of serious color attenuation and blue (green) tone during capturing underwater images, we propose an underwater image enhancement method based on improved dark channel prior and color correction. By considering the propagation characteristics of light in water, we improve the dark channel prior dehazing algorithm for underwater image, and by considering the inverse channel of red channel, we obtain the underwater dark channel and image background light. It is proposed to use the improved underwater dark channel prior to remove the backward scattered light firstly, and then use the white balance algorithm to correct color. The experimental results reveal that the method proposed in this paper, compared with traditional methods, can acquire images with higher resolution and higher contrast during processing serious backscatter of underwater images.

To verify whether the ground ultraviolet detection system can be used to detect the high altitude and high speed reentry targets above the ozone layer, mechanism research of UV radiation caused by high altitude high speed reentry target and feasibility of ground UV detection are studied, and in the return test of the shrinked ratio cabin of a multipurpose spacecraft, an equipment of UV optical imaging is used to carry out the verification test, which obtained above 40 km of high altitude and high speed reentry target surface UV detection image, broke the blank record of ground optical detection which had never get UV image of high-altitude above ozone layer of high speed target. The results show that under the condition of thin atmosphere, the high altitude high speed target can produce strong ultraviolet radiation. And it is feasible to use the ground ultraviolet detector to detect the high altitude high speed reentry target. Moreover, the imaging target can be effectively separated and extracted from the background. It has the important reference value for the further development on practical application and further research on the theory of ultraviolet detection.

Optical frequency scanning interferometry (OFSI) is a high-precision absolute distance measurement technique with broad prospect. However,in practice, it is sensitive to the drift of optical path difference. Due to the influence of the amplification factor, the slight drift of optical path difference in measurement process is enlarged, which introduces obvious measurement error. The influence of Doppler effect introduced by the optical path drift of measuring arm on the measurement result has been detailedly analyzed in this paper. According to that the sign of the frequency drift error is consistent with direction of frequency scanning, the drift error is compensated by optical frequency scanning method with triangular wave. The experimental results show that, for the target at 3.6 m, the standard deviation decreases from 21.51 μm to 2.85 μm with 10 times of measurements. Furthermore, the contrast experiment result with interferometer shows that the standard deviation of residual error of the proposed method decreases from 18.6 μm (before compensation) to 5.6 μm, and the measurement accuracy is enhanced.

For the three-dimensional measurement method of projection structure light, the three-dimensional measurement of the objects with specular surface can’t get a clear coded image because of the specular reflection. It affects the calculation of the three-dimensional coordinate signal, and the whole three-dimensional morphology cannot be fitted. A new type of digital micromirror device (DMD) applying DMD as core is proposed. The problem of over saturation in acquiring coded image for traditional charge coupled device (CCD) is solved with high dynamic range imaging of DMD. A new system is built based on traditional monocular three-dimensional measurement system. The digital model of three-dimensional measurement and high dynamic range imaging of DMD camera are established. Combined with the matching method of Gray Code, three-dimensional coordinates of the object are obtained. The experimental results show that the system has a positive effect on the imaging of object with specular surface, and relatively complete bitmap is obtained. The system is able to measure three-dimensional morphology with high dynamic range.

A novel plasma curved waveguide filter based on microcavity coupling structure is presented, which consists of two rectangular waveguide and a rectangular resonant cavity. When light passes through the structure, surface plasmon polaritons (SPPs) can be excited. The propagation properties of the SPPs with this structure are investigated by the finite difference time domain (FDTD) method. The results show that, compared with the traditional straight waveguide structure, the single microcavity curved waveguide structure can generate stronger resonant interaction and higher coupling effect for the bilateral coupling effect induced by the structure. The numerical simulation results show that the resonant wavelength of the filter can be adjusted linearly by changing the cavity length of the resonator. In addition, based on the above design idea, a dual microcavity structure is also proposed. The structure consists of a bent waveguide and two resonant cavities at left and right, which can be used to produce dynamically tunable plasma induced transparency by the superposition of two microcavity transmission waves.

From the perspective of mechanical structures, the problem of the thermal stress within the mechanically mounted KDP crystals after the energy absorption from a high-energy laser beam is studied. The temperature rise within KDP crystals induced by the laser energy absorption is theoretically analyzed, and the generation mechanism of the thermal stress within the mechanically mounted KDP crystal is investigated. By the finite element method, the temperature distribution and the thermal stress distribution in KDP crystals are simulated, and the effects of the structural parameters of the mounting set, the preforce, the friction coefficient, and the elastic modulus on the thermal stress are investigated. The results show that the mechanical mounting action provided by the mounting set is a major factor which causes the thermal stress within KDP crystals, and the thermal stress value is associated with the structural parameters of the mounting set.

Based on the removal of insulating layers by CO2 lasers, the shielding layer material of cables is removed by YAG lasers, and the theoretical model of YAG laser peeling of metallic materials is established. The process of laser peeling of materials is simulated and analyzed by ANSYS software, and the experimental parameters are determined. With the experiment of YAG laser peeling of the shielding layers of RG113 cables, the influence rules of laser power, scanning speed, pulse width and defocusing distance on the incision width, the incision depth and the surface quality are obtained. With the range analysis of the orthogonal tests, the influences of different process parameters on the incision width and depth are studied. The results show that the laser power has the greatest impact on the incision width, the incision depth of insulating layers and the incision depth of the metallic shielding layer, while the laser scanning speed has the greatest impact on the incision width of the metallic shielding layer.

The causation of the generation of strong amplified spontaneous emission (ASE) in the Nd∶YAG slab is analyzed, and the academic and experimental researches on suppressing ASE in high gain laser slab are developed. The pumping duty ratio is 8%, the peak pumping power is 21.38 kW and the injecting intensity of 1064 nm continuous detecting laser is 4 W/cm2. The output power amplifying ratios of the detecting laser are 1.82 and 1.92 respectively with common evanescent coating and multilayer coating. The total stored energy in the slab increases by 4.6%. The experimental results show that ASE effect can be suppressed and the total stored energy in the laser slab can increase to a certain extent when special multilayer coating is coated on the top and bottom surfaces of the slab.

Single infrared or visible light band imaging technology is insufficient to meet the requirements of color imaging in night and foggy days. To overcome this problem, a color video construction method for fixed area in night and foggy days is proposed, which combines the characteristics of infrared imaging and visible light imaging. The proposed method constructs the visible background image by using visible light sensor during daytime with high visibility, and extracts the infrared moving target by using infrared sensor in night and foggy days with low visibility. Then, according to the registration parameters of the visible background image and the original infrared image, the two images fuse in the same proportion, and the construction of color video is thus completed. The experimental results show that the proposed algorithm can accurately complete the construction of color video with infrared target. It fully reflects the color feature information of the moving target and its scene in night and foggy days, and enhances the recognition and perception of target and scene information by the human eye. For a video sequence with an image size of 720 pixel×576 pixel, the running speed of the algorithm can reach 40 frame/s and meet the the need for real-time construction of color video.

In order to analyze the three-dimensional information such as the position, shape and direction of the ceramic bowl, the reconstruction algorithm of a local point cloud based on image sequences is proposed. Firstly, a calibrated binocular camera is used to collect many images of surface defect from different angles. Then the image feature point detection and matching algorithm of FAST+SURF+FLANN is used to get high precision matching point pairs. Finally, the local three-dimensional reconstruction of two-dimensional surface defects is realized through the structure from motion algorithm combined with patch-based multi-view stereo algorithm. However, the direction and location of the defect cannot be perfectly described through the local three-dimensional reconstruction. So the proposed algorithm of manually added feature points is adopted to realize the global reconstruction of the surface of the ceramic bowl. The results show that the defect reconstruction results are clear, and the defect position, direction and shape information is complete.

For depth estimation from monocular infrared video, a method based on bi-recursive convolutional neural network (BrCNN) is proposed considering the uniqueness of a single frame and the continuity of the entire infrared video. BrCNN introduces the sequence information transfer mechanism of recurrent neural network (RNN) on the basis of the single frame feature extracted by the convolutional neural network (CNN). Thus, BrCNN possesses the feature extraction ability of CNN for a single image, which can automatically extract the local features of each frame in the infrared video, and the sequence information extraction ability of RNN, which can automatically extract the sequence information contained in each frame of the infrared video and recursively transfer this information. By introducing the bi-recursive sequence information transfer mechanism to estimate the depth of monocular infrared video, features extracted from each image containing the context information. The experimental results show that BrCNN can extract more expressive features and estimate the depth from the infrared video more precisely than the traditional CNN, which estimate the depth by extracting the feature of a single frame.

Most of the current stereo matching algorithms have high matching accuracy, but there are very few of them can realize real-time matching with video level frame rate. We present the multi-scale optimization algorithm based on Bayesian reasoning, which can be used to improve the matching accuracy, while maintaining the real-time performance. This algorithm obtains disparity maps with different scales by setting different window sizes. Based on this, the joint optimization based on Bayesian reasoning is proposed to optimize the disparity maps with scale information and complementarity. And then the high precision disparity maps are obtained. Test results of Middlebury stereo vision datasets show that the proposed algorithm has better accuracy and higher efficiency than several real-time algorithms.

The design of the effective feature vectors is the key to the saliency detection algorithm, which determines the upper bound of the model effect. A new saliency detection algorithm based on global model and local search is proposed by combining the deep convolution neural networks and the hand-crafted features. In the global model, the initial saliency map is generated from designing the extra convolution layers for VGG-16 network training, and thus the saliency value of each object candidate region can be predicted from a global perspective. In local optimization model, the super-pixel region with multi-degree segmentation is described by designing the contrast descriptors and region characteristic descriptors, and the saliency score of each region is predicted. Finally, a linear fitting method is used to fuse the result generated from two models, and the final saliency map is obtained. Contrast experiments for four data sets are demonstrated and the results show that the proposed algorithm has the highest precision.

The adaptive tracking algorithm for the aerial small target is proposed based on the multi-domain convolutional neural networks (MDNet)and the autoregression (AR) model,to solve the tracking drift problem that the pseudo targets and the small target converge in the sky background. Firstly, the positive samples of the first frame in the image sequence are collected by MDNet to train the bounding-box regression model. Secondly, the AR model with its order and parameters determined by the Akaike information criterion and least squares method is trained to estimate the target track and to predict the target position. Finally, the region of sampling candidate is constrained since MDNet collects samples centered on the predicted target location, and then the target position is adjusted by the bounding-box regression model. Eight groups of benchmark video sequences are tested with the proposed algorithm and another seven classical tracking algorithms, and obtained results are compared. The experimental results show that the success rate and the average overlap rate of the proposed adaptive tracking algorithm are higher than those of other algorithms, and the proposed algorithm has higher accuracy and robustness.

Aiming at the problem that the stereo matching algorithms have high mismatching rates in non-occluded regions of the images, especially the weak texture regions, a stereo matching algorithm based on adaptive color weights over cross window and tree dynamic programming is proposed. Firstly, we combine color, gradient information and census transform as similarity measure function to propose the cost calculation function. Then the adaptive cross window is constructed with distance and color information of the image, and the cost aggregation based on color weights is proposed. Instead of using winner-take-all strategy solely for global optimization of disparity, the dynamic programming algorithm based on tree structure is introduced to calculate disparity. Finally, the dense disparity maps are obtained by the process of disparity refinement. The experimental results demonstrate that on the Middlebury test platform, the average mismatching rate evaluated with proposed algorithm in non-occluded regions of four standard images is 2.45%. Meanwhile, the other ten images are compared and evaluated. The proposed algorithm effectively improves the accuracy of stereo matching in non-occluded regions.

To enhance intensity of Cerenkov fluorescence and promote clinical transformation of Cerenkov luminescence imaging (CLI) technology, we propose an enhanced Cerenkov luminescence imaging (ECLI) technology by utilizing radioluminescence microparticles (RLMPs) in previous study, and the technolgoy can enhance the intensity of Cerenkov fluorescence effectively. To extend the application of ECLI technology to the field of three-dimension imaging, we propose a novel single-view enhanced Cerenkov luminescence tomography (ECLT) reconstruction method. In this method, single-view data acquisition is used, and sparse Bayesian learning (SBL) reconstruction algorithm combined with the strategy of iterative-shrinking permissible region is adopted to solve the inverse problem. Non-homogeneous cylinder simulation and physical phantom experiments are designed and conducted to verify the accuracy and stability of the proposed method. The results indicate that the proposed method can improve the reconstruction accuracy and speed, and the method has good stability and can effectively mitigate the ill-posedness of the inverse problem.

By modeling with COMSOL software, we carry out the thermal analysis of optical parametric oscillator (OPO). We find that under high pump power, temperature near the central axis of crystal increases due to the absorption of parametric light. Because the refractive index of MgO∶PPLN crystal is very sensitive to temperature, a small temperature rise will cause the non-negligible wavefront distortion of parametric light, which leads to the decrease of beam quality. The temperature rise of crystal is caused by the absorption of parametric light, and such temperature rise is unavoidable. The deterioration of beam quality in high power OPO/OPA (optical parametric amplifier) is inevitable.

Retinal projection display (RPD) is one of the research hotspots in the field of head-mounted display (HMD), it can overcome the vergence-accommodation conflict (VAC) of the traditional HMD. In order to make the RPD small and portable, a catadioptric laser RPD system, which uses a micro-electromechanical system (MEMS) scanning mirror as the spatial light modulator (SLM), is designed and fabricated. Firstly, the vergence-accommodation conflict is introduced and the basic principle of RPD is analyzed. Then the feasibility of Maxwellian view principle that uses a MEMS laser scanning projection as the image source is experimentally verified by the traditional RPD optical structure. And the black speckle problem is analyzed and solved. The optical design of the RPD system is finished, and the performance of the system is analyzed and evaluated. Finally, the small portable prototype is fabricated, whose interpupillary distance is adjustable, with a viewing angle of 30°(H)×22°(V) and distortion-free. Display performance of the prototype is demonstrated through experiments.

Against the design scheme of 8 m ring solar telescope (8m-RST) of Chinese Giant Solar Telescope (CGST), the model of the edge sensor and optical sensor response in segments displacement is established, and the influence of the segments in-plane displacement on the surface shape of ring segmented primary mirror is analyzed. Considering the segments' displacement introduced by zenith angle change and temperature change for telescope truss, it's found that the displacement will destroy the effect of surface shape keeping of 8m-RST. To avoid the adverse effect of the in-plane displacement on 8m-RST, an improvement plan of the corresponding control model for the primary mirror is proposed. Which is correcting the set value of the edge sensor reading and the optical sensor reading, by adding two groups of gap sensor and one group of shear sensor among the edge of neighbor segments and using increased detection amount to estimate the in-plane displacement. When the detection accuracy of the gap and shear is less than 23 nm, the improved plan can meet the surface shape requirement of 8m-RST scheme for CGST.

A new inverted trapezoidal cavity receiver with absorber of a bundle of tubes used in trough solar concentrator system is designed. The influence of some geometric parameters in the cavity design on the optical performance of the receiver are theoretically analyzed and optimized by theoretical analysis, Tracepro software simulation and experimental methods. Furthermore, the influence of installation position deviations on optical performance of the cavity receiver are studied. Simulation and analysis results show that the wall surface on inverted trapezoid is reflex arc surface, whose optical efficiency can be improved nearly 4%. The optical efficiency increases gradually with the increase of opening angle, and firstly increases and then decreases slowly with the increase of arc radius. The combination of the opening angle of 65° and the radius of the reflection arc 90 mm has higher optical efficiency, more uniform energy flux distribution on the wall of collector tube, and better photothermal transfer efficiency. When vertical installation position is 5 mm below focal length, the optical efficiency is better than that at the focal length. By using the normalized temperature difference and the test method of the instantaneous heat collection efficiency, the optical efficiency of the trough concentrating collector is 73%, which is in good agreement with the ideal value of the simulation.

The surface albedo is a prerequisite for the research of the information such as land radiation budget, regional and global climate, and geomorphology. Therefore, its precise measurement is the premise and guarantee of relevant research. Due to the influence of atmospheric absorption, the surface albedo cannot be directly measured. Therefore, the method of measuring vertical downward and upward scattered light and combining SCIATRAN radiative transfer model is adopted to obtain the surface albedo. Based on this method, the observation experiment is carried out on the airborne platform. The vertical downward and upward radiance values are obtained. The iterative retrieval method is used to obtain the surface albedo of the area from Shijiazhuang to Baoding in the ultraviolet band from 350 nm to 395 nm. And the surface albedos of different underlying surfaces and different bands are compared. The transition from the center to the edge of the city is analyzed in detail. The results show that the surface albedo of ultraviolet waveband ranging from 350 nm to 395 nm goes up slightly with the increase of the wavelength, which is in good agreement with medium resolution imaging spectrometer (MODIS) data. In addition, the comparison of different observation areas shows that the surface albedo of urban area is larger than that of farmland, and the surface albedo of the center of the city is about 0.014 higher than the edge of the city. The difference also goes up with the increase of the wavelength.

Based on wavelength modulation spectroscopy (WMS) theory, a measurement method for gas parameters in combustion flow is put forward using spectroscopy fitting. Taking advantage of fitting harmonic signals of spectral lines, the integral absorbance, Doppler width, and collisional width are measured, and then the gas temperature, pressure, and vapour concentration are measured. The numerical simulation is used to investigate the effect of the integral absorbance and collisional width on harmonic signals, and experiments are conducted in a sample cell. The results show that the sensitivity of the harmonic signal spectrum to integral absorbance is nearly 1. However, the sensitivity of the harmonic signal spectrum to collisional width firstly increases and then keeps invariant approximately with the increase of collisional width. The fitting method has a high measurement accuracy. Compared with predicted values, the maximum measurement errors of measured temperature, pressure, and vapour mole fraction are less than 4%, 6%, and 5.5%, respectively.

In order to achieve the 3D distribution of internal components of the sample, we study the confocal 3D imaging reconstruction algorithm based on the synchrotron radiation infrared source. A confocal 3D imaging model is built by the single point detector scanning imaging on the basis of the principle and no-linear characteristics of synchrotron radiation infrared confocal spectroscopy 3D imaging. Based on the characteristics of the forward model, the model of the test sample (an absorption shell) is established to simulate the infrared confocal spectroscopy 3D imaging measurement process, and the raw data of the infrared confocal spectroscopy 3D imaging is obtained. Levenberg-Marquardt algorithm and the modified Gauss-Newton algorithm are used to reconstruct the sample model data. The reconstruction results show that the Levenberg-Marquardt algorithm can reconstruct the 3D information of the sample when no random noise is added to the forward model. When 1% noise is in the forward model, the reconstruction result of Levenberg-Marquardt algorithm has a relatively large deviation with the actual result, but the reconstruction result of modified Gauss-Newton algorithm is accurate.

The optical and mechanical properties and the microstructures of Nb2O5, Ta2O5 and SiO2 thin films prepared by the ion beam sputtering (IBS) method are studied. The effect of the assisted ion source voltage on the film properties is analyzed. The comparison among films prepared by the electron beam evaporation, ion beam assisted deposition and IBS is also conducted. The study results show that the films prepared by IBS possess better optical performance and microstructures, as well as higher compressive stress, hardness and Young modulus. The assisted ion source is beneficial to improving the optical performance, adjusting the film stress and reducing the surface roughness of thin films, however, it has relatively small influence on the hardness and the Young modulus. Under different assisted ion source voltages, the stress of prepared Nb2O5 by IBS is -152--281 MPa, that of Ta2O5 is -299--373 MPa and that of SiO2 is -427--577 MPa. Under proper process parameters, the extinction coefficient can be smaller than 10-4 and the film surface is smooth with a root mean square roughness of smaller than 0.2 nm.

A photonic approach for generating phase-coded waveform with continuously tunable carrier frequency based on cascaded Mach-Zehnder modulators (MZMs) is proposed. By setting the direct-current bias point and modulation index of the cascaded MZMs, we can select ±4 th-order optical sidebands. Finally, a phase-coded signal with a frequency multiplication coefficient of 16 can be generated at the output of the unbalanced temporal pulse shaping (UTPS) system through the controlling of the phase of the radio frequency driving signal applied to the cascade MZMs. In addition, when the dispersion values of the two linear chirped fiber Bragg gratings in the UTPS system are adjusted to be appropriate, the carrier frequency of phase-coded signal can be continuously tunable. Phase-coded signals with the carrier frequencies of 32 GHz and 19.2 GHz are generated with a 2 GHz radio frequency signal in the simulation. The simulation demonstrate the effectiveness of the proposed method, and also shows superior compression of the generated signal.

To satisfy the diagnosis requirement of laser driven inertial confinement fusion (ICF), an X-ray streak camera system with large dynamic range and high temporal-spatial resolution is developed. By optimizing the design of electronic optics, improving the production process of streak image converter tube, as well as producing and using the efficient devices, we improve the dynamic range and temporal-spatial resolution of the streak camera. The photocathode working length of the camera is designed to be 30 mm, and the focus voltage is designed to be 12 kV. Static and dynamic calibration systems are set up with the help of femtosecond and picosecond lasers. The results show that the spatial resolution of the camera system is greater than 20 lp/mm, the temporal resolution reaches 5 ps, the dynamic range reaches 2237∶1, and the sweep speed nonlinearity is smaller than 3%. Ultra-fast signals can be achieved by four gears of the camera with four sweep speeds. The camera has good performance and can be used for precise diagnosis of temporal-spatial and energy spectrum resolution in laser fusion research.