In recent years, the resolution of far field optical imaging has broken the limit of diffraction limit with the appearance of new fluorescent probes and the improvement of imaging methods. The fluorescence microscopy based on structured illumination has become one of the main superresolution imaging techniques owing to its advantages such as fast imaging speed and light toxicity. The most critical technologies in structured illumination microscopy are concentrated on rapid control of the high-quality illumination pattern and the image reconstruction method. An unsuitable method will generate the artifacts in the reconstruction super-resolution image and that may confuse the scientific evaluation of biological morphology. Several typical structured illumination microscopic super resolution reconstruction algorithms are compared. It is proven that the image transform of the structured illumination microscopic super resolution reconstruction algorithm based on image recombination transform can effectively solve the low structure light modulation under the super-resolution image reconstruction problem, and reduce the excitation power of the structured illumination microscopy.

In recent years, people pay more and more attentions to the development of reversibly photoswitchable fluorescent proteins (RSFPs), because those fluorescent proteins can greatly promote the development and application of live-cell super-resolution imaging techniques. Reversibly photoswitchable fluorescent protein can be repeatedly and reversibly modulated by different wavelength lights, thus it can be widely applied in many fields such as high density data optical storage, the measurement of photochromic fluorescence resonance energy transfer, and super-resolution imaging based on reversible saturable linear fluorescence transitions (RESOLFT) principle. The recent development of green RSFPs is reviewed from the key amino acids involved in RSFPs design. We also make a brief discussion of the relationship between structure and optical characteristics of RSFPs, serving as the reference for further structure-guided development of better RSFPs.

Structured illumination fluorescence microscopy (SIM) is a wide field microscopy imaging technique which can circumvent the Abbe diffraction limit. It has various potential applications in the biomedical research due to its advantages of non-invasivity, fast imaging speed and low photon damage and so on. Based on the basic principle of structured illumination microscopy imaging system, the super-resolution image reconstruction algorithm, the sources of artifacts in reconstructed image and optimization methods are analyzed. The SIM imaging system based on laser interference is discussed combining custom built linear/non-linear structured illumination microscope. Design of the control system for hardware synchronization, and few key questions during optical alignment are discussed in detail. Imaging of cell skeleton is done with the linear SIM microscopy system and the comparative tests verify the reliability of the independent-developed image reconstruction algorithm. At last, the future improvement and application in biology of the SIM technique are discussed.

Super-resolution localization microscopy can achieve the ultra-high spatial resolution up to several nanometers by the single molecule localization and reconstruction from thousands or even tens of thousands of single molecule image, which provides unprecedented opportunities for studying the cell structures and biological phenomenon. However, the aberration (originating from the imperfection of the optical system or the inhomogeneity of the sample itself) distorts the raw images from single molecules, which decreases the final spatial resolution and even results in wrong results. The effects of several representative aberrations on super-resolution localization imaging are quantitatively characterized, and an aberration correction method based on the sample image itself is proposed. Simulation and experimental results show that the aberrations cause distortion of the point spread function and the decrease of the spatial resolution. The image quality can be restored by using the proposed aberration correction method.

Optical microscope is widely used in many fields such as life science, biomedicine, clinical diagnosis and material science. But due to the optical diffraction limit, the traditional optical microscope cannot reach the nanoscale resolution, which greatly restricts the development of science and medicine. In recent years, a number of super-resolution fluorescence microscopy techniques are invented to overcome the diffraction barrier and the imaging resolution are improved at various degrees. Among various super high resolution microscopes based on different principles, the stochastic optical reconstruction microscopy (STORM) reaches the highest spatial resolution, about tens of nanometer, and therefore it can realize the single molecule detection. The principle, experimental method and its application of STORM super resolution microscope are mainly introduced.

Single image super-resolution (SR) restoration is an ill-posed inverse problem, in which regularization restriction is done with image priori knowledge. One single image SR method is proposed which simultaneously taking external example and internal self-similarity into account. Here the external knowledge is learned by convolutional neural network from external low-resolution-high-resolution image pairs, while the internal prior is utilized by cluster and low-rank approximation. The experimental results show that the proposed method outperforms many other existing super-resolution methods in recovery effect and robustness.

In the past two decades, laser scanning confocal microscope has been the standard tool for observing the process of life at cellular and sub-cellular level. The optical sectioning capacity of pinhole-based confocal microscope comes at the price of unwanted excitation of fluorophores out of focal plane and phototoxic damage to biological samples. As a new type of fluorescent microscope, light-sheet fluorescence microscope (LSFM) uses side illumination to conduct surface imaging of the samples directly. As compared to the point-scanning imaging mode, LSFM excels at its imaging speed, which is much higher than that of laser scanning confocal microscope, thus making it possible to study some high-speed fine life activities. Another advantage of the light-sheet fluorescence microscope is that only the sample at the light-sheet is excited and the sample outside the light-sheet is not excited, so there is less phototoxic dosage and we can observe the sample in a longer time scale. The special illumination and imaging mode of the light-sheet fluorescence microscope make it play an irreplaceable role in three-dimensional high-speed imaging of big biological samples. The history and research status of light-sheet fluorescence microscope are reviewed with the purpose of providing a personal perspective of current situation and future direction of LSFM.

Cell biology has come into a new era by the hand of super-resolution microscopy. However, how to improve the temporal-spatial resolution of super-resolution microscopy is still an important problem to be resolved in the field of optics. So far, almost all of the super-resolution microscopies are depend on fluorescent probes. As a unique series of fluorescent probes, light regulated fluorescent proteins can be activated by excitation light of different wavelengths to produce stochastic or specially patterned signals. Using this information, the spatial resolution of lens system is improved. By summarizing different parameters of light regulated fluorescent proteins, we focus on fluorescent probes to discuss the way of improving spatial resolution of imaging system. It provides advices to choose proper fluorescent probes, and the relationship between fluorescent proteins and super-resolution microscopy is expounded.

Stimulated emission depletion microscopy (STED) is a powerful far-field technique for super-resolution optical imaging with a few tens even a few of nanometer spatial resolution, thus it is extensively used in investigation of cell biology and so on. The spatial resolution of STED highly depends on the intensity distribution of the doughnut-shaped depletion spot, near the objective focus. The polarization state of field has influence on focal intensity focused with high numerical aperture objective. The off-axis aberrations of microscopic system bring serious damage to the central symmetry of doughnut-shaped depletion spot. The influences of different polarization states of incidence vortex beam and aberrations of coma and astigmatism is of the optical system on the intensity profiles of doughnut-shaped depletion spot simulated by using the vectorial diffraction theory. In the experiment, the deformed depletion spot is optimized by utilizing a pure phase spatial light modulator to correct the aberrations of the optical system. A fluorescent nanoparticle is used as a probe to scan the focal region to obtain a high spatial resolution of intensity distribution. The measured results are in good agreement with those predicted by the vectorial diffraction theory.

Multicolor imaging is an important extension of super-resolution microscopy, which greatly enhances the investigation ability of the relationship between localization and interaction of sub-cellular structures. Therefore, it is beneficial for researchers to go deep into understanding of complicated biological phenomena and processes in cells. Based on the special working principle of single-molecule localization super-resolution microscopy (SMLM), several characteristic multicolor imaging technologies including excitation-dependent, activation-dependent, split-dependent multicolor SMLM and so on are accomplished. In this paper, we analyze the advantages and disadvantages among six main multicolor SMLM from the views of color separation ability, spectrum cross-talk and data collection efficiency. In addition, we discuss the cell fixation method in relation to the multicolor imaging. This review may be helpful for researchers to seek suitable and reliable multicolor imaging methods to study corresponding scientific problem according to their own experimental conditions.

An improved image super-resolution algorithm based on the convolutional neural network is proposed to overcome many problems such as more parameters, a large amount of calculation, longer training time and fuzzy texture combined with the present image classification network model and visual recognition algorithms. The proposed algorithm adjusts the convolution kernel size to reduce parameters in the original three layers of convolutional neural network. Pool layers are added to reduce the dimension and decrease the computational complexity. The learning rate and size of input sub-blocks are improved to reduce the training time. The training database is expanded to provide extensive and comprehensive characteristics. Experimental results show that the proposed algorithm achieves good super-resolution results, and the subjective visual effect and objective evaluation indices are both improved obviously. The image resolution and edge sharpness are enhanced significantly.

Face targets in image taken by the current surveillance video are small and difficult to identify. The image super resolution processing has become the technology and means to solve the image practical application problems of surveillance video. A super resolution technology for outdoor surveillance video face image is proposed. The prior knowledge is used to construct the image training set, and some pre-processing operations like the image space conversion and denoising are operated. The convolutional neural network with eight layers is designed, and its layer types and connection mode are set. Meanwhile, the activation function types and the transmission mode functions among layers are set. The network parameters are initialized and the network is trained according to the training set. The convolution kernels and bias parameters are adjusted reversely by the loss function, and the image output is implemented. Through a large number of actual monitoring video image tests, and compared with other existing methods, the experimental results show that the proposed method has certain advantages in the effect of image super resolution and processing speed.

In order to form an extended light sheet to provide high resolution across a large field of view (FOV), the combination of Airy beam light sheet illumination and deconvolution arithmetic is proposed. The light intensity distributions of Gaussian beam and Airy beam in the focal plane of the objective are numerically simulated. The light sheet microscope setup is constructed. Airy beam is created by modulating the wavefront with a cubic polynomial function, and fluorescent microspheres and stained zebrafish musculature are imaged. Finally, based on the pictures imaged by Airy light sheet microscopy, deconvolution arithmetic is built to improve the axial resolution. For fluorescent microspheres (the amplification is 42), the FOV is extended from 25 μm to 208 μm. For labeled zebrafish musculature (the amplification is 53), the FOV is extended from 20 μm to 167 μm.The experimental result and numerical simulation all suggest that the FOV of the system can be extended with high resolution.

The decline in the quality of retinal imaging will reduce the visual acuity and other visual function, which directly affects people′s daily lives. Human eye aberration is the main factor affecting the image quality. In order to improve the retinal image quality, the adaptive optics is commonly used to correct the aberrations of the human eye. Thus, the resolution or ideal image quality which is close to the diffraction limit is obtained. But the process of correcting the wavefront aberration needs to evaluate the system resolution or the image quality as a basis. Mostly, the evaluation methods can be divided into two types which are based on the pupil plane and imaging plane. Although these two methods can mainly represent the resolving power of imaging system, they cannot directly and quantitatively evaluate the image quality. The optical transfer function (OTF) method based on the imaging plane needs the aid of the curve area to quantitatively evaluate the image quality. And the image quality of different spatial frequencies is different. Thus, the OTF of one eye aberration is calculated, which can simulate the process of human eye imaging. It is used to directly characterize the change of image quality. And the root mean square error and correlation coefficient are defined to analyze the influence of the aberration on them, which can quantitatively reflect the changes of the image quality before and after correcting each aberration.

Aerosol scale height is one of the important quantities, and it indicates the aerosol concentration and the vertical distribution characteristics of extinction coefficient profile. The aerosol extinction coefficient profile is obtained with micro pulse lidar in northwest China, and the atmosphere vertical optical thickness is measured by sun-photometer. We can obtain aerosol scale height by fitting the measured data. Results show that the aerosol scale heights obtained by the two different methods have a good linear correlation, and the correlation coefficient is larger than 0.87. A statistical analysis for aerosol scale height in northwest China is carried out. The variation in aerosol scale height with time, region and climate condition is presented. The research has significance in studying the characteristics of aerosol variation in northwest China and establishing aerosol patterns.

For digital speckle correlation methods, the reliable initial displacement estimation is particularly important to achieve sub-pixel accuracy. Utilizing marker matching technology, a new initial value estimation method is proposed. This method sticks circular markers with high reflection coefficient on the speckles. In order to eliminate the effects of speckles on extracting points, an improved scale invariant feature transform algorithm detecting extreme points within significant edge region is presented, so the redundant feature points are greatly reduced in this way. At last, the homography transformation is executed to complete full field deformation, and the initial value of each pixel in the region of interest is estimated rapidly. Experiments are performed through making speckle plate and the results show that the initial values obtained by the proposed method can reach a fast convergence within 3-4 iterations and consequently get accurate and reliable measurement results.

In order to find the optimum threshold noise ratio (TNR) of the airborne infrared search and track (IRST) system, the operating range model based on the signal to noise ratio (SNR) is adopted to combine with the function relationships of false alarm probability, recognition probability together with SNR and thus the probability model of detection point target of the airborne IRST system is established. The relationship between the recognition probability and the operating range, the target speed and the oblique angle is mainly analyzed. Combining with the infrared radiation characteristics of aircraft, the mathematical detection probability model of airborne IRST system is established. The influence of the change of target azimuth and pitching angle on the operating range is simulated and analyzed under different TNRs. According to operational requirements, it is concluded that the range of the optimum TNR is from 8 to 10. The results provide references for increasing the detection efficiency of airborne IRST system in practical applications.

The reference interference fringe is used to detect the parameters of the holographic grating. The distortion of the interference fringes is caused by the system aberration, and the grating tiling-error is generated. In order to compensate the grating tiling-error, the relationship between the aberration and the distortion of interference fringe is studied. The reference interference fringes are simulated by computer, and the grating tiling-error is per-modulated into the interference fringe image. To verify the effectiveness of the proposed method, the grating splicing experiment is carried out under the condition of system aberration whose peak to peak value is 0.72λ. The maximum variation of the -1st diffraction wave front at the grating seam is 0.36λ. When the prefabricated reference interference fringe images are used as the reference, and the maximum variation is 0.097λ. The experimental results show that the modulated interference fringes can effectively control the tiling-error of the grating in the exposure system with large aberration.

According to the coupled mode theory, a three-core fiber based three-mode multiplexer/demultiplexer supporting to transmit three spatial modes (LP01, LP11a, LP11b) is designed. This three-core fiber consists of a three-mode core located in the center and two single-mode cores deployed in the outer layer. The structural parameters of the three-mode core are chosen, and then according to the effective refractive index matching principle of mode, parameters of two single-mode cores in the outer layer are designed. The power conversion processes between LP11a mode and LP01 mode, as well as LP11b mode and LP01 mode, are analyzed by simulation, respectively. The optimal fiber length obtained is 5.2 mm. Finally, in C+L band, the working bandwidth provided by the three-mode multiplexer/demultiplexer is 50 nm for mode conversion efficiency of 90%. This three-core fiber based three-mode multiplexer/demultiplexer has plenty of advantages, such as simple structure, high mode conversion efficiency, small splice loss, and wide band.

In order to meet the need of slope monitoring, a new fiber Bragg grating (FBG) differential tilt sensor combining mercury column and piston is proposed. When the tilt angle of measured object changes, the pressure to the piston changes as the effective mercury column length changes in the metal shell. The pressure of the mercury is passed to the free end of the cantilever beam by the piston and leads to the central wavelength change of FBG, which is pasted on the left and right ends of the cantilever beam. By use of the finite element method, the stress characteristics of the cantilever beam and FBG are analyzed with two point bonding method at different tilt angles. According to the simulation parameters, the tilt sensor is developed. Angle calibration test and temperature sensitivity test are carried out on the sensor. Experimental results show that the measuring range of the FBG tilt sensor is 0°-40°; sensitivity is 21.8 pm/(°); the fitting degree R2 is 0.998, and the repeatability is 1.12%.

A kind of magnetic-fluid-cladded photonic crystal fiber coupler is designed and investigated experimentally. The coupling ratio of the coupler can be adjusted online by magnetic field. Experimental results indicate that the smaller the coupling area size of the as-fabricated coupler is, the higher the sensitivity of the coupler with respect to magnetic field is. The coupling ratio of the as-fabricated coupler with coupling area size of 27 μm changes linearly with the external magnetic field within the range of 0-14 mT. The achieved sensitivity is 0.0289 mT-1. But the influence of ambient temperature on the coupling ratio will increase when the coupling area size of the as-fabricated coupler decreases. For practical applications, the trade-off between the magnetic tunability of the coupling ratio and the temperature response should be considered comprehensively and the photonic crystal fiber coupler with appropriate parameters should be chosen.

The transient optical signal with wide bandwidth widely exists in laser interference velocity measurement, optical communication and some other fields. It has wide bandwidth and non-repeatability. The existing direct or indirect optical pulse measurement technologies cannot be applied to the measurement of such signal. We use time-wavelength mapping chirped optical pulse to sample the transient optical signal at a high rate under the four wavelength mixing effect, and magnify the signal time scale by using time-stretching technology and dispersive medium. The bandwidth and sampling rate of the back-end electrical record system are increased by times. The experimental results show that the measurement of light beat-frequency signal with bandwidth of 56.978 GHz in real-time is feasible and the equivalent measurement time resolution is 3.7 ps, which verifies the validity of the proposed method.

A portable biosensor with high sensitivity is proposed, and the biosensor is based on a micro/nano-fiber coupler. When the biosensor is used in label-free biodetection, its sensitivity and reusability are studied. A fiber coupler with the waist diameter of 3 μm is fabricated by the fused biconical taper method, and then the coupler is used in refractive index detection experiment. A refractive index sensitivity of 1402.3 nm/RIU is measured in experiment. The detection results are fitted, and the correlation coefficient of the fitted curve is 0.99459.The calculated results of the weak coupling model of the micro/nano-fiber coupler are in good agreement with the experimental results. The detection limit of 2 pg/mL is obtained when the micro/nano-fiber coupler is used to detect goat immune globulin (IgG) antigen. The reusability of the biosensor is verified by 10 dissociation experiments. These experiments demonstrate that the micro/nano-fiber coupler can obtain a high detection sensitivity in the label-free biosensing and has good practical value.

High precision optical frequency transfer in fiber link is of great significance to optical clock comparison. Bidirectional Er-doped fiber amplifier (EDFA) contributes to loss compensation and high precision transmission of signal in long distance optical frequency transfer. Based on the basic principle of stimulated amplification for erbium particles, a bidirectional EDFA with low noise and high gain which can be used in fiber optical frequency transfer link is designed, and its parameters are simulated and optimized. Experimental results show that the noise index and the gain of the bidirectional EDFA is 3.86 dB and 20.14 dB, respectively. The phase noise introduced is only 0.1 rad2/Hz at the frequency of 1 Hz. The bidirectional EDFA is used as an amplified compensation device in 200 km fiber optical frequency transfer link, and a second frequency stability of 3.8×10-16/s and a ten-thousand second frequency stability of 2.8×10-19/(104 s) are obtained. It has broad application prospects in the field of frequency signal transfer and optical clock comparison.

In order to realize high resolution angular measurement, a subdivision method based on the photographic sensor is proposed. A more adaptable high-resolution subdivision algorithm is proposed to eliminate the subdivision error caused by the irregular distance between the photographic sensor and the grating center during installation and adjustment of the encoder. The mathematical model of the proposed subdivision algorithm is established, and the error of the method is analyzed. Based on practical working conditions, an error model which takes many aspects into consideration is built and the effect of various elements on the subdivision algorithm is analyzed. According to the error analysis results, how to reduce the subdivision error of photographic encoders is recommended. Experimental results show that the subdivision error can be ignored when the line number of circular grating is more than or equal to 128. The results can provide theoretical foundation for develop the miniature photographic encoder.

In order to avoid the error introduced by the imaging system in traditional pinhole type point diffraction interferometer, a lensless imaging algorithm is proposed. The algorithm can retrieve the phase of the interferograms captured by CCD without the imaging system. On the basis of the plane wave spectrum theory, we can obtain the interferogram of the exit pupil conjugate location by constructing a virtual imaging conjugate relation. Simulation and experimental results show that the proposed lensless imaging algorithm can remove the edge diffraction fringes and the central diffraction effects in the interferogram, and the coherent noise can be eliminated as well. With this lensless imaging method, we can remove the interference imaging system in the pinhole type point diffraction interferometer which is difficult to design and fabricate. Besides, this algorithm can also effectively guarantee the high accuracy of pinhole type point diffraction interferometer.

In order to guarantee the polarization accuracy of atmospheric synchronous corrector, it is necessary to test and filtrate its filters in the process of corrector development. The polarization measurement model of corrector and the filtration conditions of optical filters are designed, and the main factors which affect the polarization accuracy of corrector are analyzed. Taking the 490 nm channel as an example, using the spectral transmittance of optical filters and the spectral radiance data of typical targets, the relative transmittance between polarized channels and the changes of ratio of out-of-band response to total response with the spectral radiance shape of typical targets are analyzed, and the filtration of optical filters is performed. At last, the polarization measurement precision of the corrector is verified. The results show that the polarization accuracy of the corrector can be guaranteed when the changes of the relative transmittance between polarized channels and the changes of out-of-band response with total response are all not more than 0.4%, and the combined uncertainty is less than 0.5% in the range of atmospheric polarization 0 to 0.4. The uncertainty of a corrector with optimal selection filter is obviously smaller than that of a corrector without optimal selection filter, which proves the effectiveness of the filtration method.

Pulse wave measurement of carotid artery is the foundation of the prophylaxis and treatment of cardiovascular diseases. Three dimensional digital image correlation system is applied to measuring the human carotid artery based on the water transfer printing (WTP) technique, which can transfer optimized digital speckle patterns for different skin colors. Compared with natural textures of the skin, digital speckle patterns fabricated from the WTP technique are proved to be more stable, accurate and efficient. Experiments are performed to demonstrate the validity of the proposed method. Various information of the pulse wave such as shapes, frequencies and deformations in both time domain and spatial domain can be extracted from the full-field of carotid arterial surface. The research may be beneficial to cardiovascular disease monitoring.

Utilizing two section passive mode-locked laser structure, a monolithic high-power quantum well passive mode-locked laser operating in the near-infrared band is realized by cascading two quantum well active regions by the tunnel junction. The mode-locked characteristics of the model-locked laser are measured. The results show that the center wavelength of this laser is 1038 nm, the pulse repetition frequency is 24.37 GHz, and when the working condition of lasers changes, the pulse width varies from 2.19 ps to 9.27 ps. The peak power variation range is from 76 mW to 308 mW. It shows that the mode-locked laser has the advantages of monolithic integration, compact size, high power and high repetition frequency. Meanwhile, under certain reverse bias conditions, the mode-locked laser exhibits the characteristics of power bistability and enlarges the effective mode-locking range.

As for the discharge uniformity and impedance matching problems during multi-group electrode discharge at 3 kW high-power radio frequency (RF) slab CO2 lasers, the multi-group electrode discharge model is built with the Maxwell time domain difference method and the shunt inductor value is gained under the optimum discharge uniformity efficiency. The impedance of total load in the discharge electrode is calculated with the combination of the shunt inductor value and the equivalent impedance of discharge area. The impedance matching network parameters are given with the Smith circle and the impedance of total load, and the sum of impedance of total load and matched network impedance is 50 Ω which makes the RF input power completely fed. The actual measured standing wave ratio (SWR) is 1.18 when the discharge uniformity and impedance matching are satisfied simultaneously. The results show that, with the matched shunt inductors added in the time domain difference method, the fluctuation of the overall discharge uniformity of electrode slab decreases from previous 15% to 2.8% and the discharge uniformity difference between each electrode group is within 0.3%. The discharge experiment shows that, under the discharge uniformity and impedance matching conditions, the laser output power of 3.1 kW within 2 h can be stable with the fluctuation within ±1%.

Due to the large distortion of fisheye camera images, the two-dimensional calibration board approaches are often difficult to obtain reliable corners around the board boundary, which may lower the calibration accuracy. Traditional three-dimensional calibration methods have some challenges on building a complex calibration field to obtain sufficient number of feature points. Therefore, a three-dimensional fisheye camera calibration method based on the laser scanner is proposed to achieve more accurate results. First, a laser scanner is used to obtain the point clouds of the fixed indoor environment. Second, an improved scale-invariant feature transform matching method is used to detect the corresponding features between point clouds and camera images. A two-round random sample consensus (RANSAC) algorithm is applied to rejecting outliers among the correspondences, and the internal and external parameters of fisheye camera are estimated based on the image and physical coordinates remaining corresponding points after the 3D RANSAC screening. Compared with the classical Zhang Zhengyou calibration method and its improved methods for various cameras, the proposed method can get more effective feature points to make calibration accuracy much higher, which can provide good lens distortion correction for fisheye camera. The proposed method is convenient, precise and extensively applicable.

Focusing on the issue that spatially regularized discriminative correlation filter (SRDCF) tracking algorithm has poor performance in handling rotation, out of view and heavy occlusion, we propose a visual tracking approach based on adaptive convolutional features. First, based on the principal component analysis of conv3-4 layer features in the VGG-NET model, the dimension of conv3-4 layer features is reduced from 256 to 130 by adaptive dimension reduction technique. Then, we maximize classifier score in the detection area and get the location and scale of target. In order to redetect the target in the case of tracking failure and achieve long-term tracking, we compare the confidence of the location with maximum score and train an online support vector machine (SVM) classifier. Finally, the tracking model is updated by the reliable tracking results which are determined by peak-to-sidelobe ratio. To verify the feasibility of the proposed algorithm, the results are compared with those obtained by thirty-eight kinds of tracking algorithms in one hundred video sequences of OTB-2015 benchmark. Experimental results indicate that the precision and success rate are respectively 0.804 and 0.607. The proposed approach has a ranking of one. Compared with SRDCF tracking algorithm, the proposed approach improves the precision and the success rate by 1.9% and 1.5%, respectively. In addition, the proposed approach is robust for rotation, out of view, heavy occlusion and other complex scenes.

To solve the problem of image matching failure under complex conditions including extreme non-rigid transformation, partial occlusion, and imbalanced illumination, a more robust template matching algorithm than the original best-buddies similarity (BBS) algorithm is presented. The similarity measurement between image patches is represented by Manhattan distance instead of Euclidean distance. On this basis, the new confidence map is constructed by sliding window to compute the BBS response values. The center of the brightest connected region is determined to be the last matching location after threshold filtering the confidence map by filtering process to eliminate the minor response values. Experimental and analysis results show that the proposed algorithm may be used to match the images with elastic deformation, similar region interference, partial occlusion, and extreme illumination change, etc.

Aiming at defects of the poor uniformity of spotlight distribution and the mismatching between the focusing spot shape and solar cells for the traditional point focusing Fresnel lens, a Fresnel lens is proposed to uniformly focus square spot by sub-region multi-focus superposition. Four angle-missing isosceles right triangles were intercepted from the traditional circular concentric rings spot focusing Fresnel lens and spliced seamlessly to form a Fresnel lens uniformly focusing square spot by sub-region four-focus superposition, and the concentration uniformity is improved effectively by light superposition in the four parts. The ray tracing method was used to simulate and analyze the influence on the optical performance parameters such as the shape of the focal spot, the concentration uniformity and the irradiance, which is caused by the structure parameters of lens such as ring spacing, angle-missing chord length and waist length. The results show that the focus spot shape of the lens designed by this method is square, and the concentration uniformity is higher than 90%.

This paper introduces the basic structure of a long depth of field array optical components defect detection system used in SG-Ⅲ. The phase modulation used on the system aperture can realize the depth of field extension effect, which can achieve large depth of field imaging detection from 3~13 m object distance. The performance index, design ideas, design result and simulation evaluation of the optical detection system are expounded. It can achieve the defect simultaneous detection of the array optical components in longitudinal range of 3~13 m by using proposed system.

The astigmatism Z5 in the etching process of lithographic lens silicon wafer can greatly deteriorate wave aberration of the lithographic lens. In order to compensate the astigmatism in real time, this paper proposes an astigmatism active Z5 compensation system that includes real time data platform, actuator system, flexible supporting structure and optical lens. The sphere interferometer is used as test equipment for optical lens surface shape. The least square method and the linear superposition principle are both used to determine the relationship between actuate parameter and surface shape. The experiments, including test of response function of actuator in active compensation system, test of compensation stroke, test of compensation precision, test of compensation resolution, are conducted. The results show that the astigmatism Z5 compensation stroke of the system is up to 735 nm. The astigmatism Z5 compensation precision is less than 2 nm with the high-order aberration less than 1 nm. The astigmatism Z5 compensation resolution is 2 nm. The system can effectively compensate the wave aberration in the lithographic lens system and satisfy the image quality requirement of lithographic lens.

An all-optical diode structure based on Fano micro-cavity at the side of 2D photonic crystal waveguide is designed to realize the nonreciprocal transmission of light. The key technique is to break the spatial symmetry of Fano micro-cavity structure by using a simple reflecting layer in the photonic crystal (PC) waveguide, so that the coupling efficiency of waveguide on both sides of the reflective layer and micro-cavity is asymmetric. Since the thresholds of incident light intensity required for the nonlinear Kerr material to excite the micro-cavity on both sides of the waveguide are different, the function of one-way conduction is achieved. The transmission properties and performance are numerically simulated and investigated by finite-difference time-domain (FDTD) method. The results show that forward transmission and backward cut-off all-optical diode effect can be achieved in this structure with low light intensity threshold. The structure has ultra-fast response time, which can reach the order of picoseconds. It also has high maximum transmittance (about 90%), and high positive and negative transmittance. The design based on photonic crystal structure enables the device to have good work wavelength tunable characteristics, which makes it easy to produce and integrate with other devices on the basis of current semiconductor technology.

In order to realize the measurement with high speed, high accuracy and high sensitivity of crystal electro-optic coefficients, a novel electro-optic coefficient measurement method based on photo-elastic modulation is proposed. By means of low-pass filtering and digital phase-locked technologies, the data of the DC term and the first harmonic term of the modulated signals are extracted, respectively, then the crystal electro-optic coefficient is demodulated. The principle of this new scheme is analyzed in detail, the experimental system is set up and the electro-optic coefficient of lithium niobate (LiNbO3) crystal sample is measured. The experimental results show that the measurement accuracy is 2.3%, the sensitivity is 1.7×10－14 m/V, and the measurement speed is 100 ms per data point.

Under short filament condition, the plasma with locally high density can be generated by means of the interaction between laser and medium from jet target, the acute change of dual-color-field phase difference is realized in a short distance, and the elliptically polarized terahertz (THz) radiation is easily obtained. By adjusting the back pressure of gas, THz radiation with ellipticity of 0.438 is generated. The simulated result based on the transient photocurrent model is consistent with that from the experiment, which reveals the mechanism of combined action of photocurrent and plasma dispersion.

In-flight absolute radiometric calibration is one of critical techniques for quantified information application of remote sensing. A novel calibration approach is presented based on reflected point sources. At the same time, the reflectance measurement on site is changed to accurate determination in laboratory. The top-of-atmosphere radiance of satellite-borne remote sensors can be calculated by simplified radiometric transfer code and measurement of the atmospheric optical properties. The calibration method can also isolate the response value of remote sensing image created by the targets from the response produced by sky path radiance and background radiance through in-flight point spread function estimation by point source method, which can eliminate the assumption of aerosol scattering properties. The approach can break the limit of time, locale and weather conditions at vicarious calibration sites and achieve high precision, high frequency mobile calibration for high spatial resolution satellite sensor in complex environments. Preliminary results show that its uncertainty is less than 3%. The difference of calibration coefficient is 3.02% between the proposed method and the method based on large-area gray-scale target. Simultaneously, the method can simplify the process and improve the calibration accuracy and efficiency.

Aiming at solving the problems of remote sensing images with poor clarity, contrast and color fidelity in foggy weathers, a rapid defogging method using dark channel prior is proposed based on the characteristics of large amount of data, small changes in depth and almost no area of the sky. To ensure the performance of defogging, the dark channel image is obtained by calculating the minimum value of r, g, b channels. It significantly reduces the complexity and avoids heavy computation. Experimental results demonstrate that the improved algorithm can remove the haze effectively. It can enhance the definition and color of hazy degraded image, and the processing time is 2% of the original method. The proposed algorithm can satisfy the requirement of real-time remote sensing image processing.

In order to obtain the geographic location of photography areas accurately, the geo-location algorithm is developed for time delay integral-charge coupled device(TDI-CCD) aerial panoramic camera without the laser range finder. Using aircraft position and attitude information measured by airborne position and orientation system (POS) along with the depression angle and position angle from the encoder in aerial camera, line of sight (LOS) pointing angle in geographic coordinate is solved by homogeneous coordinate transformation. According to the WGS-84 ellipsoidal earth model, longitude and latitude location is solved with the earth ellipsoid calculation theory. The influence of aircraft attitude angle measurement error and the depression angle and position angle measurement error on line of sight pointing angle calculation accuracy is analyzed with the Monte Carlo method. The impact of photography inclination angle and topographic change of target region on geo-location precision is investigated, which shows that the precision of geo-location is higher when the photography inclination angle and topographic change of target region are smaller. The validity of the geo-location algorithm is verified by the flight test in which the plane flies at an altitude of 17750 m and the photography inclination angle ranges from 63° to 75°. The circular error probability of target geo-location is less than 212.96 m, which meets the requirement of the project.

Based on the vector-ray tracing (VRT) model, the optical caustic structures of homogeneous ellipsoid droplets in the third-order and fourth-order rainbow regions are studied. The evolution process of the optical caustic structure is investigated in detail. The simulated result and fringe curvature of the third-order rainbow fringe as well as the relationship between the cusp location of the third-order rainbow and hyperbolic umbilic fringes and the aspect ratio are obtained.

Aluminum oxide (Al2O3) thin films are prepared by radio frequency magnetron sputtering at room temperature, and the optimal control of the properties of these films is realized by adjusting the sputtering pressure. The prepared films with the best thickness uniformity are obtained under the sputtering power of 120 W and Ar pressure of 0.13 Pa, whose ratio of aluminum atoms to oxygen atoms is 1∶1.67, density is 3.21 g/cm3, and surface roughness is 0.62 nm. This smooth and compact Al2O3 thin film can significantly reduce the defects, and it also possesses the characteristics of high breakdown voltage, high relative dielectric constant and low leakage current. The flexible α-IGZO-TFTs are prepared on a polyimide (PI) substrate at room temperature by utilizing the optimized Al2O3 films as the gate insulation layer, and exhibit electrical performance with a mobility of 2.19 cm2/(V·s), an on/off current ratio of 105, a sub-threshold swing of 0.366 V/decade, and a threshold voltage of 3.01 V.

The experiments of second harmonic generation (SHG) of 57.4 fs pulses at 1550 nm wavelength for type Ⅰ(o+o-e) and type 0 (e+e-e) are achieved using Mg-doped lithium niobate crystal with mole fraction of 5% of magnesium. For SHG of type Ⅰ, the second harmonic pulse with spectrum width of 28 nm and pulse width of 79 fs is obtained at the peak density of 4.3 GW/cm2, and the conversion efficiency is up to 54%. For SHG of type 0, the second harmonic pulse with spectrum width of 2.1 nm is obtained at the peak density of 3.7 GW/cm2, and the conversion efficiency is up to 40%. The key points of the evolution for fundamental pulse and second harmonic pulse in SHG process are analyzed, which are the phase-match of femtosecond pulse multi-wavelength components of fundamental wave in frequency domain and the group-velocity-match of fundamental pulse and second harmonic pulse in time domain. It is found that the spectral width can keep constant in the propagation when the phase-match of multi-wavelength components is satisfied. However, the spectral width will gradually become narrower with the increase of the propagation distance when the phase-match of the center wavelength is satisfied only.