Using the fixed calibration field with automatic observation equipment and sharing the other satellite's moving targets fields, a multi-sensor cooperative radiometric calibration method is proposed to improve calibration frequency. This method is developed based on the spectral flatness of the gray-scale tarps and the consistency of observation phase, geometry and field of the multi-sensors on the same platform. The radiance relationship model between the multi-spectral camera and the high resolution panchromatic camera is established, and then the cooperative radiometric calibration of the multi-spectral camera is realized by the high resolution panchromatic camera. The synchronous experiment results of Mapping Satellite-1 show that the radiance relationship model is universal, the relative difference between cooperative and site radiometric calibration coefficient of the multi-spectral camera is less than 5.3%, and the cooperative radiometric calibration precision is less than 5%. Compared with the site radiometric calibration, the same calibration precision is obtained while the calibration efficient is improved.

Based on global vegetation fluorescence distribution, photon path probability distribution density function factor and the fluorescence intensity at 755 nm are synchronously inverted with GOSAT data. The inversion results are compared with the results of TCCON site. The results show that for the GOSAT data near the Park Falls (45.9°N, 90.3°W) site, which is greatly affected by plant chlorophyll fluorescence, the maximum deviation between CO2 inversion results before and after considering the effect of fluorescence is 1.6×10-6. While for the GOSAT data of near Sodankyla site which is slightly affected by fluorescence, the maximum deviation is 0.8×10-6. The scattering correction of the fluorescence effect can reduce the average error to about 0.1×10-6.

Based on the random phase perturbation of light waves in laser optical paths, the formation mechanism of mode fringe patterns is analyzed in the real-time image subtraction using an electronic speckle pattern interferometry (ESPI). An amplitude-fluctuation measurement method is proposed. The systems of ESPI and digital shearing speckle pattern interferometry (DSSPI) are established and used for the out-of-plane vibration analysis. In addition, the vibration properties of the intact and the cracked cantilever aluminum plates are investigated experimentally. The experimental results show that the visibility of mode fringe patterns obtained in the real-time image subtraction mode is obviously superior to those by the other methods. The obtained first 10 orders of modal fringe patterns are well consistent with the calculation results by the finite element method. Compared with ESPI, DSSPI is more sensitive to the local stiffness variation and flaws of specimens.

The stability of interference fringes in a holographic exposure system directly affects the quality of grating masks. The characteristics of the reference interference fringes are investigated and the detection methods of phase, tilt, and period of the grating are obtained. A three-dimensional (3D) locking system is designed in order to improve the stability of this grating exposure system. The reference interference fringes are detected by a camera. The optical path is finely adjusted by the micro-motors and piezoelectric ceramics. The phase, tilt, and period of the grating are effectively controlled. The experimental results show that the 3D fringe locking system improves the stability of interference fringes, shortens the setup time before exposure, and further enhances the grating contrast, which is benificial for improving the quality of a large-aperture holographic grating.

The ray solutions of paraxial equation for imaging electron optics have been explored by an ideal model of combined electromagnetic concentric spherical system in the present paper. The exact analytical expressions of rotational angle as well as two special solutions of paraxial equation in this system have been firstly derived, and the paraxial imaging properties have also been investigated. The results have been generalized to the bi-electrode electrostatic concentric spherical system, the homogeneous and parallel combined electromagnetic system, and the electrostatic proximity system.

Starting from the two special solutions of the paraxial equation of the combined electromagnetic concentric spherical system, the paraxial longitudinal and lateral aberrations of the combined electromagnetic concentric spherical system have been explored by expanding the representation of the image rotation angle at the ideal imaging position. The main contribution of this paper is to obtain the analytical expression of the two special solutions of paraxial equation at ideal image plane, and it has proved that the paraxial spherical-chromatic aberration of second order that determines the limited spatial resolution of imaging can still be characterized in terms of the Recknagel-Artimovich formula. Expressions of the paraxial lateral aberration composed of paraxial spherical-chromatic aberration, paraxial magnification chromatic aberration, and paraxial anisotropic chromatic aberration have been deduced.

The approximate expressions of special solutions of the paraxial equation and its paraxial lateral aberrations have been explored by an ideal model of a combined electromagnetic concentric spherical system in this paper. The approximate expressions of two special solutions of the paraxial equation in a combined electromagnetic concentric spherical system have been derived, and on this basis, some special types of paraxial lateral aberrations such as paraxial spherical-chromatic aberration, paraxial magnification chromatic aberration, and paraxial anisotropic chromatic aberration have been deduced. The results show that the paraxial lateral aberration deduced from the approximation of two special solutions is exactly the same as through the exact solution, which proves the feasibility of the approximation to solve the paraxial lateral aberration.

The asymptotic solutions of the paraxial equation for a combined electromagnetic concentric spherical system have been explored. Expressions of each kind of coefficients of two special solutions given by the asymptotic method for solving the paraxial equation have been deduced. By verifying the exact solution of two special solutions of the combined electromagnetic concentric spherical system, this paper proves that the approach based on asymptotic solutions given by Monastyrski ［Journal of Technical Physics, 1978, 48(6): 1117-1122］to solve the two special solutions of paraxial equation in electron optics is basically correct and feasible, and only a few aspects need to be improved.

There is an active demand for infrared chalcohalide glass fiber with ultra-low zero-dispersion point. Here, two series of Ge-Se based chalcohalide glasses are selected to compare in the research. For the aim of fiber drawing, an optimized glass composition is selected after comparing the characteristics of each composition. The results show that with the increase of the molar fraction of iodine from 5% to 40%, the infrared transmittance of glasses is significantly improved, the cut-off wavelength of near-infrared absorption appears blue-shift, the optical band gap increases gradually, the refractive index decreases, and the zero-dispersion wavelength is blue-shifted. The glass transition temperature decrease and the expansion coefficient increases gradually. The fiber loss is measured by cut-back method in the spectral range of 2.5-11.5 μm, and the minimum optical loss at 5.9 μm is about 16.9 dB/m.

A self-assessment technique based on Fourier transform is presented to evaluate test environment of fiber optic gyroscope (FOG). Test results show that the zero bias stability of FOG is decreased to 0.0019 (°)/h (100 s, 1σ, that means the standard deviation after 100 s smoothing) in environment 4 from 0.0015 (°)/h (100 s, 1σ) in environment 3, and random walk coefficient is decreased to 2.8876×10-4 (°)/h1/2 in environment 4 from 2.1565×10-4 (°)/h1/2 in environment 3. Another FOG with pulse output is tested under different environments, whose zero bias stability is decreased to 0.0021 (°)/h (100 s, 1σ) in environment 4 from 0.0013 (°)/h (100 s, 1σ) in environment 3. The experiments of two FOGs demonstrate that the proposed self-assessment technique is effective, which provides guidance for the precision test of high precision FOG.

In the indoor visible light communication (VLC) system based on white light-emitting diode (LED), considering the primary reflection of the wall, the layout schemes of different LED topology are studied. By adjusting the topological distribution, the square array combined with the circular ring layout model is proposed to reduce the power consumption of system, and the dynamic range compressions of illuminance and signal-to-noise ratio (SNR) of the receiving plane are realized by power distribution under the indoor lighting requirements. The illuminance distribution, received power distribution, SNR distribution and bit error rate (BER) performance of the square array combined with the circular ring layout model based on power distribution are simulated and analyzed. The simulation results show that the number of light sources used in the square array combined with the ring layout is reduced by 48 compared with the traditional square array layout, but the performance is still improved. At the same time, the relationships between bandwidth, BER, distance and SNR of indoor VLC link system are given, which can provide reference for users to build VLC links.

Based on the phase delay function of a quarter-wave plate and combined with a polarizer and a Faraday rotator, a nonreciprocal phase modulator is proposed for the Sagnac loop, which integrates the functions of polarization and phase delay. The proposed structure fundamentally eliminates the influence of delay coil on system sensitivity, and reduces the complexity and volume of the whole system. The measured relative insertion loss of this phase modulator is 3.6 dB, and the relative phase offset angle is 89.9605°. The experimental results are well consistent with the theoretical analysis results. A variable ratio error self-compensation technology is proposed. In the range of -40~40 ℃, according to the different fabrication process, the minimum variation ratio error of the system is no more than ±0.02% when the temperature changes.

A terahertz wave wide-beam imaging technology based on the block compressive sensing theory is proposed. Simulation results show that this technology can be used to achieve the rapid imaging with high resolution and high quality. With the CO2 gas laser light source of continuous terahertz wave and based on the wide-beam matrix modulation sampling, the different objects are imaged by block compressive sensing. The results are compared with that by the block compressive sensing method based on single-pixel random sampling. It is shown that the proposed technology has a higher imaging stability. Moreover, the sampling process is more universal for different imaging objects.

A landform image classification algorithm based on sparse coding and convolutional neural network is proposed. The non-subsampled Contourlet transform is applied to the training samples for multi-scale decomposition. The images are selected in the training samples to learn the local features by using sparse coding, and the feature vectors are sorted. The feature vectors with larger gray-scale mean gradients are selected to initialize the convolutional neural network convolution kernel. The results show that the proposed algorithm can obtain better classification results than traditional underlying visual features, which effectively avoids the problem of network training falling into local optimum, and improves the classification accuracy of unmanned aerial vehicles landing landform in natural scenes.

A fully automatic and non-rigid hierarchical registration separation ROI (Region of Interest) method based on optimal Atlas image search and local weighted B spline transform is proposed. The experimental results show that the accuracy of the registration of the proposed algorithm is 95.6%, the normalized mutual information value is 1.8432, the root mean square error is 1.12%, and the correlation coefficient is increased by 18.33%. Compared with other registration methods, the registration accuracy and precision of this registered method have obviously improved, which is of great significance for clinical assistant diagnosis.

In order to express the spatial structure information of hyperspectral image more effectively and improve the classification accuracy after dimensionality reduction, we propose a hyperspectral feature extraction algorithm based on linear embedding and tensor manifold. Different from other manifold structure expression methods, the proposed algorithm uses the cooperative representation theory to solve the weight matrix for globally linear embedding, which is more beneficial to maintain the global information of high dimensional data and improve the accuracy of manifold structure expression. At the same time, the dimension reduction framework of tensor manifold based on multi-feature description is established, and the obtained explicit mapping has strong reliability and global adaptability. Experimental results show that compared with the principal component analysis, locally linear embedding, Laplacian Eigenmap, linearity preserving projection and other algorithms, the proposed algorithm has better classification performance.

Based on the lock-in photon-counting technique, a multi-wavelength single pixel spatial frequency domain imaging (SFDI) parallel detection system is proposed. With a digital micro-mirror device (DMD) as the modulated light source of diffuse reflected light to be measured under different wavelengths and another DMD as the equipment for the acquisition, coding, and convergence of diffuse reflection light, the converged diffuse reflected light is demodulated separately under different wavelengths based on the lock-in photon-counting technique, and simultaneously the cost of the proposed system is reduced. The time consumption of single-pixel imaging caused by multiple spatial encoding is effectively compensated via the introduction of the compressed sensing theory. The experimental results show that the proposed SFDI system can realize the reconstruction of diffuse reflection light from phantom surfaces under multiple wavelengths by encoding with only 20% of total pixels.

A fast and efficient segmentation algorithm is proposed for scenes in which multiple plate-shaped objects are placed in an overlapping manner. The algorithm makes full use of the characteristics of the ordered point cloud, and combines the top-down and bottom-up segmentation strategies. The Random Sample Consensus (RANSAC) algorithm is used to quickly extract the three-dimensional planar point set according to the spatial position and normal vector of the three-dimensional point from the three-dimensional point cloud. The image coordinates corresponding to the extracted planar point set are mapped into a binary image, and are divided into a plurality of connected planar regions by the connected region analysis. Then, the glue algorithm is used to quickly merge these regions, and the larger weakly connected regions are subjected to the fracture correction, so as to obtain the final segmentation result. The experimental results show that compared with the region growing algorithm, the proposed algorithm can obtain better segmentation results, and the algorithm efficiency is greatly improved.

A lower-level unfolded and higher-level folded multi-level discrete wavelet transform architecture in single-clock domain is proposed. A three-input line-based scanning method is adopted. The first-level discrete wavelet transform is designed based on the 9/7 discrete wavelet transform lifting scheme. The partially folded second-level discrete wavelet transform is constructed according to the ratio of clock cycles to valid input data. The third-level and higher-level discrete wavelet transform architectures are folded into the second-level discrete wavelet transform architecture to reduce the consumption of hardware resources. The results show that for the input image with size of 512 pixel×512 pixel processed by the three-level discrete wavelet transform, the hardware efficiency of the proposed architecture increases over 57.1% compared with the existing architectures.

To enhance the calibration accuracy of the line structured light sensor (LSLS), we propose a novel calibration method based on the planar reference target. A method is brought out to calculate the coordinates of the reference target point centers and the laser plane equation in the camera coordinate system. These reference target center points are projected onto the laser plane. The world coordinate system is established with the projection point of the lower left corner as the origin. The corresponding world coordinates for each center of the target point, which has known pixel coordinates, can be obtained. The mapping relationship between the arbitrary pixel coordinates and the world coordinates is achieved via the polynomial fitting and the calibration of sensor is finished. The calibration results are the coefficients of the polynomials, which are convenient for storage and use. The calibration accuracy is analyzed. The results show that the root mean square of distance error between each testing points and the first point is 0.0216 mm, which is 22% lower than the classical method. The proposed method has certain advantages in high-precision calibration of LSLS.

Based on the application requirements of large aperture and short focal length Fresnel lens as an antenna in the underwater wireless optical transmission system, an experimental device was built to measure the focusing performance of Fresnel lens. The 450 nm and 532 nm lasers were adopted as the test light sources, and the relationship curves between the focusing performance of Fresnel lens with an aperture of 75 mm and a focal length of 25 mm and the lens surface, laser wavelength and incident angle were obtained. The influence of incident angle of laser on the focusing spots of the 450 nm and 532 nm lasers was measured. The experimental data results show that, under the same experimental conditions, the focusing efficiency for laser passing through the serrated surface of Fresnel lens is 10%~15% higher than that through the smooth surface. In addition, the focusing efficiency of Fresnel lens for 532 nm laser is about 5.4% higher than that of 450 nm laser. With the increase of incident angle, the focusing efficiency of Fresnel lens decreases gradually, and the focusing efficiency at a 0° incident angle is 25% higher than that at ±30°. Furthermore, when the incident angle of laser is 5°, the comet aberration occurs at the focusing spot of the 450 nm laser, and when the incident angle of laser increases to 15°, the comet aberration also occurs at the focusing spot of the 532 nm laser.

The model for simulating the influence of intensity scintillation on pulse arrival time jitter is established based on the principle of multiple phase screens and the power spectral inversion method. The pulse arrival time jitter is simulated under different receiving apertures and different transmission distances. The simulation results show that the pulse arrival time jitter shows a positively skewed distribution under the condition of aperture receiving. A large receiving aperture can effectively suppress the pulse arrival time jitter. The variation of transmission distance has less influence on the pulse arrival time jitter at the far end compared with at the near end. An experimental platform for the pulse arrival time jitter caused by atmospheric turbulence is established, the second pulse waveform propagating through a turbulence channel is collected, and the pulse arrival time is determined by threshold detection. The experimental results show that the distribution of pulse arrival time jitter is consistent with the theoretical analysis results.

A new vision-based method for measuring linear displacement is proposed. The model of mapping function between the gray value of each pixel and the linear motion displacement is established. The displacement of the linear motion platform is obtained by the numerical method. The Monte Carlo method is used to analyze and verify the feasibility and rationality of the proposed method. In addition, the influence of detection point topology on the measurement accuracy is analyzed. The experimental results show that the standard deviation of the measurement error is 4 μm when the measurement range is 10 mm.

An online optimization method for improving the performance of a range-gated imaging system is introduced by means of adaptively adjusting the operation parameters of modules. The eye pattern parameter is treated as a random variable varying with the tunable parameters of the system, and the Gaussian process regression method is applied for fitting this variation relationship. In addition, the parameter optimization procedure is also involved in the learning process. Thus the variable characteristic of a random process in the high dimensional parameter space is learnt rapidly, and simultaneously the system parameters are optimized. The experimental results show that the proposed method can be used to optimize the system parameter configuration online and improve the range resolution. The dependence of the range resolution on the configuration parameters of modules is illustrated, which provides a basis for the design of an imaging system.

The time-domain radiative transfer equation is used to simulate the propagation of pulsed laser in dispersive medium. The generalized Guess-Markov random field model is adopted to establish a regularization term and thus the ill-posed property of an inverse problem is overcome. The inverse problem is solved by the step acceleration method combined with the sequential quadratic programming algorithm. An optical imaging algorithm which does not rely on the initial values is proposed to reconstruct the distributions of absorption and scattering coefficients in the two-dimensional dispersive medium. The near-infrared optical imaging of the cancerous biological tissues is simulated. The research results show that the proposed algorithm can not only relatively well solve the initial value dependent problem, but also possesses high accuracy. The reconstructed images can show the inner structures of medium clearly.

Based on the relative position relationship between the measured object and the high-frequency fringe in the on-line three-dimensional (3D) measurement, an improved dual-frequency on-line 3D measurement method is proposed. By controlling the acquisition points when the measured object moves, the light intensity distribution of the high-frequency fringe in each deformed pattern after pixel matching can be exactly consistent. The full-cycle equal phase-shift algorithm is directly applied for phase calculation so that the filtering process is avoided and the 3D reconstruction accuracy is improved. The light intensity component of the high-frequency fringe is designed to be much smaller than that of the low-frequency fringe, and it can be considered as faint background light and the interference of high-frequency fringe on phase calculation is further reduced. The simulated and experimental results show that compared with other methods, the proposed method can improve the reconstruction accuracy effectively.

An interferometric projection Fourier transform profilometry based on cube splitting prism is proposed. The interference projection theory of cube splitting prism is theoretically analyzed and simulated based on Young's double-slit interference principle. By adjusting the installation angle of the prism, any adjustment of the interference projection fringe period can be realized. The results prove the feasibility and flexibility of the interference projection system. The three-dimensional shape reconstruction of a small foot model is carried out, and the experimental results show that the interference fringe phase is stable and the intensity distribution is uniform, the proposed method can be used to recover the three-dimensional shape of the object well.

A new type of spot-size converter is proposed based on the self-focusing effect of a graded-index waveguide. The field distributions and propagation characteristics of the modes in the uniform-index waveguides and the graded-index waveguides are investigated. In addition, the principle of self-focusing effect is analyzed. With the finite-difference time-domain method, the insertion loss and the return loss of the proposed converter are calculated. The performance of the proposed converter is as good as that of the existing best device, however the length of the proposed converter is reduced to 1/6 of the the existing best device. The novel spot-size converter can be used in the integrated optics.

The effect of working gas on the output energy stability of a repetitively pulsed HF laser is investigated. The calculated results show that the decomposition of SF6 in the discharge region of the laser is about 1.5%, a little larger than the experimental data. The gas consumption rate is about 2.2×1015 (cm-3·s-1), and the consumption is 3~4 orders of magnitude smaller than the initial content. The stability of the laser output energy can be improved by properly controlling the HF molecular concentration in the laser chamber and replenishing the working gas at a right time. The change of the laser output energy can be divided into two stages, a transitional stage and a maintenance stage. The transitional stage plays a decisive role in the determination of the output power level of a repetitively pulsed HF laser. The calculated results can provide a technical support for the design of a repetitively pulsed non-chain HF laser.

A double pulse dissipative solitons (DSs) passively mode-locked Yb-doped fiber laser with a tunable wavelength is proposed and demonstrated in the all-normal-dispersion (ANDi) system. Using a Sagnac loop as a tunable all-fiber format spectral filter and the effect of nonlinear polarization evolution (NPE) as the mode-locking mechanism, a stabilized double pulse DSs mode-locked pulse output is obtained in the ANDi Yb-doped fiber system. When the polarization state of the Sagnac loop changes, the tuning of the double pulse DSs output spectrum is realized with a tuning range of 1038.96-1044.80 nm. Furthermore, by changing the pump power of the laser cavity, the pulse-spacing of the double pulse can be changed from 0.034 μs to 0.021 μs. Under the pump power of 520 mW, the double pulse DSs pulse has the maximum energy of 34.3 nJ and the repetition rate of 2.285 MHz.

Aiming at the problems of decoupling and analysis of optical flow vectors with various motion states confronted in the pose estimation process of autonomous vehicles (AVs) using optical flow, the optical flow motion field model (OFMFM) is derived and the decoupled optical flow motion filed model (DOFMFM) is analysed as the vehicle poses change independently in six degrees of freedom. According to the DOFMFM, a simulation algorithm is designed for the vehicle platform, and the completely decoupled simulation results are presented. The DOFMFM is applied to quantify and verify the simulation results. Two real scenes of translation and rotation from the KITTI dataset are utilized for the flow-decoupled experiments. The consistency among model analysis, simulation process, real data and comparison results is verified. The results show that the proposed decoupled model analysis, simulation algorithm, simulation and real results together with comparison analysis can not only be applied in the error analysis and algorithm test of pose estimation, but also provide a reference or instruction for the improvement in the understanding of optical flow motion imaging and the research on the optical flow based applications on an AV platform.

In order to improve the accuracy of one-dimensional (1D) multi-camera calibration, a gradual calibration method based on the normalization algorithm is proposed, where the projective projection matrices are first obtained from the fundamental matrices and then transformed into the metric projection matrices. The coordinates of the image feature points of the calibration object are pre-processed by normalization, improving the accuracy of calibration and simultaneously maintaining the advantages of fast and easy implementation of the linear method. In the proposed calibration method, the 1D calibration objects can move freely without restriction of the site environment and are flexible to use as well. The simulation and real experiments demonstrate that the normalized feature point coordinates can substantially improve the accuracy and robustness of calibration results.

The spatial regularization correlation filtering tracking algorithm based on deformable diversity similarity is proposed. The spatial regularization weight and the sub-grid detection method are introduced on the basis of the kernelized correlation filtering (KCF) tracking algorithm. The target re-detection module is constructed with deformable diversity similarity matching algorithm, the nearest neighbor search problem in the matching algorithm is solved via the principal component analysis (PCA) algorithm and the k-direction tree coherence approximate nearest neighbor (TreeCANN) algorithm. Through the adaptive template updating strategy, the problem of false updating of the template under occlusion is solved. Experimental results show that the precision score and success score of the proposed algorithm are 0.825 and 0.625, which are 18.5% and 31.0% higher than those of the KCF algorithm, respectively. The proposed algorithm can better solve the tracking problems of the target scale variation, occlusion, fast motion, rotation and background clutter, showing a wide range of application prospect.

To satisfy the real-time requirements of the online object tracking algorithm and improve the robustness of the algorithm, we propose a correlation filter-based tracking algorithm with high-confidence updating strategy. Multi-features are extracted and integrated in the target region to construct robust appearance representation, and the projection matrix for dimension reduction of features is used to improve the operational efficiency of the algorithm. The correlation filter is used to localize the target at a high speed via the maximum response value. Two indicators of maximum response value and average peak-to-correlation energy are utilized to design a high-confidence updating strategy. The results show that the proposed algorithm achieves high tracking precision and success rate on large-scale public datasets while running at 122.3 frame/s on average.

In order to improve both the tracking accuracy and speed of the efficient convolution operators based tracking algorithm fusing the histogram of oriented gradient and color names features (ECO-HC), a weighted correlation filtering algorithm based on context and relocation is proposed. Considering the differences between the histogram of oriented gradient and color names features, the responses of two features are fused in different weights. The adaptive iterative method is used to predict the position of a target, which combines with the multi-scale search area, the contextual features and the relocation method when the target prediction is failure to further improve the tracking accuracy. The algorithm is evaluated on the OTB-100 dataset. The experimental results show that the average distance accuracy of the proposed algorithm is 89.2% and the average overlap rate is 80.6%, 3.6% and 2.1% higher than those of the ECO-HC method, respectively. In addition, the tracking speed on the central processing unit is 65.2 frame/s, superior to that of the other tracking algorithms compared in the experiments. The proposed algorithm effectively improves the tracking accuracy and can track the objects well under the condition of severe occlusion, illumination variation and other interferences.

Based on depth separable convolution, a small object detection network for embedded platform, MTYOLO (MobileNet Tiny-Yolo), is proposed. It divides the image into many grids and replaces the traditional convolution by the depth separable convolution, which decreases the number of parameters and computational cost. The point convolution and the feature map merging are adopted to improve the detection accuracy. The experimental results show that the size of the proposed MTYOLO network model is 41 MB, approximately 67% of that of Tiny-Yolo model. Furthermore, its detection accuracy on the PASCAL VOC 2007 dataset is up to 57.25%, superior to the Tiny-Yolo model’s. The proposed model is particularly suitable for application in embedded platforms.

An algorithm of recognition and classification of three-dimensional (3D) model based on deep voxel convolution neural network is proposed. The voxelization technology is used to transform 3D polygon mesh model into a voxel matrix, and the deep features of the matrix are extracted by the deep voxel convolution neural network to enhance the expressive ability and difference of the features. The experimental results on ModelNet40 dataset show that the accuracy of the algorithm can reach about 87% for recognizing and classifying 3D mesh model. The constructed deep voxel convolution neural network can effectively enhance the feature extraction and expression ability of 3D model, as well as improve the classification accuracy of large-scale complex 3D mesh models, which is better than the current mainstream methods.

SiO2/VO2 thermochromic coatings are prepared via VO2 sintering alone and SiO2/VO2 co-sintering, respectively. The microstructure, thermochromic properties and optical properties of samples are studied. The experimental results show that both the microstructure and the above properties of the coatings are strongly dependent on the deposition process, whereas the type of SiO2 precursor sols only have little effect on the microstructure and even less effect on the thermochromic and optical properties. When the coatings are prepared by the VO2 alone-sintering routine, the SiO2/VO2 coatings obtain high transmittance under the condition of slight loss in thermochromic properties. The double-layer coating sample exhibited good comprehensive performance, its near infrared switching rate at 2000 nm value is 39.6%, the solar energy modulation value is 8.4%, and the integral visible transmittance and the integral solar transmittance at 25 ℃ are 68.4% and 72.0%, respectively.

A dual-band chiral metasurface structure with giant optical activity and negative refractivity is proposed, which is composed of a periodic array of a middle dielectric layer and two conjugated gammadion layers. The reasons for the occurrences of giant optical activity and negative refractive indexes are clarified by analyzing the surface current density distributions. The effects of the radius of the connected circular metallic patch in the unit structure of the chiral metasurface and the thickness of the dielectric layer on optical activity and negative refractivity are studied. The numerical simulation results show that there are four resonance frequencies within the frequency range from 0.1 THz to 2 THz, and near these resonance frequencies, the average refractive index is negative and the maximal magnitude of the real part is -3.7. The proposed structure indicates its exceptional giant optical activity as well as its dual-band negative refractivity for a left circularly polarized wave and a right circularly polarized wave near these resonance frequencies. The maximum polarization rotation angle reaches 122° and the magnitude of the real part of the refractive index of a right circularly polarized wave reaches -12.74.

The finite-difference time-domain method is used to simulate uniform moth-eye structure (UMS) and bi-hybrid moth-eye structure (BHMS). The influences of height and bottom diameter on the reflectance are analyzed, and the anti-reflection properties of UMS and BHMS are compared. The superior anti-reflection property of BHMS is analyzed based on the reflectance curves and the electric field intensity distribution. The results show that the reflectance of UMS decreases with the increase of height, and decreases and then increases with the increase of bottom diameter in the wavelength range of 300-1200 nm. The average reflectance of BHMS consisting of UMS with different heights is greater than the minimum average reflectance of the corresponding UMS when the bottom diameter is 250 nm. The anti-reflection property of BHMS consisting of UMS with different bottom diameters can be further improved.

The N-order coupled nonlinear Schr dinger equation with variable coefficients is studied and its 3-soliton solutions are obtained. The soliton interaction is studied by the asymptotic analysis and graph analysis. The results show that when the eigenvalues are different, the 3-soliton solutions can be expressed as the regular solitons, the bounded solitons, and the combination of them, respectively. Under the certain conditions, both the regular bright solitons and the bounded bright solitons can realize the elastic and the inelastic interactions, respectively. However, the dark solitons only have the inelastic interaction. But for the combination of the regular and the bounded solitons, the interacting rules of the bright soliton components are very complex and significantly influenced by the value of the parameters, but the dark soliton components still maintain the elastic interaction.

In order to solve the problem that the quantum key distribution protocol based on the heralded pair coherent state (HPCS) adopts polarization coding and phase coding to bring the basis dependence, a measurement device independent quantum key distribution protocol for asymmetric channels based on the HPCS and orbital angular momentum(OAM) is studied. The relationship among the mean photon number, bit error rate, key generation rate and channel transmission loss of the protocol at different distance ratios is analyzed. The performances of measurement device independent quantum key distribution protocols for symmetric and asymmetric channels with the HPCS and OAM are compared. The simulation results show that the use of the HPCS compensates for the lack of weak coherent source and heralded single photon source, greatly reducing the vacuum pulse and increasing the single photon pulse. As the channel transmission loss increases, the key generation rate and the secure transmission distance gradually decrease, but the performance of the asymmetric channel is still better than that of the symmetric channel.

The transmission properties of the quantized probe field traveling in one-dimensional ultra-cold Rydberg atom samples are studied. Based on the dipole blockade effect, the steady-state spectrum of the electromagnetically induced transparency (EIT) of a four-level atom exhibits the typical nonlinearity at single-photon level. Namely, the field transmittance and photon correlation depend on the probe field intensity before the saturation. By changing the single-photon detuning of the two classical control fields, we realize the flexible manipulation of the nonlinear EIT behavior. By varying the ratio of the Rabbi frequencies, we can observe the transformation of nonlinear EIT from the four-level atomic system to the three-level trapezoidal atomic system.

A separation type fiber sensor structure is designed to improve sensitivity, in which double Fabry-Perot (F-P) cavities with similar cavity lengths are connected in parallel. The sensitization principle of this structure is theoretically analyzed and two groups of sensitizing structures are prepared. The experimental results show that the pressure sensitivity of the sensitizing structure is increased to 43.95 nm/MPa from 4.85 nm/MPa of the single F-P structure. The temperature sensitivity is increased to 0.40364 nm/℃ from 0.0675 nm/℃ of the single F-P cavity. At the same temperature, the use of double cavity structure can eliminate the effect of temperature cross-sensitivity on measurement results. The structure overcomes the defects of integrated sensitization structure, and increases the sensitivity without affecting the original sensor structure. The sensitivity can be adjusted by replacing the auxiliary cavity. This structure has the advantages of good portability and small cross-sensitivity, etc.

The microwave radiative transfer modeling PWR (P. W. Rosenkranz) and back propagation neural network method are used to construct the models of forward modeling of downward radiance brightness temperature and inversion of atmospheric relative humidity profile respectively, and the channel selection of inversion of atmospheric relative humidity profile by hyperspectral microwave radiometer under clear sky conditions is studied. The research results show that the information content of 200 channels is greater than the information content of 7 channels of microwave radiometer and the increase of the detection channel can improve the inversion accuracy of the atmospheric relative humidity profile. When we use the first 120 channels of information content for simulation experiments, the inversion accuracy of the atmospheric relative humidity profile in the range of 0-2 km and 6-10 km increases by 4%~10%, and the inversion accuracy of the relative humidity profile in the range of 2-6 km increases by about 10%. When the channel continues to increase, the inversion accuracy of atmospheric relative humidity profile not improves much.

We propose a triggering resample method by improving frequency modulated continuous wave laser ranging technology. In our method, the auxiliary signal triggers the data acquisition card to sample both of the measurement signal and the auxiliary signal, and then, the extreme points of the acquired auxiliary signal are used to re-sample the measurement signal. The experimental results show that the minimum standard deviation of the proposed triggering resample method can reach 12 μm.

A fully convolutional network (FCN) model based on U-Net is proposed to implement the semantic segmentation of remote sensing images with high resolution, in which the data standardization and data augmentation are adopted for data preprocessing. In addition, the Adam optimizer is used for the model training and the average Jaccard index is used as the evaluation metric. A weighted cross entropy loss function and an adaptive threshold algorithm are employed to improve the classification accuracy of small classes. The experimental results on the DSTL dataset show that the proposed method can increase the average Jaccard index of prediction results from 0.611 to 0.636, and produces an accurate end-to-end classification for high-resolution remote sensing images.

Under the two common conditions of incident natural light and linearly polarized light, we simulate the polarization degree and polarization ratio distributions of whole scattered light and various orders scattered light based on Mie theory and Debye series expansion. The results show that the scattered light is usually partially polarized when incident light is natural light. The distribution of polarization degree is consistent with Rayleigh scattering when the particle size is less than the wavelength. When the particle size is larger than the wavelength, the polarization degree of the whole scattered light is very low at the forward scattering angle, and increases with the increase of scattering angle but changes frequently, and get a high value at the rainbow angle. The scattered light is commonly elliptically polarized light when the incident light is linearly polarized light. In this case, the polarization ratios of the whole scattered light and various orders scattered light are high. A minimum value occurs only when the azimuth angle is far from 0°, 90° and 180°.

The magnetic field distributions and photoelectron trajectories of the camera are simulated, the two-dimensional distribution model of the transmission time and distance are built, and the causes and reduction methods for the magnetic focusing temporal dispersion and distortion are analyzed. The research results show that, the temporal dispersion and distortion are closely related to the magnetic field intensity along the axis of object point (cathode position), which increase with the increase of the off-axis distance of object point indicating that the magnetic focusing temporal dispersion and distortion can be reduced by the improvement of the uniformity of the axisymmetric magnetic field. As the gap of the magnetic leakage broadens from 4 mm to 40 mm, the peak value ratio between the on-axis magnetic field intensity and the 30 mm off-axis one increases from 0.82 to 0.89. Simultaneously, the temporal dispersion is reduced from 127 fs and 435 fs to 120.6 fs and 378.4 fs, respectively, and the temporal distortion for 30 mm off-axis decreases from 4.61% to 4.01%.