最新刊期
卷 34 ， 期 16 ， 2026
- 摘要：In response to the engineering challenges in the interferenceless coded aperture correlation holography （I-COACH） partial aperture imaging system （PAIS）， including uneven modulation transfer function （MTF） distribution of the conventional annular aperture array， mid-frequency information loss， and constrained imaging quality， a partial aperture imaging system integrating the Golay-6 aperture array with I-COACH was proposed. Employing a combined approach of theoretical analysis， numerical simulation， and experimental validation， wecompared the frequency response characteristics of the 8-element annular aperture array， 6-element annular aperture array， and Golay-6 aperture array， and optimized the generation conditions of the coded phase mask （CPM） and parameters of nonlinear reconstruction （NLR）. Experimental results demonstrate that the Golay-6 aperture array reduces the system aperture area by 25%， lowers the MTF fluctuation coefficient to 0.137， and minimizes the maximum mid-frequency fluctuation amplitude to 0.024， achieving a system signal-to-noise ratio （SNR） of 14.82 dB. For the complex sample of a dragonfly wing， the structural similarity （SSIM） reaches 0.975 with an SNR of 13.5 dB， enabling clear reconstruction of fine structures. This system enhances mid-frequency modulation capability and image uniformity while preserving high resolution， thereby providing experimental references and methodological support for astronomical partial aperture imaging and high-performance computational optical imaging.关键词：Interferenceless Coded Aperture Correlation Holography （I-COACH）;Partial Aperture Imaging System （PAIS）;Golay-6 aperture;Modulation Transfer Function （MTF）;Nonlinear Reconstruction （NLR）1|0|0更新时间：2026-08-26
- 摘要：Traditional reflective filters had narrow color gamut， low energy utilization and low fabrication efficiency. These problems limited their applications. This work studied micro-nano structures for reflective filtering. A one-dimensional dielectric grating-metal-dielectric metasurface was designed. The materials were photoresist， Ag， and TiO₂. The finite-difference time-domain method was used. It analyzed key structural parameters. These parameters included grating groove depth， dielectric layer thickness， duty cycle and grating period. The structure was optimized based on these analyses. The results show that the structure enables continuous color tuning. The tuning covers the whole visible band. The grating period ranges from 280 nm to 480 nm. The full width at half maximum is below 100 nm. This holds for all reflection spectra. The electric field is confined in the TiO₂layer under guided-mode resonance. This confinement leads to high energy utilization. The measured central wavelengths of the RGB peaks are obtained. They are 613 nm， 562 nm， and 491 nm. The corresponding peak reflectivities are measured. They are 55%， 57%， and 54%. These values agree well with simulations. A time-division multiplexed multi-interference optical field is used. This field generates multiple interference patterns. It fabricates a 3×3 pixel structure in one step. This method greatly improves fabrication efficiency. It avoids multiple exposures and reduces fabrication time significantly. The proposed structure shows excellent spectral selectivity and high color purity. The colors are vivid and stable. The structure is compatible with micro-nano fabrication. It can be integrated with existing processes. It can support display imaging applications， infrared detection and biosensing. These fields benefit from its high performance. The structure offers a promising solution. It addresses key challenges in traditional reflective filtering.关键词：reflective filtering;one-dimensional metasurface;color filter;micro-nano structure;time-division multiplexed multi-Interference optical field1|0|0更新时间：2026-08-26
- 摘要：Vortex beams possess helical phase wavefronts and carry orbital angular momentum （OAM）. According to polarization characteristics， they can be classified as scalar and vector vortex beams. Scalar vortex beams usually have nearly uniform polarization distributions. Vector vortex beams additionally exhibit spatially varying polarization structures. These distinctive field properties support applications in microscopy， material processing， optical communication， and quantum information. In recent years， intracavity generation techniques for vortex beams have continued to develop. They offer advantages in mode purity， output power， and system stability. This review summarizes recent progress in vortex-beam lasers. It introduces the generation principles of scalar and vector vortex beams. It also discusses their implementation methods and technical characteristics. Intracavity gain modulation， intracavity loss modulation， and extracavity device-based control are compared. Their advantages， limitations， and applicable conditions are evaluated. Special attention is given to mid-infrared vortex-beam lasers. The review outlines nonlinear optical up-conversion and down-conversion routes. It also summarizes their main implementation approaches and representative research progress. Different schemes are compared in terms of wavelength extension， OAM mode control， and pulse characteristics. Particular emphasis is placed on relationships among laser sources， modulation methods， and conversion processes. The comparison highlights how each route balances flexibility， efficiency， stability， and implementation complexity. The review further distinguishes source-generation problems from vortex-mode control problems. It also discusses the coupling between these two aspects. Key challenges include crystal compatibility， phase matching， thermal effects， and optical damage. Other challenges involve spatial mode matching， mode purity， and faithful OAM transfer. The review finally discusses future development trends. Possible directions include higher mode purity， broader wavelength coverage， shorter pulses， and greater system integration. These discussions support a clearer understanding of current opportunities and remaining limitations. This work aims to clarify the main technical routes in this field. It also provides a reference for future research on mid-infrared vortex-beam lasers.关键词：vortex beam lasers;scalar vortex beams;vector vortex beams;mid-infrared band;orbital angular momentum1|0|0更新时间：2026-08-26
Modern Applied Optics
- 摘要：To enhance the tracking performance and disturbance rejection capability of a piezoelectric Fast Steering Mirror （FSM）， this paper proposed a μ-synthesis-optimized high-gain loop shaping （Oμ-HGLS） method. First， based on hysteresis inverse compensation， the dynamic characteristics of the piezoelectric FSM were identified， and the modeling errors were treated as model uncertainties. Then， the μ-synthesis method was employed to perform robust performance optimization of the Q filter， increasing the loop gain while accounting for model uncertainties， thereby improving tracking accuracy and disturbance rejection without compromising system robustness. Finally， comparative experiments were conducted on the piezoelectric FSM to validate the effectiveness of the proposed Oμ-HGLS method. Experimental results demonstrate that when tracking a 100 Hz sinusoidal trajectory， the root mean square error of the Oμ-HGLS method is reduced by approximately 53.4% compared to the conventional HGLS method， and by approximately 18.7% compared to the H∞-optimized high-gain loop shaping （Oh-HGLS） method. Under a single-frequency disturbance， the root mean square error of the Oμ-HGLS method is reduced by approximately 54.8% relative to the HGLS method and by approximately 17.6% relative to the Oh-HGLS method. The above results demonstrate that the proposed method has significant advantages in improving the control performance of piezoelectric fast steering mirrors， providing an effective control strategy for high-precision motion control.关键词：piezoelectric fast steering mirror;μ-synthesis optimization;high-gain loop shaping;high-precision motion control1|0|0更新时间：2026-08-26
- 摘要：To investigate the influence of servo frequency in the control system of a multi-axis coordinated ultra-precision machine tool on its dynamic performance， this study established a servo frequency matching and performance evaluation method combining simulation and experimentation. A simulation model of the ultra-precision machine tool servo control system was developed， and the effect of servo frequency on the output accuracy of the linear motor-driven guideway was examined using a single-variable approach. The simulation results indicate that a higher servo frequency helps suppress the influence of external disturbances on motor thrust output. In this study， tracking error and repeat positioning accuracy were used as performance indices to validate the machine tool system. A position-control experimental platform was built， and the experimental results show that increasing the servo frequency reduces tracking error， improves response speed， and enhances positioning accuracy. However， due to hardware limitations， when the servo frequency exceeds 8 kHz， the proportion of dead time becomes too large， thereby reducing the closed-loop stability margin； therefore， the servo frequency should not exceed 8 kHz. To address the difficulty of directly correlating servo frequency with machine tool performance， linear correlation analysis was employed to establish a direct relationship between the positional experimental results and the simulation results， thereby identifying 8 kHz as the optimal servo frequency. In the single-crystal nickel turning experiments， the workpiece machined under the optimal servo frequency exhibited improvements of 23% in Sa and 13% in P-V compared with those machined under non-optimal servo frequencies. The results demonstrate that the optimal servo frequency can effectively improve surface quality and enhance the machining performance of ultra-precision machine tools， providing a theoretical basis for the hardware design and performance optimization of ultra-precision machine tool control systems.关键词：ultra-precision machine tools;servo frequency;linear correlation degree;repetitive positioning accuracy;following error1|0|0更新时间：2026-08-26
- 摘要：Electrohydrodynamic Direct-Writing （EDW） technology can efficiently fabricate micro-nano structures in one step and is widely applied in fields such as micro/nano sensing and flexible electronics. However， charged jet of EDW is susceptible to charge interference， which limits the stability， uniformity and positioning accuracy of the printed micro-nano structures. This is a major challenge of EDW in the micro-precision system integration manufacturing process. In this study， an auxiliary needle tip was introduced for electric field focusing. The anti-interference ability of the charged jet was enhanced. The deposition and printing behavior under the influence of the focused electric field was investigated. Experimental results indicated that electric-field-focused EDW was beneficial to uniform deposition of fibers. Under the condition of a collector velocity of 20 mm/s， non-spiral micro-nano structures could be obtained. The average diameter of the micro-nano structures printed by electric-field-focused EDW was reduced from 4.84 μm to 2.85 μm， and the diameter standard deviation was reduced from 2.03 μm to 1.17 μm. With the help of the precise motion platform， cross-grid patterns were printed successfully. The error between the average size of the grid and the preset size is only 0.4%. A complex emblem pattern and a phoenix flower pattern were fabricated. The deposition morphology was complete， and the fibers were smooth and continuous. Focused electric field could significantly enhance the controllability of printing of charged jet， which was of great significance for the popularization and application of EDW technology.关键词：Electrohydrodynamic direct-writing;electric field focusing;auxiliary needle tip;patterned printing1|0|0更新时间：2026-08-26
Micro\/Nano Technology and Fine Mechanics
- 摘要：To address severe video degradation caused by high-energy particle radiation in real underwater environments of nuclear power plants， this study proposed a zero-shot underwater video denoising method that reduced the dependence on paired clean samples and task-specific training data， and improved the usability of visual information for underwater inspection. A pre-trained image restoration diffusion model was used as the backbone. Low-quality noisy videos were processed in batches， and key frames were selected as reference inputs. In the early denoising stage， hierarchical latent deformation and optical-flow alignment were used to impose global structural constraints and local motion constraints. In the middle and later stages， a hybrid flow-guided spatial-aware token merging strategy was introduced. Optical-flow matching was used in the downsampling stage to maintain motion consistency， spatially constrained similarity was used in the upsampling stage to preserve structural details， and a merging-ratio annealing mechanism was adopted to dynamically balance temporal consistency and detail fidelity.On the collected underwater noisy video dataset from a nuclear power plant， the proposed method achieved MaxVQA， CLIPVQA， and Ewarp values of 34.2， 0.97， and 1.2， respectively， outperforming the compared video restoration and denoising methods. The proposed method enables effective video denoising in underwater nuclear power plant scenarios where annotated training data and clean reference videos are difficult to obtain， and provides technical support for underwater optical imaging enhancement， intelligent image analysis， and equipment defect detection.关键词：nuclear radiation;zero-shot learning;underwater video denoising;diffusion model;token merging2|0|0更新时间：2026-08-26
- 摘要：Aimed at the problem of low segmentation accuracy caused by the similarity between dynamic ground targets and background feature classes， as well as insufficient weak edge perception， a feature modeling link was constructed， which included edge prior explicit selection， spatial channel dynamic complementary aggregation， and dual domain downsampling high-frequency compensation. An edge guided selection and feature aggregation method for dynamic object instance segmentation in aerial photography of patrol aircraft was proposed. Firstly， a multi-scale edge guided selection perception module was designed， which took the residual response between multi-scale features and local means as the edge prior and generated an edge response map. Then， a dual domain selection mechanism was applied to enhance the model's ability to perceive weak edge information of dynamic targets. Secondly， an adaptive feature aggregation module was constructed to utilize channel level interaction through a channel shuffling mechanism to activate cross-group semantic information， and to aggregate spatial semantic information extracted adaptively by dynamic convolution and deep dynamic convolution with complementary channel detail features， in order to improve feature responsiveness and solve the problem of feature similarity between target classes. Finally， to overcome the limitations of ordinary convolution in extracting multi-scale spatial information， a dual domain downsampling adaptive module was designed to fuse the hierarchical feature information extracted from dual paths， achieving adaptive aggregation and feature enhancement capabilities of multi-scale information. The experimental results show that the edge guided selection and feature aggregation proposed in this paper for the dynamic object instance segmentation models mAP50-95 and mAP50 in aerial photography achieve segmentation accuracies of 51.8% and 84.3%， respectively， which are 5.1% and 4.6% higher than the benchmark model. Deploying the models on the Jetson Nano B01 mobile development platform can achieve dynamic object segmentation.关键词：instance segmentation;dynamic object;edge guided selection;adaptive feature aggregation;model deployment1|0|0更新时间：2026-08-26
- 摘要：In medical image computer-aided diagnosis （CAD）， deep learning models make predictions based on the spurious correlation between lesion areas and background areas， which leads to a decline in the model's generalization ability and robustness. To address this issue， the paper proposed a quad-tree token splitting-merging debiasing method called QSMFormer， which was specifically designed for computer-aided diagnosis. Its main contributions were as follows： Firstly， designing a spatial-semantic coupled heterogeneity evaluation module. By jointly modeling the differences in regional significance distribution and semantic consistency， the complexity of the region could be adaptively evaluated. a quadtree-based recursive token splitting method driven by region heterogeneity is constructed， which performed fine-grained partitioning on structurally complex regions to preserve critical lesion features. Finally， building a semantic- and saliency-aware region token merging module. By dynamically merging adjacent regions based on the similarity and significance between tokens， and using the Top-K strategy to select key Tokens， redundant background information was reduced and the consistency of regional semantic expression was enhanced. This method was verified on the Kaggle Chest X-ray dataset and the Sharmin 9+1 dataset （a ten-class ophthalmology dataset）. Specifically， on the Kaggle Chest X-ray dataset， an accuracy rate of 98.51% and an F1 score of 97.02% were achieved； on the Sharmin 9+1 dataset， the accuracy rate and F1 score reached 98.26% and 92.60% respectively. The experimental results show that QSMFormer improves classification performance and enhances stability under background perturbations， while reducing the influence of non-subject background information to some extent.关键词：medical image classification;potential spurious correlation mitigation;vision transformer;quadtree token splitting;region token merging1|0|0更新时间：2026-08-26
- 摘要：Current referring expression segmentation methods had key limitations. They showed spatial redundancy in visual features. Language descriptions had referential ambiguity. Lexical-level semantics were often ignored. Fusion modules failed to fully exploit text information. This paper proposed fine‑grained referring expression segmentation via bidirectional feature modulation and text-guided progressive fusion （BFMPF）. It used a pre-trained BEIT3 as an image-text encoder. It designed a bidirectional feature modulation module （BFMM）. BFMM used global features of one modality to modulate local features of the other. This decoupled strategy suppressed noise before cross-modal interaction. Specifically， text-guided visual modulation filtered visual spatial redundancy. Vision-guided text modulation reduced the weights of irrelevant words. Then a text-guided progressive fusion module （TPFM） was introduced. TPFM embedded convolutions to compensate for the transformer's local modeling. It re-injected text semantics at multiple abstraction levels. It used multi-layer cross-modal attention to progressively fuse text features with high-level image features. A multilayer perceptron finally produced the segmentation mask. Experiments are conducted on three classic datasets. On RefCOCO， the method achieves 77.12%， 79.46% and 73.45% on val， testA and testB. On RefCOCO+， it reaches 69.72%， 72.86% and 62.42% on val， testA and testB. On RefCOCOg， it reaches 69.12% on val-u and 69.58% on test-u. The proposed BFMPF has strong generalization. It enables fine-grained cross-modal alignment. It performs robust referring expression segmentation.关键词：referring expression segmentation;cross-modal alignment;bidirectional feature modulation;text-guided progressive fusion;fine-grained understanding2|0|0更新时间：2026-08-26
- 摘要：To address the limited accuracy of 6D object pose estimation under complex backgrounds and weak-texture conditions， an uncertainty-aware 6D object pose estimation framework was proposed. Firstly， a multi-scale receptive field enhancement module was introduced into the encoder–decoder architecture at the feature extraction stage. By integrating contextual information with different convolutional scales and dilation rates， the network was able to effectively improve the multi-scale feature representation capability for weak-texture objects. Secondly， a pixel-wise confidence-aware voting mechanism was designed to model the reliability of the predicted dense keypoint unit vector fields， and the contribution of each pixel was adaptively weighted during the voting process， thereby enhancing the robustness and accuracy of keypoint localization. Finally， the uncertainty of keypoint localization was explicitly incorporated into the pose estimation stage by constructing an uncertainty-aware PnP framework， which effectively reduced the influence of noisy keypoints on pose estimation results. Experimental results on the LineMOD， Occlusion LineMOD， and YCB-Video datasets demonstrate that， in comparison with several representative methods such as BB8， DenseFusion， and PoseCNN， the proposed method achieves significant improvements in pose estimation accuracy， with an average increase of approximately 7% in the Average Distance （ADD） metric， and exhibits strong robustness under complex background and weak-texture scenarios. These results demonstrate the effectiveness， robustness， and adaptability of the proposed method under complex background and weak-texture conditions.关键词：6D object pose estimation;keypoint localization;pixel-wise confidence-aware voting;uncertainty-aware modeling;PnP pose estimation1|0|0更新时间：2026-08-26
Information Sciences
- Address：No.3888 Dong Nanhu Road, Changchun, Jilin, China Postal code：130033
- Tel：0431-86176855 Email：gxjmgc@ciomp.ac.cn
- Technical support is provided by Beijing Founder electronics co., LTD 吉ICP备11002662号-17
京公网安备11010802024621 - It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
- Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.
0