摘要:At present, the liquid crystal materials used in microwave phase shifters are mostly high birefringence polybiphenyls or polyphenylacetylene liquid crystal compounds as the main components. But these liquid crystal compounds inevitably have some disadvantages such as large viscosity, slightly higher dielectric loss or poor photothermal stability. The trans-cyclohexyl(tri)biphenyl liquid crystal compounds with relatively low viscosity and high thermal stability are widely used in displays devices,but their microwave dielectric properties studies have been fewer reported. In this paper, six compounds of trans-cyclohexylbiphenyl and tri-biphenyl isothiocyanate were synthesised selectively, and their liquid crystal phase and microwave (10~30 GHz) dielectric properties were tested. Their dipole moments and polarizabilities were calculated by molecular simulation using density functional theory (DFT), and compared with the properties of the related polybiphenyl liquid crystal compounds to explore the effect of the trans-cyclohexyl structure on the microwave dielectric properties of liquid crystal molecules. The experimental results show that the isothiocyanato-trans-cyclohexyl (tri)biphenyl liquid crystal compounds not only have relatively low melting points (<90 ℃) and wide nematic temperature ranges (86~160 ℃), but also have low maximum dielectric losses (≤7.25×10-3) at 10~30 GHz, which makes their microwave quality factors relatively high (η≥30) and suitable to be used as effective components in microwave-usable liquid crystal materials.
关键词:trans-cyclohexyl(tri)biphenyl;Liquid crystals for microwaves;low dielectric loss;Wide temperature nematic phase;synthesis
摘要:In order to study the dynamic diffraction characteristics and grating profile of phase gratings in liquid crystals (NLC), a new method used for recording the dynamic phase grating with an asymmetric profile in C60 doped homeotropically aligned nematic liquid crystal (NLC) is presented, the generation mechanism of grating is analyzed and the mathematical model of asymmetric profile is established. First, an oblique incidence parallel beams is used to record thin asymmetric dynamic phase holographic grating. Then, a beam of light is used as a probe to read out the grating and record the diffraction intensity with a photodetector. Experimental indicates that the maximum diffraction efficiency is more than 60%, and the diffraction intensity distribution is asymmetrical. Based on the characteristics of asymmetric diffraction distribution, the physical mechanism of producing asymmetric diffraction distribution and diffraction efficiency exceeding the theoretical limit of Raman-Nath diffraction of symmetric gratings is proposed based on photorefractive effect. Finally, the hypothesis of a grating with asymmetrical fringe close to saw-tooth profile formed in the NLC is given, and the mathematical model of describing asymmetric phase gratings is proposed. The results indicate that the high diffraction efficiency of the phase grating originates from the non-sinusoidal modulated space charge field in the liquid crystal. The asymmetric electric field is mainly the result of surface charge modulation and Carr-Helfrich effect. The proposed model of asymmetric saw-tooth profile grating can well describe the reason why the asymmetric diffraction and diffraction efficiency exceed the theoretical limit of symmetric profile gratings of Raman-Nath diffraction theory.
摘要:The interdigital electrode is expected to improve the optical performance of the liquid crystal clad optical waveguide by simplifying the waveguide layer structure. Therefore, a physical model of the liquid crystal clad optical waveguide based on the interdigital electrode is established in this paper, and the changes of the effective refractive index modulation of the waveguide mode under this structure are studied. Firstly, the electric field distribution characteristics of the interdigital electrode are defined by theoretical simulation. Then, the configuration of liquid crystal molecules under electric field and its effect on the waveguide mode effective refractive index are analyzed, and the effects of the initial orientation angle, Δn, Δε on the waveguide mode effective refractive index are quantitatively studied. Finally, the effective refractive index modulation performance of the liquid crystal optical waveguide is compared between the interdigital electrode and the traditional two-sided electrode structure. The results show that the interdigital electrode produces periodic electric field distribution, and the rotation angle of the liquid crystal and the effective refractive index distribution of the optical waveguide are periodic. Compared with parallel orientation (initial orientation angle 2°), the effective refractive index modulation amplitude of vertical orientation (initial orientation angle 88°) is larger, and the effective refractive index modulation amplitude of vertical orientation is 10.1 times that of parallel orientation at 40 V voltage. Larger Δn and Δε are helpful to obtain larger effective refractive index modulation of optical waveguide. Compared with bilateral electrodes, interdigital electrodes can achieve a larger effective refractive index modulation at the same voltage.
摘要:Adaptive optics is widely used for aberration correction. But in practical applications, closed-loop correction requires a longer time, and open-loop correction is affected by the hysteresis effect of deformable mirrors. In this research, we propose a wavefront correction method for pyramid wavefront detector based on the P-U-net. P-U-net establishes a model of the entire adaptive optical system through data training, and constructs a direct nonlinear mapping relationship between the pupil image of the corner cone wavefront detector and the control voltage of the deformable mirror. The control voltage of the deformable mirror can be calculated through the pupil image of the pyramid wavefront detector. The feasibility of this method is verified by using simulation adaptive system theory. It can complete aberration correction within 50 ms, and the correction accuracy can reach 0.1 μm. Compared with traditional closed-loop adaptive correction algorithms, it can complete aberration correction in a shorter time, achieve higher accuracy, and has good application prospects.
摘要:The stability of thin film transistor inverters affects their further applications in the fields such as system on glass (SOG), etc. In this study, a simulation model of amorphous IGZO TFT devices was extracted based on the experimental data. In addition, the stretched-exponential equation between the threshold voltage change (ΔVTH) and the bias stress time was obtained by fitting. Then, the variation of the electrical stability of traditional pseudo-CMOS inverters with bias stress time was explored, and an improved TFT inverter circuit was proposed, followed by the channel width adjustment and layout design. The revised inverter improves the high output voltage by 18.47% by delaying the pull-down transistor in the output stage. Meanwhile, the proposed inverter alleviates the effect of equivalent resistance increase caused by the threshold voltage shift on the output stage current through feedback, significantly improving its speed stability. When the bias stress time is 2.56×107 s, the variation rate in its rise time is only 4.09%, far lower than the 296.11% of the traditional pseudo-CMOS inverters.
关键词:inverter;thin film transistor;stability;amorphous InGaZnO;system-on-glass
摘要:Healthcare of display points out the impact of display technology on human health and has attracted widespread attention within the display industry. However, the current display industry does not attach enough importance to the concept and connotation of health display, blindly pursuing display performance parameters, and ignoring the high-level needs of consumers for display health. Therefore, this paper takes health display as the theme and elaborates on the concept of health display. Health display provides a comprehensive description from the top-level health display requirements to the display hardware technology requirements. Health display clearly states that health is a high-level pursuit of display, and proposes new display technology requirements and corresponding improvement strategies. Furthermore, an overview of the effects of various display parameters such as low blue light emission, red light, high color gamut, high dynamic range, low glare, no screen flicker, and high refresh rate on human health are summarized. On this basis, we explore three technical dimensions for improving healthcare of display, including spectral adaptability, brightness adaptability and frequency adaptability. The main technical strategies and development directions are proposed to improve healthcare of display in the future. We hope this article can attract the attention of display peers and transmit the concept of health display to display practitioners and consumers by combining television manufacturers and media promotion, alleviating the troubles of myopia and other eye diseases caused by the display technology.
关键词:healthcare of display;display parameters;human eye health
摘要:Fracture often occurs in the metal layer of the COF (Chip on Film) transition zone when OLED(Organic Light-Emitting Diode) screens are at the module-bending stage and the reliability verification stage. Based on the design of experiment (DOE), this paper uses a L9(34) orthogonal test scheme to optimize four influence factors: the thickness of the foam, the offset of the foam, the thickness of the MCL(Metal Cover Layer), and the offset of the U-film. Two indexes, the metal layer stress in module-bending stage and the reliability state, respectively are employed and optimized to be minimal. In the L9(34) orthogonal DOE test scheme, finite element method are used to analyze nine groups of COF zones with different influence-factor combinations, and experimental tests are also carried out for verification. The results show that the U-film attachment offset has the greatest influence on the bending stress of the COF metal layer, while the foam thickness has a weaker influence, and the MCL thickness and foam attachment offset have the least influence. For the optimal influence-factor combination, both the module-bending stage and the reliability state have the least metal layer stress, with values of 57 MPa and 523 MPa, respectively. Compared with other combinations, the optimal influence-factor combination have 100% of the yield rate during the experimental tests,satisfying design and production requirements.
摘要:Compared with single-aperture optical imaging technologies, coded aperture imaging (CAI) could achieve higher light intensity and resolution, thus attracting extensive attention in recent years. However, existing CAI technologies have limitations in low temporal resolution and inferior imaging performance. Deep learning technologies have been widely applied in various signal processing fields due to their powerful capabilities in modeling complex features. Therefore, we develop a convolutional attention mechanism based generative adversarial network (CAM-GAN) model to utilize deep learning technologies to address the above problems, which could adapt to different scenarios and task requirements to improve the effects and stability of CAI. Convolutional attention mechanism module is introduced in this model so that the generator can selectively focus on specific areas of the data to recover details and structures of the original image. On this basis, the network is trained in the form of generative adversarial networks to generate more realistic and higher quality images. Experimental results on public datasets show that compared with other methods, CAM-GAN performs excellently on image quality and achieves the highest peak signal-to-noise ratio, improving by about 0.32 over the suboptimal UNet-GAN algorithm, which fully demonstrates the application potential of deep learning technologies in the CAI field.
摘要:Transformer has a wide range of applications in image classification tasks, but in small dataset classification tasks, Transformer is affected by factors such as small amount of data and excessive amount of model parameters, which leads to low classification accuracy and slow convergence speed. Therefore, a progressive hybrid transformer model with hourglass attention is proposed. Firstly, the global feature relationships are modeled by the hourglass self-attention with down-up sampling, and up-sampling is used to supplement the information lost by the down-sampling operation, while the learning temperature parameters and negative diagonal mask are used to sharpen the fractional distribution of the attention to avoid excessive smoothing due to the excessive number of layers. Secondly, progressive down-sampling modules are designed to obtain the fine-grained multi-scale feature maps, which can effectively capture the low-dimensional feature information. Finally, a hybrid architecture is used, where the designed hourglass attention is used in the top stage, the pooling layer is used in the bottom stage instead of the attention module, and layer normalization with deep convolution is introduced to increase network locality. The proposed method is experimented on T-ImageNet, CIFAR10, CIFAR100, and SVHN datasets, the classification accuracy can reach 97.42%, and the computation and parameters are 3.41G and 25M. The experimental results show that compared with the comparison algorithms, the classification accuracy of the proposed method is significantly improved, with a significant reduction in computation and parameters, which improves the performance of Transformer model on small datasets.
关键词:image classification for small dataset;Transformer;hourglass attention;multi scale features;hybrid architecture
摘要:To address the issue of low robustness in visual place recognition due to environmental changes like weather, season and lighting, we propose a solution called parallel omnidimensional dynamic attention (POD-Attention). In order to achieve dynamic and fine-grained exploration of convolutional kernels across all dimensions and enhance the feature extraction network’s ability to capture invariant features like buildings, a complementary attention mechanism is incorporated into the omni-dimensional dynamic convolutional block. This mechanism operates on all dimensions of the convolutional kernels, including input/output channels, convolutional space and kernel quantity, enabling comprehensive attention across the entire kernel space. Furthermore, the parallel fusion of the 1×1 convolution, skip squeeze-and-excitation (SSE) module and omni-dimensional dynamic convolutional block yields notable benefits in terms of both feature extraction speed and the expansion of the receptive field within the visual place recognition network. By combining these components in parallel, the network gains the ability to capture more comprehensive information, resulting in enhanced accuracy for visual place recognition tasks. Experiments conducted on public datasets show that the visual place recognition method based on VGG16 and Patch-NetVLAD feature aggregation improved by the POD attention mechanism, achieves 9.7% increase in Recall@1 on the Nordland dataset and 1.8% increase on the Mapillary Street-Level Sequences dataset. These results demonstrate that the proposed POD attention mechanism effectively enhances the robustness of visual place recognition in different environmental conditions, laying a foundation for more accurate visual localization and map construction in visual SLAM.
关键词:Visual Place Recognition;Environmental robustness;deep learning;Parallel Omni-Dimensional Dynamic Attention;Parallel strategy
摘要:A prototype correction method based on Gaussian distribution is proposed to address the issues of class prototypes being prone to bias and poor network generalization that arise in metric-based few-shot learning methods due to the scarcity of support samples. The proposed method first obtains the class prototypes based on the prototypical network, and perform nearest neighbor matching on the query samples through the class prototypes to get the pseudo labels of the query samples. Then, the Gaussian distribution information of the pseudo labeled sample features is acquired, namely mean and variance. Finally, enough samples are generated by sampling from these distributions to expand the support set, thus obtain more accurate class prototypes and improve classification performance. At the same time, orthogonal constraints are introduced into the existing feature extraction network to improve the generalization of the model. The few-shot classification experiments and further ablation experiments are performed on common point cloud datasets. On the ModelNet40 and ModelNet40-C datasets, the average classification accuracy of the proposed method is comparable to the existing method. On the noisy ScanObjectNN and ScanObjectNN-PB datasets, the average classification accuracy is better than the existing method by 1.36%. The further ablation experiments verify the effectiveness of prototype correction and network parameter constraints. The proposed method can effectively alleviate the overfitting problem in few-shot point cloud classification and has strong robustness against perturbed data.
关键词:3D point cloud classification;few-shot learning;Prototype correction;feature enhancement;Gaussian distribution
摘要:An under-sampled non-uniform density multi-station 3D point cloud alignment method based on manifold clustering is proposed, to address the problem that the point cloud data from each view overlap with each other, and the uneven point cloud density caused by different overlapping areas directly affect the multi-station cloud alignment accuracy. First, the geodesic distance is used as a similarity measure to cluster the unbalanced point cloud data to achieve a streamlined point cloud data. Then, the K nearest neighbour (KNN) method is used to calculate the number of points within the radius of each point, and the point cloud is divided into denser and less dense point clouds. Next, the denser regions are clustered and the surfaces are fitted to each cluster, and the curvatures of all points on the surfaces are calculated. The points with greater curvature are extracted, so that the denser regions and the points with greater curvature are extracted so that the number of point clouds in the denser regions and the less dense regions are balanced, resulting in more balanced point cloud data. Finally, the point clouds are undersampled using manifold clustering and clustered using K-means clustering, which updates the clustering centres and the rigid transformation matrix to achieve non-uniform density multi-station cloud alignment. Compared with the random sampling method and the uniform sampling method, the proposed method has a smaller chamfer distance and preserves the local feature information of the point cloud. The experiments on the Bunny dataset in the Stanford University public dataset indicate that the proposed method improves the alignment efficiency by more than 60% while ensuring the accuracy of the alignment.
关键词:point cloud registration;multi-station point cloud;manifold clustering;point cloud simplification;k-means clustering
摘要:The light field image can record the light information of different position and direction in space simultaneously, which provides rich information for estimating accurate depth map. However, in complex scenes such as occlusion and repeated texture, the lack of feature extraction will lead to the loss of detail in depth map. An optical field depth estimation network based on correction convolution is proposed to make full use of the rich structural information for optical field images to improve the depth estimation of complex areas such as occlusion. The occlusion mask is generated by using the initial disparity map and subaperture image, and the spatial information of the occlusion area is perceived by correcting convolutional discrimination and encoding, and multi-scale features are combined to supplement the edge details that are easily lost.The spatial attention mechanism is used to give more weight to the occlusion area, eliminate redundant information and optimize the subpixel cost body globally. Experimental results show that average MSE and BadPix (ε=0.03) of the proposed method on 4D optical field reference platform are 0.951 and 4.261, respectively. The proposed method can achieve depth estimation with minimum error in most scenes, and shows high robustness to the occlusion area, which is better than other algorithms.
摘要:To address the severe degradation in image quality caused by sand and dust weather conditions, a sand and dust image enhancement method based on the Lab color space is proposed. The enhancement process is decomposed into two steps: color correction and detail enhancement. The color correction part includes color bias removal and brightness stretching. Firstly, the shift characteristics of the sand and dust image histograms in the Lab and YUV color spaces are studied. Then, a Lab space color correction algorithm is proposed to correct the histogram shift, and brightness stretching is applied to enhance the image contrast after color bias removal. For detail enhancement, a haze removal method based on the estimation transmission map of saturation is introduced to further enhance the image’s detail information. Experimental results indicate that compared to other algorithms, the proposed algorithm can effectively remove the color bias brought by sand and dust at different levels,and demonstrates the best performance in terms of time efficiency for small and medium-sized images. In terms of quantitative evaluation, the method proposed in this paper achieves a 3.2% improvement based on a no-reference perception-based image quality evaluator and a 10.7% improvement based on an entropy-based no-reference image quality assessment. Therefore,it can effectively remove the color bias and restore clear images.
关键词:image enhancement;sand-dust image;lab color space;color correction;image defogging