In order to accurately measure the concentration of NO2 gas by mid-infrared differential absorption lidar, the absorption spectrum characteristics of NO2 in mid-infrared band were measured and analyzed. The absorption cell experiments were conducted using two wavelengths λon and λoff emitted from OPA lasers and OPO lasers respectively. The absorption cross section of NO2 gas in the range of 3 410~3 433 nm is measured by a laser λon whose spectral line-width is less than 0.05 nm. The absorption cross sections at 291 K, 308 K and 363 K were acquired respectively. And the absorption spectra from 3 400~3 435 nm were also acquired by a laser λoff with spectral line-width being 10 nm. The experimental results show that the absorption cross-section is negatively correlated with temperature in the range of 3 410~3 433 nm, consisted with HITRAN database and the correlation coefficient between the measured data and the HITRAN data is above 0.92. A new convolution correction method is used to calculate the absorption spectrum line under laser λoff. The correlation coefficient between the measured results and the fitting results is 0.97. The measured absorption cross sections at the on and off wavelengths are applied to the simulation of the mid-infrared differential absorption lidar, the concentration error of differential absorption lidar is simulated, which verified the feasibility of the scheme of differential absorption lidar.

In order to improve the detection ability of the terminal sensitive projectile to the armored target under complex background conditions, the linear array LiDAR is used as the detector and the convolutional neural network is combined to classify and identify the range profile of the linear array LiDAR. The range imaging of the scanning area is realized by using the steady-state motion characteristics of the terminal sensitive projectile dropping while rotating. The original range profile is constructed into a range profile suitable for convolutional neural network by sampling rate control and interpolation. The convolutional neural network consisting of two convolutional layers and one full link layer are established to meet the requirements of high real-time, small size and low power consumption on missile-borne. Xilinx ZYNQSoC chip is selected as hardware platform, and hardware acceleration is achieved by placing convolution operation on PL end of ZYNQSoC based on HLS technology and SDSoC development environment. The scaled simulation experiment proves that the method has high target recognition accuracy and can effectively recognize armored targets in complex background. Through placing convolution operation on PL end of ZYNQSoC, the performance acceleration is five times faster than that of CPU, which can meet the requirements of missile-borne.

A CMOS single photon avalanche diode based time-of-flight sensor is presented for light detection and ranging applications. The sensor integrates 16 structure-optimized single photon avalanche diode pixels and a dual-counter-based 13-bit time-to-digital converter. Each pixel unit has a novel active quench and recharge circuit. The dark noise of single photon avalanche diode is reduced by optimizing the guard ring of the device. The active quench and recharge circuit with a feedback loop is proposed to reduce the dead time. A dual-counter-based time-to-digital converter is designed to prevent counting errors caused by the metastability of the counter in the time-to-digital converter. The sensor is fabricated in 180 nm CMOS standard technology. The measurement results show the median dark count rate of the single photon avalanche diode is 8 kHz at 1 V excess voltage, the highest photon detection efficiency is 18 % at 550 nm light wavelength. The novel active quench circuit effectively reduces the dead time down to 8 ns. The time-to-digital converter with 416 ps resolution makes the system achieve the centimeter-accuracy detection. A 320×160 depth image is captured at a distance of 0.5 m. The maximum depth measurement nonlinear error is 1.9% and the worst-case precision is 3.8%.

A single-photon avalanche diode with double charge layers is designed with 180 nm standard CMOS technology, which is able to improve photon detection efficiency. PN junction is formed by deep N-well with retrograde doping P-charge layer. The different doping concentration of P-charge layers are selected to optimize the breakdown voltage. When the P-charge layer concentration is 1×1018cm－3, the simulation results show that the breakdown voltage is 17.8 V and the electric field intensity is 5.26×105V/cm. Further study shows that the position of N-charge layer affects drift current density and diffusion current density. The device performance is optimal when the N-charge layer is doped at the contact of the deep N-well and the N isolation layer, in other words, when the peak of the N-charge layer is 2.5 μm from the surface of the device. By using Silvaco TCAD simulation analysis, which can get a conclusion that at a wavelength of 500 nm, the detection efficiency peak is 62% under bias voltage of 1 V, while the photon detection efficiency in the range of 300 nm to 700 nm is greater than 30%.

A method for detecting small and weak damaged targets based on Neighborhood Vector PCA (NVPCA) image enhancement was proposed. The main idea is that each pixel and its 8 neighborhood pixels in the damaged image are treated as a column vector to participate in the operation. All column vectors generated by each pixel will construct a 9-dimensional data cube. After PCA transformation, the correlation between the middle and neighborhood pixels is eliminated, so that the main information of the 9-dimensional data cube will be set in the first dimension, and the transformed first dimension data is NVPCA image. When the NVPCA image is processed again by LCM method, a better image enhancement effect is obtained. In addition, the region growth method is used to separate the damage target from background. The experimental results show that the method can detect the damage target with the size of 1 pixel and located in local bright area, and meet the requirement of on-line optical damage detection system for the accuracy of the damage target.

In this paper, two chaotic systems with common parameters were used to generate new templates that could replace the random phase templates of the traditional double random phase encoding system. The plaintext image was encoded as phase information to overcame the insensitivity of the first random phase template of the original traditional double random phase encoding system. A generalized Fibonacci chaotic system, which could generate uniform and unrelated random sequences, was constructed to generate evenly distributed phase templates for image encryption and improve the efficiency of key transmission and the sensitivity of the keys. The complex-valued image obtained by the first encryption was replaced by extracting its original amplitude and phase for re-encryption. Therefore the problem that the pixel values of the complex-valued image could not be substituted by bitwise-operation of "xor"was solved, and the pixel value distribution of the ciphertext image became more uniform. The entropy of information reached 7.995 8, which could effectively resist statistical analysis attacks. In the second encryption, the chaotic initial value of the encryption template was connected with the first encryption ciphertext. The number of pixels change rate of 0.995 239 was closer to the ideal expected value. The connection enhanced the sensitivity of the plaintext and the ability of the resistance to choice plaintext attack. Simulation results show that this method can effectively increase the key space and the sensitivity of the keys. Moreover the encryption efficiency and the security of the encryption system are improved effectively.

A multi-focus image fusion method based on super-resolution combined with group sparse representation model is proposed. First, the bicubic interpolation method is used to enhance the resolution of the source image and the source multi-focus image information. Then, the adaptive sparse representation learning dictionary is used to learn the image blocks without obvious dominant direction and specific dominant direction respectively. The sparse coefficient representation of the source multi-focus image is conducted by the group sparse representation model. Finally, the maximum l1 norm is used to select the final representation coefficient vector. The experimental results show that the proposed method restrains the shortcomings of low spatial resolution and blurring that are easy to appear in multi-focus image fusion, and has better contrast and sharpness. Subjective visual effects and objective indicators show that the proposed method has certain advantages over traditional multi-focus image fusion methods, especially in the mutual information index of the three sets of image fusion results leading 0.37, 0.38 and 0.32 respectively.

A method of adaptive threshold segmentation, detection and tracking for infrared sequence images is proposed to solve the problems that the current infrared target detection and tracking algorithm has weak scene adaptability, strong specificity and high false detection rate of small target in the first frame image under large field of view. The density, rectangularity and Hu invariant moments are selected as static variables to establish static decision criteria. The target moving speed, area and perimeter of target contour, adaptive segmentation threshold and location of ROI are selected as dynamic variables to establish dynamic decision criteria. The first frame target detection algorithm is used to calculate the target static variable and some of the dynamic features of the first frame image. The subsequent frame images are segmented by the improved local adaptive threshold segmentation algorithm and then the static and dynamic decision criteria are used to screen out the segmentation. Finally, the dynamic decision criteria are calculated and updated. The infrared target test results show that the method has good adaptability to different scenarios. By using this algorithm, the average tracking accuracy of the four scenarios is 95.81%, the average processing time per frame is 10.93 ms on microcomputer platform and 26.79 ms on embedded platform respectively.

Aiming at the uneven distribution of haze concentration and color imbalance in haze weather images, a block-based image enhancement algorithm is proposed, which combines dark channel prior algorithm with multi-scale retinex with color restore algorithm. The idea of the algorithm is to map the atmospheric scattering model linearly to the scale of the Retinex image model, and get a new multi-parameter model with both transmittance and illumination value in the atmospheric scattering model and the incident image in the Retinex image model. The new original image of removing atmospheric scattering image is obtained according to the model, and the incident image of the whole image is calculated by using different scales of guided filtering. Then, the final high-frequency detail image is obtained by combining the atmospheric scattering image and color restoration factor. The whole image is divided into several regions in view of the uneven distribution of haze concentration in the image and the dynamic truncation value of each regional block is calculated by using fusion algorithm. According to different dynamic truncation values, the high frequency details of the whole image can be dynamically adjusted to obtain multiple local optimal images. These images are equal-weighted at the pixel level, and finally the optimal image can be obtained which can retain the local details. The proposed method is experimentally compared with other existing methods on subjective visual effect and objective evaluation, the results show that the method can effectively solve the problems of non-uniform fog and color imbalance in images, and obviously improve the quality of the image after fog removal.

In order to improve the fusion efficiency and solve the problem of wrong selection of fusion coefficients in the fusion method based on multi-scale transform, a method based on non-subsampled dual-tree complex contourlet transform combined with adaptive block is proposed for infrared and visible image fusion. Firstly, the source image is decomposed into low-frequency and high-frequency parts by non-subsampled dual-tree complex contourlet transform. For the low frequency coefficients, an adaptive block-based fusion technique is applied, where the optimal block size can be calculated by using the improved drosophila algorithm, and the low-frequency fusion results are refined to obtain a label map which can accurately indicate the origin information of each pixel. Then, the neighborhood coefficient difference of the high-frequency component is used to combine the label map height. Finally, the fused image is reconstructed. The experimental results demonstrate that the proposed method can accelerate the process of fusion and solve the problem of block effect in spatial block fusion.

Aiming at the low accuracy of traditional brain tumor detection, a three-dimensional brain tumor detection method based on deep learning was proposed. Firstly, the magnetic resonance images of different modal brain tumors were fused to obtain the three-dimensional features of brain tumor focus under different modalities. Then, an instance normalization layer was added between the convolution layer and the pooling layer to improve the convergence speed of the network and relieve the problem of overfitting. And the loss function was improved, the weighted loss function was used to enhance the feature learning of the focus area. Finally, the problem of more focuses in the false positive brain tumor was solved combining with the post-processing method. The experimental results show that the proposed brain tumor detection method can effectively detect the tumor focuses. The Dice coefficient, sensitivity and specificity of the three evaluation indexes reach 0.926 7, 0.928 1 and 0.997 7 respectively. The three indicators improve 4.6%, 3.96% and 0.04% compared with the 2D detection network, and improve 13.2%, 10.42% and 0.12% compared with the initial single modal brain.

Based on the surface-enhanced Raman scattering enhancement effect of SiC@Ag substrate and Ag nanoparticles, an ultrasensitive miRNA-106a detection protocol was proposed by using the second surface-enhanced Raman scattering amplification of the silver-biotin-streptavidin nanoaggregates. First, SiC@Ag@anti-digoxin/digoxin-DNA substrate was prepared by specifically binding of a capture DNA modified with a digoxigen group and an anti-digoxin immobilized on a SiC@Ag substrate. The Ag@4MBA@DNA-biotin probe was prepared by linking the amino/biotin-modified probe DNA with the 4-mercaptobenzoic acid (4MBA) immobilized silver nanoparticles. Then, the "sandwich" structure was constructed by Ag@4MBA@DNA-biotin probe, target miRNA-106a and SiC@Ag@anti-digoxin/digoxin-DNA substrate to amplify the surface-enhanced Raman scattering signal. Finally, the streptavidin and the extra probes were alternately added into the “sandwich structure” to form the silver-biotin-streptavidin nanoaggregates and achieve secondary amplification of the detection signal. The experimental results show that the ultrasensitive detection of miRNA-106a has been implemented by using the dual surface-enhanced Raman scattering amplifications of SiC@Ag substrate and silver-biotin-streptavidin nanoaggregates, arriving an ultralow detection limit of 0.579 fmol/L, which has a potential for early diagnosis of tumors.

An Ag-coated polystyrene pellet hexagonal close packed array (Ag/PS HCA) substrate was prepared by L-B film method and magnetron sputtering technique. And Au@4MBA probes were obtained by linking the synthesized urchin-like Au nanoparticles with 4-mercaptobenzoic acid (4MBA). Then the single-stranded oligonucleotide DNA21 was linked with the substrate and the Au@4MBA probes, respectively, to construct the Au@4MBA-DNA21-Ag/PS HCA sandwich structure. After DNA21 and miRNA-21 hybridization, the DNA phosphodiester bond is cleaved using a double-strand specific cleavage enzyme (DSN), and finally SERS signal detection was performed. The experimental results show that the detection scheme of the tumor marker miRNA-21 based on the above-mentioned sandwich SERS structure and enzymatic cleavage technology has extremely high sensitivity and excellent specificity, and the detection limit reaches 0.853 fmol·L－1. In addition, compared with real-time fluorescence quantitative polynucleotide chain reaction (RT-qPCR), our scheme not only has the same results also has higher sensitivity.

In order to further grasp the subsurface damage mechanism of ultrasonic vibration-assisted grinding of optical glass materials, the Vickers indentation experiment under non-ultrasonic vibration and ultrasonic vibration conditions was designed to investigate the indentation features of K9 optical glass under two conditions. The magnetic component fluid polishing was used to detect the depth of the median crack in the indentation area of K9 optical glass. The conventional model of median crack of indentation was modified twice to obtain the model of the Vickers median crack depth of the indentation under ultrasonic vibration conditions. The static and dynamic fracture toughness were calculated by experimental data of Vickers indentation under non-ultrasonic vibration and ultrasonic vibration conditions. The first correction coefficients of the two conditions were 0.08 and 0.06 respectively. Combined with the experimental results of the measured median crack depth, the values of the second correction coefficients under the two conditions were close under two conditions, which were 94.75 and 94.50 respectively. The results show that the new median crack depth model has a good recognition of ultrasonic vibration and processing conditions.

Aiming at the low diffraction efficiency of optical beam splitting devices, based on the traditional Gerchberg-Saxton (GS) algorithm, the initial phase and the amplitude limitation method is improved. The secondary phase is used as the initial phase of the iterative algorithm, and then the output plane is divided into the signal area and the noise area. Keeping the phase of the two parts unchanged, the amplitude in the region is multiplied by a factor which varies with the number of iterations, and the amplitude in the noise region remains unchanged. A 9×9 continuous-face beam splitting device designed by the proposed method is compared with the beam splitter designed by the traditional GS algorithm. The results show that the phase of the beam splitter designed by GS algorithm has serious mutation and discontinuity. However, the beam splitter phase designed by the proposed method is continuously smooth and can be processed by the moving mask technique.The 1×3 and 1×9 beam splitters are prepared, the measured diffraction efficiency are 83.5% and 89.4%, the uniformity error are 3.56% and 15.23%, respectively.

A large field of view Unmanned Aerial Vehicle (UAV) camera imaging system, named curved compound eye camera for the small UAV, was designed. The system consists of three subsystems, a cured microlens arrays, an optical transformation subsystem, and a data processing unit with image sensors. The designed compound camera has a focal length of 4 mm, a F number of 4, and a field of view is 106°, which makes it can resolve the ground target with a feature size of 0.5 m at an altitude of 500 m. In the design, lenslets with a doublet form were used in curved compound eye to eliminate the optical abberations. Since there is an overlap in field of view for neighboring lenslets, lenslets as much as of 7 can view the same target at the same time from different view angles, which allows the object location and speed measurement. The simulation results show that the image quality of the entire compound eye camera system meets the requirements with an acceptable tolerance, and the maximum optical distortion can be controlled under 1.2%.

To achieve efficient AlGaN-based Deep Ultraviolet Light-Emitting Diode (DUV LED), the N-polar LED structure with grading quantum barriers is proposed to manipulate the carrier transport. By adopting the N-polar structure, the hole injection and the electron overflow issues can be improved due to the reversed polarization-induced potential barrier for carrier transport in p-type electron blocking layer. Furthermore, the impacts of different grading directions and schemes on the device performance are investigated. Simulation results show that grading the Al composition linearly from 0.65 to 0.6 for the 12 nm-thick AlGaN quantum barriers along the (000-1) can well balance the quantum barrier height and slope, thus resulting in remarkable improvement of hole injection as well as 53.6% enhancement of optical output power. The proposed design provides a straightforward and effective solution to the electron overflow and hole injection issues, which shows promise in the pursuit of higher efficiency DUV LED.

Based on the small-size streak tube, the dependence of the physical temporal resolution and temporal distortion on the factors including initial energy of electrons generated from the photocathode and spherical photocathode curvature radius is numerically simulated. The simulation results show that the physical temporal resolution is greatly affected by the initial energy, which is less affected by the spherical curvature radius and off-axis distance; the temporal distortion increases with the off-axis distance. Furthermore, the streak tube with a flat photocathode has a positive temporal distortion. With the decrease of the radius of curvature of the photocathode, the temporal distortion gradually changes from positive value to negative ones. And, when the photocathode curvature radius is 70 mm, the temporal distortion can be reduced to 8 ps and the spatial resolution can reach 25 lp/mm @ MTF=10%, the curvature of the slit image can be almost neglected even though the off-axis distance is 8 mm. Besides, the static spatial resolution and luminance gain of the streak tube, the luminance photocathode sensitivity of the photocathode are tested experimentally. The experiment results show that the spatial resolution of the streak tube is higher than 28 lp/mm at the photocathode center while it is higher than 18 lp/mm at the edge. The luminance photocathode sensitivity of the small-size streak tube is 178 μA/lm and the luminance gain is higher than 12 while the luminance gain of picosecond streak tube is only 0.5. Thus, the small-size streak tube has a good weak light detection ability in the field of streak tube imaging lidar.

Two inverted non-lighting devices with the configurations of ITO/TPBi(40 nm)/C60(x nm)/ZnPc(x nm)/NPB(40 nm)/Al(120 nm)(x=0, 5, 10, and 15) and ITO/TPBi(40 nm)/LiF(y nm)/Al(2 nm)/C60(5 nm)/ZnPc(5 nm)/MoO3(3 nm)/NPB(40 nm)/Al(120 nm)(y=0, 0.5, 1.0, and 1.5) were fabricated . The experiments illustrate that the organic heterojunction of C60/ZnPc can realize charge separation under external field and demonstrate that ability of charge separation and injection can be efficiently improved by the electron injection layer of LiF/Al and hole injection layer of MoO3. C60/ZnPc was used as charge generation layer (CGL), LiF/Al and MoO3 were employed as charge injection layers. Moreover, a series of green phosphorescent Tandem Organic Light-Emitting Diodes (TOLEDs) based on the multilayer structure of LiF/Al/C60/ZnPc/MoO3 was also fabricated to further explore the effect of the CGL on photoelectronic performances for tandem devices. Experimental results show that the performance of tandem devices can be influenced by the effect of charge separation and injection of CGL on charge-injection balance inside TOLEDs. When the multilayer structure is LiF(0.5 nm)/Al(1 nm)/C60(5 nm)/ZnPc(5 nm)/MoO3(3 nm), a green TOLED with a better charge balance and an optimal device performance is achieved, the operating voltage of the tandem device is below two-fold of correspording single OLED. The peak luminance, peak current efficiency, and peak power efficiency for the green TOLED reach 84 660 cd·m－2, 94.7 cd·A－1 and 43 lm·W－1, respectively,

In order to determine the topological charge of a Perfect Optical Vortex (POV), a in-line interferometric measurement method was presented. The basic idea is to use a spatial light modulator to produce a POV and a spherical wave simultaneously. By modulating the divergence angle of the spherical wave, the POV and the spherical wave interfer. The number of interference fringes is used to realize the direct and rapid measurement of topological charge. Simulated and experimental results show that the interference fringes obtained by the method can be used to determine the topological charge of POVs, including the magnitude and symbol. Furthermore, the proposed in situ method is demonstrated to determine the topological charge of the POV array by measuring the interference pattern of a perfect vortex array and a spherical wave. The proposed in situ method is simple and effective. Therefore, it is of a significance for the application of orbital angular momentum control and information coding based on POVs.

To elucidate the influence of laser polarization state on optical trap stiffness, the three-dimensional stiffness of four kinds of optical traps with different polarization states (azimuthally polarized beam, radially polarized beam, linearly polarized beam, and circularly polarized beam) to trap silica beads with different sizes was investigated. The result show that, when the silica bead's size is equivalent to the laser wavelength, the three-dimensional stiffness of circularly polarized beam and linearly polarized beam are larger than those of radially polarized beam and azimuthally polarized beam. With the increase of the silica bead's size, the three-dimensional stiffness of azimuthally polarized beam and radially polarized beam are larger than those of circularly polarized beam and linearly polarized beam. In addition, the experimental results also show that, when using an oil-immersion objective lens to trap particles, the spherical aberration caused by the mismatch of the refractive index of oil and water will reduce the utilization of the numerical aperture of the objective lens. The work reported in this paper provides a guidance and reference to the force measurement when using trapping beams with different polarization states.

The micropyramid structures was integrated into the silicon nitride thin film. The refraction, diffraction and interference of the thin film optical microstructures which contains interface were comprehensively utilized to realize the regulation of transmission and reflection.The preparation of micro-pyramid array in SiNx film by single-point diamond turning technology combined with nano-imprint and inductively coupled plasma etching was presented. The large-area, high-efficiency, low-cost microstructure preparation method was applied to generate thin film optical microstructures. The thin film optical micropyramid structures with different sizes were fabricated. The structural unit size could be adjusted between 1.5 μm and 10 μm. Spectral characteristics test results proved the ultra-wide band anti-reflection of the SiNx thin film optical micropyramid array during near-infrared to long wave infrared region. In the near-infrared band of 0.8~2.5 μm, the reflectivity is less than 1.0%. In the mid-infrared band of 3~5 μm, the reflectivity is less than 2.5%. In the long-wavelength infrared band of 10~12 μm, the average reflectance is less than 5%. Compared to antireflection film,the anti-reflective effect of SiNx thin film micro-pyramid structure array can be achieved without the refractive index matching in the design of the films, which simplifies the structure of the films. It is also found that the SiNx thin film micro-pyramid structure array has induced transmission enhancement characteristics at near-infrared. The SiNx thin film micro-pyramid array with a height of 2 μm to 4 μm have obvious transmission-induced enhancement effects at the wavelength of 2.1μm, and the micro-pyramid structure array with a height of 4 μm and a bottom width of 8 μm has the most obvious transmission enhancement effect, and the transmittance reaches 96% at least. The position and intensity of transmission enhancement are determined by the morphology, size and structure proportion of the micro-structure based on experimental test and simulation analysis.