A new method is proposed to improve the monitoring accuracy of the thin film thickness in the photoelectricity extremum method. By improving the accuracy of the interpretation points, the random errors occurred when film coating is stopped in the optical extremum method are avoided. The nonlinear relationship between the monitoring signals and the optical thickness is transformed into a linear one and the optimum starting time is deduced by the treatment with algorithms, which avoids the influence of the nonlinear errors on the discrimination of extreme value points in the process of the monitoring thin film deposition.

The output characteristics of the one-dimensional In0.3Ga0.7As solar cells irradiated by a continuous wave (CW) laser with a wavelength of 1070 nm are studied by the finite element numerical simulation method. The validity of the model is verified by the calculation of the distribution of the recombination rate of the internal carrier in solar cells under a zero-bias. The effect of the bias voltage on the carrier recombination rate distribution and the influence of the structure of the solar cell on the photoelectric conversion performance are studied, and the current density-voltage curve of the solar cell is obtained under a laser power density of 100 mW·cm-2. The research results show that, with the increase of the forward bias, the recombination rate in the space charge region increases rapidly, so the forward bias is the main factor which influences the conversion efficiency. The short-circuit current of the cell decreases exponentially with the increase of the depth of the pn junction, however the open circuit voltage increases first and then tends to saturation with the increase of the thickness of the base area. These results provide a reference for the design of solar cells.

In order to configure the LED display control system dynamically and flexibly to satisfy different displaying requirements, we propose a new field-programmable gate array (FPGA) system structure, which achieves the 2048 pixel×32 pixel point to point addressing with minimal hardware resources based on the principle of two level caching and addressing code. Following the principle of gray controlling with high priority, we generate the control signal under the precondition of ensuring the display performance by taking 250 MHz high frequency as basic clock, associating with the synchronize control module. Experimental results verify that the display control system with the proposed FPGA frame could drive different resolutions, different scanning modes and different LED panels successfully. It improves the compatibility and flexibility of LED display control system.

Aiming at the current RGBW-based multi-primary display, we analyze the manufacturing problems of RGBW sub-pixel arrangement and the shortcomings of common RGBW permutation on displaying images, and present a new RGBWRGB sub-pixel arrangement structure. We also analyze the shortcomings of common gamut mapping algorithms with RGB to RGBW in the face of complex images, and try to improve the algorithms to fit the proposed sub-pixel arrangement, while maintaining the color saturation of the displayed image and lifting the brightness.

The photon counting image is scanned by multi-pixel photon counting detector point by point under the environment of 10-4 lx according to the principle of photon counting. To present more details and get a high definition image, the Bayes-Shrink threshold and the improved new symbol functions are used to realize image de-noising preprocessing first. Then, in the stage of image reconstruction, the low-frequency coefficients are set to zero to reconstruct the image with high-frequency coefficients after processing and it is set as a virtual channel to make the number of observation signals equal to the number of signal sources. Finally , the fast independent component analysis noiseless separation model is used to separate the photon counting image from noise by blind source separation. The experimental results show that the peak signal to noise ratios of the image are improved by 16.39%, 10.18%, 5.20%, respectively, compared with the soft, hard and new symbol function de-noising algorithm. The image after removing noise is also good to protect the edge details, and the visual effect is good.

The traditional random sampling consistency (RANSAC) algorithm can only perform coarse registration at low efficiency. To address this problem, an improved RANSAC fast point cloud registration algorithm is proposed herein. The proposed algorithm first combines the intrinsic shape signatures and fast point feature histogram algorithms to obtain feature descriptors and then employs pre-estimation and three-dimensional (3D) grid segmentation to improve the RANSAC algorithm. Finally, it is compared with the traditional sample consensus initial alignment algorithm. Our experimental results demonstrate that the proposed algorithm can quickly and accurately eliminate false matching points and solve the affine transformation matrix without secondary registration. In comparison with the traditional registration algorithm, the proposed algorithm demonstrates good robustness in large-scale 3D point cloud registration and significantly improves the registration efficiency while ensuring accuracy.

Starting from the Smoluchowski and Nernst-Planck equations, we demonstrated that the effective refractive index exponentially depends on the optical intensity in nanosuspensions with negative polarizability. We proposed a method for inducing curved waveguide structures by Airy beams in the nanosuspensions, and we theoretically analyzed the size of the curved waveguides. Moreover, we simulated various curved waveguides induced by different Airy beams, and discussed the relationship among the curved waveguide, Airy beams and the parameters of nanosuspensions. As shown in this research, by the control of the incident propagation behavior of Airy beams and the nanosuspension, the curved waveguides with different-sized structures can be induced. The results provide a novel method for creating a waveguide structure in a volume medium, which may find practical applications, especially for coherent beams combination, integrated photonics integration and optical manipulation.

The soliton self-frequency shift in a single-mode optical fiber has been studied by numerical simulation and experiment, with emphasis on the influences of various parameters of optical fiber and soliton pulse on it. Firstly, with split-step Fourier method for numerical simulation, it has been found that the soliton self-frequency shift increases with the increase of soliton peak power and nonlinear coefficient of the transmission fiber, and decreases with the increase of soliton pulse width and group velocity dispersion. Secondly, the soliton self-frequency shift effect in a 2-km-long single-mode fiber has been experimentally studied. By adjusting the peak power of the soliton, continuously tunable self-frequency shift with central wavelength from 5.44 nm to 26.64 nm has been achieved. The experimental results are consistent with the numerical simulation results. It has been shown that by flexibly adjusting the parameters of soliton pulse and optical fiber, the soliton self-frequency shift can be effectively tuned, which provides guidance for many practical applications of soliton self-frequency shift in optical fibers.

In space laser communication terminal with coaxial two pieces of reflective mirror of optical antenna, the highest energy part of the incident Gaussian beam can be obscured by the secondary mirror and its bracket, which leads to the loss of energy. In order to reduce loss of energy and improve the transmitting efficiency of terminal, we use an aspheric homogenizer in beam collimation system. The aspheric homogenizer can convert a Gaussian beam to a flattop beam, reduce the loss of energy led by the obscure of the secondary mirror and its bracket and collimate the beam. In the designed aspheric homogenizer system, incidence diameter d is 3 mm, exit diameter D is 12 mm, wavelength λ is 1550 nm, the glass material is BK7, the light intensity distribution is close to uniform on the output surface, and the divergence angle θ is 1.216 mrad. Finally, the spherical beam expander is contrasted with aspheric homogenizer, and the transmitting efficiency of terminal with aspheric homogenizer can be improved by 11.1% under the same condition.

Optical power density monitoring and temperature monitoring play an important role in the industrial production and day-to-day life. In this study, an all fiber-optic Mach-Zehnder sensor with PbSe quantum dots as the sensing material is fabricated on the basis of thermo-optical characteristic of PbSe quantum dots, and the sensor is tested under different optical power densities and temperatures. The results show that the power density sensitivity of the sensor to a light with a wavelength of 473 nm is 1.455 nm·(mW-1·mm2) and the temperature sensitivity is 0.67 nm·℃-1. This study not only realizes a high-sensitivity temperature sensor, but also lays the foundation for the thermal-optical devices made by PbSe quantum dots.

An optical time domain analysis system that is based on the interaction between the Brillouin gain and loss is proposed. A continuous wave propagates with reverse pulses at its Stokes and anti-Stokes frequencies in an optical fiber. Further, a narrow linewidth absorption peak is observed at the center frequency of the Brillouin gain spectrum. The linewidth of the absorption peak is approximately 1/5 of that observed in the Brillouin gain spectrum, and the frequency of the absorption peak is related to the Brillouin frequency shift. Therefore, the frequency resolution of a Brillouin sensor can be improved by the narrow linewidth absorption peak; thus, highly accurate temperature or strain measurements can be achieved. The experimental result depicts that the accuracy of the temperature measured by this scheme is more than doubled, compared with the traditional Brillouin gain spectrum based sensor.

Organic-inorganic hybrid perovskite is a promising new generation of photodetection materials due to its advantages of direct bandgap, high absorption coefficient, and high carrier mobility. Researches have shown that the perovskite nanomaterials have various morphologies, such as quantum dots, nanowires, nanorods, and nanosheets. In particular the anisotropic nanowires photogenerated carriers transmit efficiently along the axial direction, which enhances the charge extraction efficiency of the photodetector. In this paper, the perovskite nanowire arrays were prepared by the self-assembly growth method. When the mass fraction of precursor solution decreased from 10% to 0.2%, the diameter of nanowires was reduced from micron to hundred nanometer scale. A photodetector based on nanowire arrays obtained from 0.5% precursor solution was reported. The light-to-dark response ratio of the device was as high as 3.7×104, the external quantum efficiency reached up to 180.88%, and the dark current density was as low as 1 pA. All data were measured under 2 V bias voltage and 660 nm red light. The remarkable improvement of the photodetector performance is attributed to the radial uniform distribution of the nanowires and high crystallization quality.

The resonant frequency of dither mechanism of laser gyro varies with the temperature and other environmental factors, resulting in dither bias instability. The resonant frequency tracking technology of dither mechanism of laser gyro based on phase-locked loop is investigated. Based on the analysis of the transfer function of the dither mechanism and the phase-locked loop, the general relationship between the performance of the frequency tracking accuracy and the range of synchronization and the dither control parameters is worked out with the automatic control theory. The results show that the increase of the open-loop gain of phase-locked loop can reduce the steady-state error of frequency tracking and the synchronization range of phase-locked loop simultaneously. The low open-loop gain is adopted to initiate laser gyro, and then high-open loop gain is used to track frequency of dither mechanism. This strategy can guarantee the gyro resonance frequency being tracked automatically within 100 ms. The frequency tracking accuracy of the dither mechanism of a laser gyro machine based on phase-locked loop is better than 0.015 Hz with temperature ranging from -40 ℃ to 70 ℃.

A two-temperature model of copper is established based on the COMSOL simulation software. The effects of spot radius and laser energy on the temperatures of electron and lattice are numerically investigated via the control of variables, and thus the ablation morphology is predicted. The results show that the larger the spot radius of the single-shot laser, the smaller the ablation depth and the larger the ablation area. In contrast, the higher the laser energy, the larger the ablation depth and the larger the ablation area. The reliability of the simulation is verified by the experimental results.

The S136 mould steels with different particle sizes are fabricated by the selective laser melting (SLM) technique and the effects of particle size on phase compositions, microstructures, wear and corrosion resistance performances are studied. The results show that the forming parts with an average particle size of 22.8 μm have nearly no micro-pores, and their friction coefficient and wear rate are relatively small, but the corrosion resistance performance is the best and the corrosion weight-loss is 31.51×10-4 g·mm-2. If the average particle size is too small, there occur more cracks in the forming parts, however there occur more micro-pores when the average particle size is too large.

The application of a long-short composite pulsed laser in aluminum damage is explored. The influences of the reflectivity of material surface and the laser-material-coupling coefficient on damage mechanism are also discussed based on the thermal effect in laser-material interaction. The impact pressure and damage time of a short pulsed laser on aluminum and the thermal effect of a long pulsed laser irradiation on 2A12 aluminum are numerically calculated. A long-short composite pulsed laser irradiation experimental system is designed. The results show that the designed long-short composite pulsed laser can be used to damage 2A12 aluminum effectively and rapidly. Moreover, the damage effect is superior to that by the single long or short pulsed laser, and the damage time is effectively reduced.

The simulation experiment on the laser cutting of the AZ31B magnesium aluminum alloys is conducted based on the thermodynamic equations and it is found that the cutting quality of samples is better when the laser cutting speed is 8 mm·s-1. The process experiment on the laser cutting of the 1.0-mm-thick AZ31B magnesium aluminum alloys is carried out and the single-factor experimental data related to the effects of laser power, cutting speed and defocusing is obtained. The scheme of three-factor orthognal experiment is also proposed. The optimum process parameters for a laser cutting of the 1.0-mm-thick AZ31B magnesium aluminum alloys are a laser power of 120 W, a cutting speed of 0.5 m·min-1, and a defocusing of 1.2 mm, which are basically consistent with the theoretical calculation values.

The entanglement dynamics problem between two non-identical qubits in the Tavis-Cummings model without rotating wave approximation is discussed by the extended coherent state (ECS) method. The effects of the qubits with different transition frequencies but with a same coupling strength and the optical fields on the entanglement evolution between two qubits are investigated. The research results show that, in the case of the weak coupling, the entanglement evolution between two qubits is the same when the transition frequency of one qubit is identical to the optical field frequency but the transition frequency of another qubit is symmetric detuning from the optical field frequency. In contrast, in the case of the strong coupling, the entanglement evolution between two qubits is no longer same under the cases of two symmetric detuning due to the effect of the non-rotating wave term.

Based on the direct quantum feedback method for controlling the quantum coherence of a qubit, the dynamical evolution of the quantum coherence of a qubit which interacts with a single mode cavity is investigated. By using the l1 norm of coherence and the relative entropy of coherence to quantify the quantum coherence, the effects of the quantum feedback and the external driving on the evolution of the quantum coherence are analyzed. The research results show that, the dynamical characteristics of these two types of coherences are the same. The quantum feedback slows the decay of the quantum coherence and plays a certain protective role. When the external driving is considered, the coherence of the steady state in the long limit becomes zero under the strong-driving condition, while the quantum coherence reaches a stable maximum value under the non-strong-driving condition.

By the analysis of the signal transfer model of a complementary metal oxide semiconductor (CMOS) camera, the transfer function model of signals and noises in this CMOS camera is builded and the linear relationship between the output image signal and the image noise of this camera is deduced. The testing system of the performance parameters of this camera is established, and the conversion gain, the maximum signal-to-noise ratio (SNR) and the dark noise in time domain of this CMOS camera are tested. The test value is compared with the index value, and the results show that, the performance parameters of this camera can be can accurately tested with this proposed test method, which is suitable for the test of the photoelectric parameters of most types of cameras.

Hash method is an effective method for generating hash codes in large-scale image retrieval. The current hash method extracts the characteristics of the whole image first and then generates hash code, but the obtained hash code is not very precise to obtain more precise retrieval effect. Aiming at this problem, we propose a new method. First, we use a convolutional neural network to extract image features. Then, we adopt hash algorithm and generative adversarial network of input binary noise variable to learn image binary hash code, and carry out image similarity comparison by using hamming distance. Finally, we complete the effective retrieval of image data. Experiments on standard image data sets show that this method can effectively perform image retrieval, and the retrieval performance is improved than other methods.

In order to detect objects more accurately in images, an object detection algorithm based on hard example mining and residual network is proposed, which takes faster regional convolutional neural network (Faster R-CNN) as a benchmark. The working principle of Faster R-CNN is described based on deep learning, and the shortcomings and improvement methods of the algorithm are analyzed. Specifically, a deeper residual network is adopted to replace the original ZF or VGG network to extract more effective deep convolution features. In order to enhance the generalization ability of the learning network model, the network parameters are updated with hard examples during training. The experimental results on Pascal VOC2007, Pascal VOC2007+Pascal VOC2012 and BIT show that compared with Faster R-CNN, the proposed method improves detection accuracy by 3.5%, 7.1%, 6.4%, respectively, on the above three datasets.

Aiming at the problem that the face recognition algorithm based on double space local directional pattern (DSLDP) only uses differential operation to extract features, we present a novel approach based on the double operation local directional pattern (DOLDP) for face recognition. Firstly, 3 pixel×3 pixel neighborhood of facial image are convolved with eight Kirsch template operators to obtain eight directions of edge response values. Then, the neighboring edge response values are countered and summed in counterclockwise directions to obtain two sets of eight-direction edge response differences and sums, and the two sets of values are taken as absolute values. Finally, the directions of the maximum values of the two sets of edge response values are encoded into a two-digit octal number to form a DOLDP code. The experimental results on YALE, ORL, AR and CAS-PEAL face databases show that the proposed method combines the sum space and the difference space face feature information, and achieves a better recognition effect. Compared with the intensity space, the sum space face feature information plays a smooth role and shows stronger robustness to light, expression, and occlusion.

Information entropy weighted by image variance is a robust quantitative indicator describing the complexity of image. It can achieve good results by using information entropy weighted by image variance to detect the infrared small target. However, it is difficult to apply in engineering application due to its complex calculation and poor real-time performance. To recognize the small target regions under infrared complex sky background quickly, we improve the traditional image filtering algorithm, which uses information entropy weighted by image variance. Images are segmented according to their saliency first. Significant regions are selected roughly, and only the information entropy weighted by image variance of the dual-mode regions of the salient regions is calculated. Then, the candidate target regions are recognized according to the typical regional features of the information entropy weighted by image variance of the dual-mode regions under complex sky background. The experimental results show that the proposed algorithm can eliminate the disturbance of the complex sky background, and reduce the running time of the algorithm.

Fluorescence molecular tomography is a robust molecular imaging technology with low side effects which is a very hot topic in photobiology always. The fluorescence molecular tomography has better reconstruction results generally, if the excitation plane is close to the fluorescent targets. To find better excitation planes, we propose a location method of excitation planes based on mixture Gaussian distribution. Firstly, the method uses several excitation sources to obtain the living organism external surface distribution of the emitting light. Secondly, the Gaussian mixture model with pruning strategy is used to fit the distribution. Finally, the number and locations of the excitation planes are automatically determined according to fitted peak values. Fluorescence molecular tomography inverse problems are built based on the excitation light source of new excitation planes, and fluorescent targets are reconstructed using the inverse problems. Experimental results demonstrate that the fluorescence molecular tomography reconstruction results depending on the new excitation planes are much better than the results depending on original excitation planes.

We study a high-resolution melting (HRM) analysis method based on laser signal in self-made Fabry-Perot (FP) optical microcavity with high quality. We utilize the F-P optical microcavity as the micro-laser cavity, intercalated saturation dyes as the gain media and emitted laser signal as the detection signal to achieve high-resolution detection and screening of mismatched DNA by temperature scanning. We investigate on the melting curves of target DNA and mismatched DNA with 25 base-pairs and 50 base-pairs, respectively. Theoretical and experimental results indicate that the HRM detection method based on laser signal has lower melting temperature and higher signal-to-noise ratio.

Aiming at the large space indoor high-precision navigation application in complex environment, we propose an algorithm for integrated mapping and positioning based on workshop Measurement Positioning System (wMPS) and lidar. The wMPS carries out precise estimation on the position and orientation of the lidar, and integrates the point cloud data of the lidar to complete the mapping of the grid map. Therefore, the robot can recognize the information of the surrounding environment when navigating. Then considering that the wMPS measurement information is easy to be missing during navigating, we use particle filter algorithm to inversely calculate the position and orientation of the lidar according to the real-time point cloud data from lidar and the gird map. Finally, the results of particle filter are processed with the linear Kalman filter, and the algorithm simulation and experiment are conducted. The simulation and experimental results show that the integrated navigation system ensures the reliability of the map and the accuracy of the dynamic navigation, which greatly improves the overall performance of the navigation system.

An experimental platform of laser ultrasonic is built based on the thermoelastic effect and the laser interference reception method. The directivity of transverse waves generated by the incident laser and the effects of the inner surface defect on the ultrasonic signal are observed. According to the B-scan image, the defect location at the inner surface threaded part is detected. The experimental results show that the transverse wave with the largest power generated by the laser point source is in the direction of normal angle of 32.1°, which is consistent with the result of the theoretical analysis. The B-scan image combined with the peak-angle variation of the signal is used to locate the inner surface defect and to measure the defect width. This study can provide an experimental basis for industry application and popularization of laser ultrasonic.

The phase extraction technique based on the traditional spatial-temporal fringe pattern method is studied, which is perfected by correcting the phase-shift of the phase-shift fringer pattern with a linear carrier frequency. This technique is also applied to the fiber-optic interferometer projection for three-dimensional surface measurements. Compared with the traditional four-step phase-shifting method, overlapping averaging four-frame algorithm, and the advanced iterative algorithm, the phase extraction technique based on the spatial-temporal fringe pattern method can effectively suppress the phase-shifting error caused by phase-shifting inaccuracy of the phase-shifter in the fiber-optic interference fringe projection device and environmental factors. Further, it can eliminate the high-frequency noise caused by factors, including stray light.

A Rudin-Shapiro photon sieve (RSPS) based on the Rudin-Shapiro (RS) aperiodic sequence is proposed. The RSPS can be fabricated conveniently and generate two images with low chromatic aberration. The plane wave angular spectrum theory is used to study the focusing properties of the RSPSs with different ratios of the diameter of the smallest hole of the sieve to the width of the outermost ring of the corresponding zone plate. The numerical simulation results show that when the ratio is 1.398, the RSPS generates twin foci with the same high intensities in the axial direction. Compared with the RS zone plate (RSZP), the RSPS has more tolerance in the fabrication. It is found that the chromatic aberration generated by RSPS is lower than that by Thue-Morse zone plate (TMZP) and approximately the same as that by RSZP. Hence, the RSPS will have potential applications in the fields of the polychromatic imaging, X-ray microscopy and so on.

To improve teaching video compression ratio and production efficiency, considering that effective information of teaching video normally remains for a long time and locates in fixed area, a teaching video compression algorithm based on key frame detection and indicator movement modeling is proposed. Firstly, the projection area is detected as the effective region of each frame to reduce the knowledge redundancy. Then, the type of each frame is determined using change detection, and the indicator motion model is established for the mouse cursor and the laser point to reduce more knowledge redundancy. Finally, a corresponding play algorithm based on the key frame coding and indicator motion model is designed to replay the videos by use of OpenCV and MFC. The experimental results showed that, for the teaching videos with effective information in the projection area, the proposed method achieved an average of 88% bitrate reduction than that of H.264 under the same peak signal to noise ratio value. Besides, the encoding and decoding efficiency of this algorithm could meet the real-time requirement. Without extra manual editing, the production and transmission efficiency of online courses could be significantly increased.

This paper presents an image recognition method based on feature matching fusion and improved convolutional neural network. Aiming at the problem that the texture features extracted by the local binary pattern (LBP) descriptors are limited and cannot describe the image edge and direction information effectively, the feature extraction of the training set is performed in the convolutional neural network by the histogram of oriented gradient (HOG) and LBP hierarchical feature fusion method. Then the extracted feature pictures are input into the improved convolutional neural network for training and recognition. The simulations are performed on ORL, YALE and CAS-PEAL face databases with ReLU as the activation function and the output layer with the Softmax classifier, and trained on the TensorFlow framework. The recognition rate of the proposed method reaches 99.2%, 98.7%, and 97.2% respectively, which is higher than other algorithms for comparison.

Optical tweezer has become a powerful and flexible tool for trapping and manipulating the micro-nano particles through a gradient force well formed by a highly focused laser beam, and it has a wide applications in the fields of biology, physics, chemistry, and medicine. Based on the 4π focusing system, the tight focusing characteristics of the radially polarized Gaussian beam and its radiation force to the metal particles are theoretically studied and compared with the results of traditional single-lens focusing system. Furthermore, the influence of the off-focus distance and the off-axis distance on the trap stiffness is also investigated in detail. Numerical results show that a focal spot with three-dimensional and spherical structure can be obtained in the 4π focusing system via the suitable parameters. This spherical focal spot can largely enhance the transverse and longitudinal trapping forces, and consequently enhance the trapping stability of metal particles of optical tweezer system.

The lidar intensity value can reflect the target reflection characteristics to a certain degree, but it cannot be directly used as an important feature of target classification because of the influence of various factors, such as distance, incident angle, and atmospheric attenuation effect. For the lidar commonly used in vehicle, based on theoretical analysis of the lindar intensity value, the influence of other factors on the intensity values is fixed by the experimental method, and the relationship between the lidar intensity value and the parameters of the target reflection characteristics is established. The experimental results show that the angle factor of semi-elliptical model can better describe the variation of the target diffuse intensity value with the incident angle. And the lidar intensity value linearly transforms with the received power. Outside the blind area, the variation of target diffuse reflection value with distance follows the negative index law. In practical applications, we combine lidar motion to obtain intensity values of the same target at different relative positions. We can use the obtained intensity values to substitute into the model, invert the parameters related to the target reflection characteristics, and realize the differentiation of different targets.

As part of the next-generation American environmental meteorological satellites, the NPP/Joint Polar Satellite System （JPSS） has introduced the Visible Infrared Image Radiometer Suite （VIIRS） Day and Night Band （DNB） low-light day/night imaging band. DNB employs the ability of defense meteorological satellite program （DMSP）/Operational Linescan System （OLS） to collect global low-light imaging data, which has resulted in a considerably optimized performance. The nighttime city lights can be detected using the VIIRS DNB data. Because of the urgent requirement for city lights in the rapidly developing regions that are located in mid-eastern China, this study comprehensively considers the influence of various important factors, including clouds, moonlight, sunlight, lightning, and fire, while identifying and rejecting low-quality data and non-light characteristic data in addition to adopting VIIRS DNB multi-day nighttime data fusion. The average of multi-day non-moonlit effective radiance data is used to generate monthly nighttime city lights in mid-eastern China. The results that are obtained by analyzing the radiance value of the product are comparable with those obtained by analyzing the National Oceanic and Atmospheric Administration （NOAA） city lights business product. The city lights fusion product algorithm, which is developed in this study, provides the technical foundation to develop the back-end generative application and for the inversion of the remaining low-light remote-sensing nighttime products.

Balanced coherent detection technology is widely used in coherent optical communication and microwave photonic systems because the sensitivity can be significantly improved by the technology. Balanced photodetector, as a key device for coherent detection, has been the workhorse of research in recent years. This paper introduces the principle, structure and some parameters of balanced photodetectors and provides an overview on recent advances of the device, with analysis on current problems and development trends. With the development of optical communication technology, the performance of balanced photodetectors will be promoted in terms of speed, responsivity, radio frequency power, and common mode rejection ratio. Moreover, the package module should be highly integrated, with lower cost and lower power consumption in the future.

The active support system can provide the capacity of surface figure control and positioning for large optical telescope primary mirror, hence the active control technology of monolithic mirror surface figure and co-phasing technique of segmented mirror has been developed rapidly. The applications of active support system and actuator in large optical telescope in recent years are reviewed and summarized. Several common active support systems and actuators are concludes. Their features are compared, and the relationship between active support system and actuator is proposed. The future development of active support system and actuator applied to telescope primary mirror is prospected in the end.

The research progress in key technologies on various types of near infrared microspectrometers, such as array detector, scanning grating, filter, Fourier transform, and Hadamard transform, is reviewed. Different instruments are discussed in detail, including their advantages and disadvantages, suitable application fields, and current problems. Finally, the application and development tendency of near infrared microspectrometers are summarized and prospected.

Brillouin optical time-domain analysis (BOTDA) is an important distributed optical fiber sensing technology based on optical time domain reflectometry (OTDR) and stimulated Brillouin scattering (SBS) effect. It has many advantages such as multi-parameter sensing, high accuracy, and long sensing distance. Accordingly, it is especially suitable for fault location and health monitoring in high-risk fields such as large infrastructure, petrochemical industry, power communication network and undersea optical fiber cable. The traditional BOTDA is only capable of static measurement, which seriously restricts its application scenarios and development prospects. The dynamic measurement techniques using BOTDA are summarized and the measurement principles are also introduced in detail. The basic factors and research difficulties which affect its sensing speed are discussed, and the future development prospect is forecasted.

The rapid generation technology of computer-generated hologram (CGH) is the key technology of the holographic three-dimensional display system. The calculation of holograms based on point source model is an important branch of computational holography because of its simple model and flexible operation. However, the calculations of point source model for holograms are enormous. In order to meet the requirements of real-time display, the methods of improving the calculation speed of hologram have been put forward continuously. The development of the fast algorithm of point source model is reviewed. Based on the principle of CGH, according to the research direction of the fast algorithm of point source model, it is divided into two categories: algorithm, and combination of algorithm and high-performance hardware. Through the analysis of the implementation methods and existing problems of all kinds of algorithms, the direction of improvement is pointed out, and the contribution of high-performance hardware to CGH is introduced. Finally, the future research direction of point source model fast algorithm is prospected.

Owing to the special structure between fiber and crystal, single-crystal fiber has the merits of both fiber and crystal medium. For its advantages of high gain, simple thermal management, and small nonlinear effect, single-crystal fiber is widely used in ultrashort pulse amplification. The experimental research progress in single-crystal fiber amplifiers is reviewed. Firstly, two kinds of single-crystal fiber preparation methods including micro-pulling down technique and laser heated pedestal growth technique are introduced. Then along the principal clue of various means to improve the amplification performance of single-crystal fiber amplifiers, we summarize the research progress in Nd∶YAG and Yb∶YAG doped single-crystal fiber amplifiers and some new amplifiers with special structures in ultrashort pulse amplification. Finally, the applications and future development trends of single-crystal fiber amplifiers are prospected.

Because of the influence of certain operational parameters, such as sensor status, imaging mechanism, climate, and illumination, hyperspectral remote sensing images suffers from serious distortion. Intrinsic image decomposition (IID) is an extensively used image processing technology in the field of computer vision and graphics because it can acquire the essential features of the images that are being processed. IID is introduced to hyperspectral image procesing to decoposite the original images. Accordingly, we propose a hyperspectral IID method based on automatic subspace partitioning. Firstly, the hyperspectral image is divided into subspaces, and the optimal decomposition-based IID method is applied to each subspace. The reflectance intrinsic image that is obtained from the decomposition is further subjected to hyperspectral image classification processing. The experimental results obtained from this study indicate that the proposed method can considerably improve the accuracy of hyperspectral image classification.

Based on the thermodynamic theory of the optical thin-film components, the thermal analysis model of thin-film components irradiated continuously by a high-power laser is established and the processes of the melting damage and the thermal stress damage induced by different kinds of surface impurities are analyzed. The statistic number of impurities inducing thermal damages on thin-film components with different sizes and different surface cleanliness levels is shown and the total thermal damage area of thin-film components induced by impurities is analyzed quantitatively. The exposure time when the thermal damage area of thin-film components exceeds 3% of the total area is calculated as well. The research results show that, under the high-power laser continuous irradiation, the melting damage and the thermal stress damage on thin-film components can be induced by impurities with sizes within a certain range. The damage way is closely related to the impurity type. The larger the aperture and the higher the cleanliness level of thin-film components, the larger the number of impurities with sizes within a certain range where the thermal damages on thin-film components can be induced. Furthermore, the damage point area of the thermal stress damage induced by a single impurity is larger than that of the melting damage.

A real-time and high-speed all-optical quantization scheme by slicing supercontinuum spectrum is studied experimentally in time domain. The optical filters are used to slice the supercontinuum spectrum generated by different intensity optical sampling pulses, so as to realize the real-time quantization of the optical sampling pulses. Optical sampling pulses with the repetition rate of 10 GHz are coupled into a 400-m highly nonlinear fiber after power amplification to generate supercontinuum spectrum, which is sliced by three tunable optical filters at different wavelengths. Real-time all-optical quantization with sampling rate of 10 GSa/s and quantization accuracy of 2 bits is realized based on slicing supercontinuum spectrum, and the extinction ratio among the quantized outputs can reach up to more than 10 dB.

Fast and accurate detection of maneuvering armored targets is an important performance requirement for low altitude unmanned aerial vehicles, but the rotation invariance of the current mainstream detection methods is not enough to deal with the challenge effectively. Combined with deep convolution neural network (CNN), we propose a low altitude armored target detection method based on rotation invariant Faster R-CNN. This method introduces the rotation invariant layer on the basis of the original frame of Faster R-CNN to strengthen the invariance of the target′s CNN feature before and after rotation by adding regularization constraints on the objective function of the model. In the experiment, three typical models of armored target are selected to simulate the low altitude reconnaissance environment under different scenes indoors and outdoors, reconnaissance simulated images of the targets are used as sample data for model verification, which are obtained by using a polarizing hyperspectral camera. In the multi model comparison test, the improved model increases the mean average precision by 2.4% on the original basis and achieves the best test result, which preliminary verifies the effectiveness of the improved method.

In the current saliency detections based on deep learning, how to make full use of the convolution features at all levels is the key issue. In order to solve this problem, we propose a saliency detection method based on full convolution neural network, which is a fusion of all convolutional features. Firstly, all the convolution features are mapped to multiple internal scales, and the saliency maps are predicted by combining the convolutional features of each level on each scale. Then the fused saliency maps are obtained by fusing the saliency maps of each scale. Finally, smooth saliency maps and optimized salient boundaries are obtained through full connected conditional random fields. Experimental results show that the proposed method has higher accuracy, recall rate and lower average absolute error in ECSSD database and SED2 database, and provides more reliable pretreatment results for target recognition, machine vision and other applications.

Human fatigue detection based on the machine vision methods is non-invasive, fast, and accurate and is unhindered by weather conditions. Owing to these advantages, this technique has gradually become a hot research topic worldwide. However, it is easily affected by complicated illumination and changes in the pilot position. To solve this problem, on the basis of previous studies on driver fatigue detection under complicated illumination conditions and postural changes, we propose a fatigue detection method based on the real-time enhanced constraint local model. First, the collected images are subjected to real-time high-dynamic-range enhancement. Then, the enhanced image is used to model the driver′s face in order to extract his/her vision and percentage of eye closure characteristics. Finally, the fatigue state is detected and an identification method based on Bayesian confidence networks is established. Our experimental findings show that the proposed method robustly detects the fatigue states of drivers under complex illumination and change in position.

In order to improve the accuracy and real-time performance of iris location, an iris boundary location algorithm based on double circular compensation is proposed. The approximate radius compensation and approximate circle center compensation are used as the core to locate the inner and outer boundary of the iris by the proposed algorithm. When locating the inner boundary, the minimum gray average method is used to perform coarse localization, and the approximate circle center compensation method is used for locating the inner boundary exactly after extracting the inner boundary images. When locating the outer boundary, the approximate radius compensation method is used to perform coarse locating, and the exact locating is performed by the approximate circle center compensation method. Experimental results show that 2000 different iris images from CASIA for verification are randomly selected. The average locating time of the algorithm was 0.29 s, and the locating accuracy rate is 98.1%. The proposed algorithm is more accurate and faster than the contrast method.

In order to meet the detection and identification requirements of products in the ultraviolet (UV) band, we design a set of UV industrial inspection optical system with wide spectrum, large field of view, large aperture and compact structure. Its design requirements are: UV industrial inspection lens with working wavelength of 240-320 nm, full field of view angle of 40°, system focal length of 15 mm, F number of 3, and total system length of less than 60 mm. The system uses MS20-UV type UV charge coupled device (CCD) with a resolution of 1920 pixel×1080 pixel, pixel size of 5.48 μm×5.48 μm. Considering the cost and image quality, the design adopts the global transmissive and non-glued solution. The system uses an anti-telescopic objective lens as the initial structure and uses Zemax optical software to design. The tolerance analysis is performed on the design results to determine the source of the tolerance error in combination with the requirements of the process requirement. And the relevant structure is optimized. Monte-Carlo simulation results before and after optimization are compared. Finally, the UV lens is designed with the modulation transfer function (MTF) of full field of view of more than 0.5 in the range of 100 lp/mm, the field curvature of less than 0.1 mm and the distortion of less than 1.3%. Compared with other UV systems, this system has the advantages of high image quality, high resolution, low distortion, short focal length and compact structure.

A uniform square spot lens is designed by the supporting quadratic method (SQM) and an algorithm for the interpolation and reconstruction of the curved surfaces based on the Zernike polynomials is developed. The mapping conservation points are chosen on the initial base surface for each sub-surface and the 9-point system is partitioned according to the adjacency relationship. As for this 9-point system, the linear equation group is established by using the lower order terms included in the polynomials and the partial curved surface is constructed with its unique solution. All the partial curved surfaces are spliced together and thus a complete and continuously smooth curved surface is obtained. This algorithm can simplify the later optimization of the initial base surfaces in SQM to a certain extent, which can ensure a good design effect of the uniform square spot lens.

We propose a double-stub resonator (DSR) based on a metal-insulator-metal (MIM) surface plasmon waveguide. The plasmon-induced absorption (PIA) effect is achieved by mutual interference between the two stub cavity modes. The transmission properties of the system are numerically simulated by the finite-difference time-domain (FDTD) method. Anomalous dispersion phenomenon can be achieved with the PIA windows based on the simulation analysis of phase response characteristics. Such anomalous dispersion effect can be used for realizing fast light effect in surface plasmonic waveguide. In addition, a multi-switch function based on the PIA effect is proposed. By simulating the influence of the changes of stub cavity′s length and the refractive index on the PIA window, the structural parameters of the switch function are optimized. Structures with these characteristics have potential applications in surface plasmon optical switches and filters and so on.

It is crucial to enhance the far-field directional emission of quantum dots (QDs) in order to improve the sensitivity of bio-detection. In this paper, we propose a novel and effective photonic structure composed of periodically distributed silver nanodomes on a SiO2 film and silver back reflectors. The radius of each nanodome structure is 162.5 nm, and the QD is located in the gap between adjacent silver nanodomes. The finite-difference time-domain (FDTD) method is used to investigate the effect of different QD positions and different quantities of nanodomes on the emission intensity of QDs. The results show that when there are four nanodomes and the QD is located in the gap between adjacent nanodomes, the emission intensity of QDs in the far field is improved by more than four times compared to that of the nanodome free structure. The proposed structure will improve the sensitivity of biological detection.

The sodium lidar can achieve atmospheric parameters measurement in the mesosphere by emitting a single-mode narrow-linewidth 589 nm pulsed laser to stimulate the alkali metal sodium atoms in the atmosphere at 80-110 km altitude and obtain the resonance fluorescence scattering echo signal. Laser frequency stabilization and shifting are the key technologies for the narrow-band sodium Doppler lidar to realize mesosphere atmosphere wind and temperature detection with high resolution. This paper introduces the methods of seed laser frequency stabilization and shifting applied in a sodium lidar system. The precise locking of 589 nm seed laser frequency is achieved using the normalized saturation absorption spectrum signal, and the long-term stability of laser frequency is about 2.2 MHz; three laser operating frequencies on the sodium D2 line spectrum are obtained by designing a cascaded double-pass acousto-optic frequency shift device. Through the experiment with sodium Doppler lidar, the wind and temperature measurement results with high resolution are obtained, and the temperature profile from lidar is compared with the result measured by satellite.

For coastal waters, the radiation signal received by a sensor is affected by the high reflectance of the terrestrial land, which can result in a reduced image contrast and significant adjacency effect. Eliminating atmospheric attenuation and the adjacency effect and obtaining accurate water surface reflectance are important prerequisites for quantitative water color remote sensing. Based on the in situ spectral data of the coastal waters of Taihu Lake on April 29, 2016 and a synchronous Gaofen-1 wide field of view (GF-1/WFV) image, the 6S model was used to eliminate atmospheric attenuation, and a nuclear function is used to express the atmospheric point spread function to correct the adjacency effect. The experimental results show that the image after correcting the adjacency effect is clearer with an increased contrast and more abundant information regarding the water body. Compared with the correction results of the 6S model, the average relative errors of the three in situ synchronous samples after correcting the adjacency effect are reduced by 10.8%, 5.24%, and 10.39%, respectively. Correcting the adjacency effect improves the accuracy of radiation detection of the remote sensing reflectance above water surface.

Based on the Rayleigh scattering theory, the distribution pattern of the sky light polarization is simulated. The distribution pattern of the sky light polarization in the East China Sea is tested by the photographic full-sky polarization pattern testing system. A method for determining the meridian by the polarization azimuthal angle information is proposed. The research results show that, when the solar altitude angle is larger than 85° or less than 25°, the meridian direction can be positioned by means of the polarization azimuthal angle. The meridian direction is consistent with the theoretical simulation result.

Graphene is a two-dimensional semiconductor material with carbon nanostructure, and it has a wide application prospect in the fields of microchip and materials science. In recent years, the research production of graphene in terahertz band have been used in optoelectronic devices , while the study of other two-dimensional sulfur materials are rarely reported. Therefore, we mainly study the absorption mechanism of graphene and graphene-like MoS2 WS2 on high resistance silicon substrate in the terahertz band from theory and experiment. We measure the terahertz spectrum of graphene-like material with terahertz time-domain spectroscopy system at different pump light powers, and obtain their transmission spectra, photoconductivitis and absorption spectra with calculation and analysis. Results show that we can similarly modulate the two kinds of graphene-like materials by changing the condition of light field. In addition, the absorption of WS2 and MoS2 are caused by the exciton of disulfide when imposing light field.

Band selection can preserve the physical meaning of hyperspectral data while reducing dimension, and has application in many aspects. The cluster of affinity propagation (AP) algorithm is according to the correlation of data points, and the AP algorithm regards all data points as potential clustering centers. We propose a band selection method based on AP clustering, which uses spectral information divergence and spectral correlation angle (SID-SCA), and spectral information divergence and spectral gradient angle (SID-SGA) to improve the similarity calculation in AP algorithm. The reducing dimension results are input into the support vector machine （SVM） classifier to classify, and the classification accuracy is calculated and verified using the data set Indiana Pines. The experimental results reveal that the proposed method can better extract the information of the band and obtain a high classification accuracy.

The methods that are used to produce microwaves can affect the strength and stability of plasma spectral signal in laser-induced breakdown spectroscopy (LIBS). Leaf vegetables that contain cadmium (Cd) at mass fraction of 1.5×10-5 are used as the subjects of this study. The influences of microwave probe with four shapes, such as the single-needle, double-needle, single-ring, and double-rings microwave probes, are studied. The three microwave loading positions, which include the front, middle, and end of the probes, are also investigated. The form of microwave function in the microwave assisted LIBS (MA-LIBS) experiment is optimized to achieve better auxiliary enhancement effect. After analyzing the spectral lines of the MA-LIBS, we find that the plasma spectral signal of Cd element is increased by 2.58 times, whereas Si and P elements is increased by 2.70 and 3.08 times, respectively. Suitable microwave probe shapes and loading positions can improve the effects of microwave-assisted enhancement, which further improves the performance of the LIBS technology in the application fields.

A high power portable femtosecond fiber laser system is demonstrated. This system uses semiconductor saturable absorb mode-locked fiber laser as the seed source and selects the wavelength through the fiber Bragg grating. The seed light output by the oscillator enters the main amplifier system after being pre-amplified by two-stage single-mode ytterbium-doped fiber and first-order double-clad ytterbium-doped fiber. The main amplifier uses a large mode field of ytterbium-doped photonic crystal fiber to amplifier, and effectively reduces the effect of nonlinear effect on the pulse by controlling the nonlinear accumulation in the amplification process. An acousto-optic modulator is added to make the output repetition frequency adjustable, and the output pulse is compressed through the transmission grating to obtain the laser output. The final output has an average power of 1.34 W, a repetition rate of 300 kHz, a working wavelength of 1030 nm and a pulse duration of 202 fs, and the corresponding single pulse energy is 4.5 μJ, peak power is 22 MW. The complete laser system is portable, stable, and cost-effective, and can be widely used in production.