最新刊期
卷 34 ， 期 3 ， 2026
- “在高精度光学加工领域，专家们建立了细束径离子束变去除函数补偿模型，通过控制驻留时间和进给速度，有效提高了加工收敛率，为亚纳米光学加工提供了稳定可靠的解决方案。”摘要：Deterministic ion-beam figuring of high-precision optical surfaces requires stable and accurately controllable removal functions. To improve in-process control of the ion-beam removal function and to establish a variable removal-function compensation model for fine-beam-diameter ion beams， a systematic framework was developed for investigating， compensating， and optimizing removal-function variability during machining. First， the practical instability of the removal function was analyzed theoretically. Subsequently， the governing factors were derived and classified into direct effects （thermal accumulation and energy distribution） and indirect effects （dwell time and lateral velocity）. Dynamic removal-function experiments were then performed to validate the influence of dwell time on the removal rate， and tests using two beam diameters confirmed the generality of the observed behavior. Finally， a compensation strategy and machining guidelines were proposed based on the effect of lateral velocity on the removal rate. The results demonstrate that regulating dwell time and lateral velocity can significantly enhance the convergence rate in fine-beam-diameter machining of high-precision optical surfaces. In sub-nanometer machining validation， a 0.332 nm RMS figure error was achieved on a 100 mm ULE flat mirror， satisfying practical requirements for stability， reliability， high precision， and controllability of fine-beam-diameter ion-beam processing.关键词：optical manufacturing;ion beam figuring;fine beam diameter;variable removal function;dwell time;nonlinear compensation18|12|0更新时间：2026-02-14
- “针对工业现场大粒径煤样研磨难题，专家构建了基于NIRS与XRF融合的PLS-AE-RR混合预测模型，为提升煤炭热值在线分析精度开辟新路径。”摘要：To address the enhanced matrix effects and intensified nonlinear spectral responses caused by the difficulty of grinding large-particle coal in industrial settings， this study proposes a hybrid PLS-AE-RR predictive model based on the fusion of near-infrared spectroscopy （NIRS） and X-ray fluorescence （XRF） spectra， aimed at improving the accuracy of on-line calorific-value analysis. The method implements a three-stage hybrid framework --linear baseline + nonlinear feature extraction + residual correction-- where partial least squares regression （PLS） first models the global linear relationship between the fused spectra and calorific value； an autoencoder （AE） then extracts low-dimensional nonlinear representations that PLS cannot capture； and finally ridge regression （RR） fits and corrects the nonlinear residuals. Experimental validation using 153 blended coal samples from power plants demonstrates breakthrough performance in calorific value prediction for large-particle-size coal. On the test set， determination coefficients （R²） for lignite and bituminous coal reached 0.974 and 0.938， respectively， with mean absolute errors of 0.233 MJ/kg and 0.216 MJ/kg. The proposed method significantly outperforms standalone PLS and alternative nonlinear correction models， confirming the generalization advantage of ridge regression in residual fitting. Consequently， this achievement provides a grinding-free， high-precision online analysis solution for raw coal calorific value in coal-fired power plants， offering critical technical support for refined fuel management and operational optimization.关键词：spectral analysis;NIRS spectra;XRF spectra;calorific value prediction;large-particle coal samples;nonlinear correction;autoencoder12|10|0更新时间：2026-02-14
- “针对多基线气体透射比测量系统存在的稳定性不足、光路对准精度难以保证、成本高昂以及难以实现紧凑型设计等问题，研究了一种能够在有限空间内实现多种基线稳定切换且接收机可共用的气体透射比测量系统。”摘要：To overcome the limitations of existing multi-baseline gas transmittance meters， including inadequate stability， difficulty in maintaining optical-path alignment accuracy， high cost， and limited compactness， a gas transmittance meter is proposed that enables stable multi-baseline switching within a confined space while sharing a single receiver. Based on time-spectral multiplexing， an integrated multi-wavelength filled-aperture laser is combined with optical components， including beam splitters and broadband reflectors， to construct folded optical paths with multiple baselines. Distinct wavelengths are gated in separate time slots to realize stable baseline switching. The multi-wavelength signals are spatially combined into a single optical path and detected by a single receiver for gas transmittance calculation. Experimental results demonstrate that， under comprehensive conditions incorporating errors from both the transceiver system and local components， the proposed multi-baseline system reduces gas-transmittance indication errors from 1.99% and 25.58% in a single-baseline system to within 0.38% and 7.23% in low- and high-transmittance tests， respectively. These results indicate effective suppression of system attenuation and local interference， providing a practical solution for accurate monitoring under complex meteorological conditions.关键词：photoelectric measurement;multiple baselines;active detection;gas transmittance17|11|0更新时间：2026-02-14
- “介绍了其在平面度误差评定领域的研究进展，专家提出了一种PSO-SQP混合算法，为解决传统智能优化算法计算精度不足、收敛速度慢等问题提供了高效解决方案。”摘要：To overcome the limited computational accuracy and slow convergence of conventional intelligent optimization algorithms in flatness error evaluation， a high-precision and high-efficiency evaluation method is developed to satisfy the stringent algorithmic requirements of a self-developed 1 200 mm-aperture non-contact flatness measurement instrument. A hybrid PSO-SQP algorithm is proposed， in which Sequential Quadratic Programming （SQP） is employed as the primary framework and Particle Swarm Optimization （PSO） is incorporated as an auxiliary global search strategy. The global exploration capability of PSO is leveraged to perform a coarse search and rapidly identify a near-global optimum， which is used as a high-quality initial point for SQP. An adaptive step-size strategy is introduced in the refinement stage to replace the conventional fixed step size， enabling rapid and stable convergence during local optimization. Experimental results indicate that the proposed PSO-SQP algorithm exhibits strong robustness to initial-point deviations， sampling scale variations， and measurement noise. Relative to a high-precision coordinate measuring machine （CMM）， the deviation of the evaluation results is below 7 nm. Engineering applicability is further validated through the evaluation of a 280 mm-diameter flat mirror in a practical measurement scenario， where the obtained flatness is consistent with the specified surface form accuracy. Overall， the proposed algorithm provides high computational accuracy， fast convergence， and strong robustness， offering a reliable and efficient solution for flatness error evaluation in precision manufacturing.关键词：precision measurement;flatness error;Sequential Quadratic Programming;particle swarm optimization;Non-contact flatness measuring instrument91|119|0更新时间：2026-02-14
- “在月球极区着陆探测领域，专家们建立了基于加权最小二乘法的位姿误差模型，为解决陨坑视觉导航中位姿估计误差传递机理不明确、工程实用化参数缺失等关键问题提供了有效解决方案，为月球极区着陆百米级导航精度要求提供了理论基础。”摘要：High-precision positioning is required for lunar polar landing， and crater-based visual navigation can be used to estimate the lander pose. However， this approach is limited by an insufficiently characterized mechanism for pose estimation error propagation and by the lack of practical engineering parameters. To address these issues， pose error modeling for crater visual navigation is investigated. A pose error model based on weighted least squares is proposed to quantify pose errors under the combined effects of multiple factors. First， traceability analysis and baseline modeling of pose errors are performed according to the pose estimation principle of crater visual navigation. Second， a pose error model is formulated by minimizing the reprojection error. Finally， weighted least squares is applied to obtain quantitative pose error estimates. Using DEM and DOM data from the lunar polar region， crater visual navigation for a lunar lander is simulated， and Monte Carlo-based experiments are conducted to evaluate the effects of different error sources and feature parameters on pose accuracy. When the pixel detection error is 5 pixels and 15 craters are used， the estimated translation error is below 95 m. The developed model enables quantitative pose error estimation under multiple contributing factors， and the proposed engineering parameter scheme satisfies the 100 m-level navigation accuracy requirement for lunar polar landing. This work provides a theoretical basis for the design of crater-based navigation schemes.关键词：visual navigation;lunar polar landing;pose estimation;error model6|4|0更新时间：2026-02-14
Modern Applied Optics
- “专家设计了螺旋驱动式胶囊机器人内窥镜，解决了传统无线胶囊内窥镜缺乏主动运动能力的问题。”摘要：Wireless capsule endoscopes provide a noninvasive approach for intestinal examination but remain limited by the absence of active mobility. To address this constraint， a capsule robotic endoscope based on an active motion platform was developed. A helical drive mechanism was designed to enable versatile locomotion through control of the rotational speed and direction of the helical drive unit. To accommodate this mechanical configuration， dual-ended receiving coils were designed， and the electrical parameters of coils with different topological structures were derived through theoretical analysis and mathematical modeling. The performance of the proposed coil topologies was subsequently assessed by simulation and validated through experimental testing. The results demonstrate that a rectifier-followed-by-series topology provides high stability for the dual-ended receiving coil， achieving a power transfer efficiency of 2.75% at 981 mW induced power. Pipeline experiments further confirm that the wireless power transfer system satisfies the power demands required for capsule endoscope robot locomotion and control.关键词：capsule robots;endoscope;wireless power transfer;Helical drive motion mechanism;Dual-end receiving coil;wireless power transfer efficiency81|122|0更新时间：2026-02-14
- “介绍了其在精密定位系统领域的研究进展，专家建立了LRGPI模型体系，为解决压电快反镜迟滞问题提供了有效解决方案。”摘要：The inherent hysteresis of piezoelectric fast steering mirrors （PFSMs） severely limits control accuracy in precision positioning systems. To address this limitation， the performance of common envelope functions is systematically compared in terms of computational complexity， inversion requirements， and error sources. Based on a comprehensive evaluation， an asymmetric linear envelope function is selected， and a rate-dependent generalized Prandtl-Ishlinskii model with an asymmetric linear envelope function （LRGPI） is developed. To mitigate rate-dependent hysteresis， a derivative term is incorporated to extend the model's applicable frequency range. The inverse LRGPI model is subsequently formulated and implemented as a feedforward compensator to evaluate hysteresis suppression. Furthermore， a composite control strategy integrating inverse-model feedforward compensation is designed to attenuate external disturbances. Simulation results indicate that， relative to the inverse models of the tanh envelope-based rate-dependent generalized Prandtl–Ishlinskii model （TRGPI） and the cubic envelope-based rate-dependent generalized Prandtl–Ishlinskii model （CRGPI）， the LRGPI inverse-model feedforward increases the hysteresis-compensation bandwidth by 5.78% and 28.69%， respectively. Comparative experiments further demonstrate that the RMSE achieved by LRGPI inverse-model feedforward control is reduced by 62.7%， 23.2%， and 26.4% compared with the PI inverse model， TRGPI， and CRGPI inverse models， respectively. These results demonstrate the superior effectiveness and robustness of the proposed LRGPI model for compensating PFSM hysteresis.关键词：piezoelectric fast steering mirror;Hysteresis characteristic;rate-dependence;asymmetric envelope function;generalized prandtl-ishlinskii model;inverse compensation;composite control80|128|0更新时间：2026-02-14
- “介绍了其在光电跟踪领域的研究进展，专家建立了望远镜跟踪架的运动学模型，探索了运动学求解方法，为高精度大范围跟踪任务提供解决方案。”摘要：To enable electro-optical tracking with high dynamic response， high precision， and wide operating range， a telescope tracking mount based on a coaxial spherical parallel mechanism （CSPM） is investigated. First， forward and inverse kinematic models are formulated using geometric vector relations and screw theory. A solution strategy is then developed in which a trust-region iterative scheme， combined with assembly-mode constraints， is applied to determine a unique kinematically feasible solution. Next， the Jacobian-based condition number is evaluated and， together with inter-limb angle constraints， is used to characterize a practical workspace free of collisions and degree-of-freedom deficiency. The results indicate that the mount provides an elevation （pitch） range of 48.4°-131.6°， while the azimuth is unbounded. Finally， a trajectory-planning method is proposed that integrates unit-quaternion SLERP with cubic Bézier velocity smoothing， thereby ensuring continuous and smooth transitions between attitudes. Experimental results show that the proposed kinematic solution method achieves an angular error on the order of 10-11 deg， demonstrating strong consistency between the kinematic model and the numerical solver. In multi-key-posture pointing experiments， within the prescribed planning time， both attitude and angular velocity remain continuous and smoothly convergent， and the entire trajectory is free of singularities， further confirming the excellent structural characteristics and motion capability of the CSPM-based telescope tracking mount.关键词：parallel mechanism;telescope mount;Kinematics;workspace;trajectory planning10|4|0更新时间：2026-02-14
Micro/Nano Technology and Fine Mechanics
- “针对天基短波红外图像中弱小目标检测难题，专家提出融合安德森加速的自正则化加权稀疏模型与相对局部对比度的改进检测方法，显著降低计算复杂度，实现多尺度目标检测，为复杂背景下遥感目标检测提供可靠方案。”摘要：To address the challenge of detecting dim and small targets in space-based short-wave infrared （SWIR） imagery-where targets are readily obscured by cloud cover and ground clutter under low signal-to-clutter ratio （SCR） conditions-an enhanced detection algorithm is proposed that integrates Anderson-accelerated Self-Regularized Weighted Sparse （SRWS） modeling with the Relative Local Contrast Measure （RLCM）. Computational complexity in background estimation is substantially reduced through the incorporation of Anderson acceleration， while multi-scale target detection is achieved via background residual maps combined with RLCM. Experiments conducted on 289 SWIR images spanning seven representative scenarios demonstrate consistently strong performance in complex backgrounds， with the AUC reaching 0.950 and remaining no lower than 0.842 under the most challenging conditions. The signal-to-clutter ratio gain （SCRG） is significantly improved relative to conventional methods， including IPI and LCM. Overall， detection accuracy and robustness for dim and small targets in space-based SWIR remote sensing are effectively enhanced， providing a reliable solution for target detection in complex background environments.关键词：short-wave infrared;weak target detection;local contrast;background estimation7|4|0更新时间：2026-02-14
- “介绍了其在遥感成像领域的研究进展，相关专家提出了一种面向TDI CCD噪声建模的物理引导深度神经网络，为解决微光环境下TDI CCD噪声分布表征难题提供了有效解决方案。”摘要：Time-Delay Integration CCDs （TDI CCDs） are widely used in remote-sensing imaging. However， complex noise sources-including dark current， reset noise， and quantization noise-hinder accurate characterization of the signal-independent noise distribution of real sensors under low-light conditions. To address this challenge， a physics-guided deep neural network for TDI CCD noise modeling （PDNN） is proposed. Signal-independent noise is learned from dark-frame images and combined with signal-dependent noise modeled by a Poisson distribution， enabling accurate representation of the TDI CCD noise distribution in low-light scenes. First， a TDI CCD Noise Decoupling （TND） module decomposes dark-frame images into pixel-level noise with spatial independence. Next， a Gain and Multistage Adaptive （GMA） module， together with 1×1 convolutional layers in the TDI CCD Noise Modeling （TNM） backbone， maps the initial noise into a distribution space that closely matches the true noise level while preserving pixel-wise independence. Finally， a Task Balanced Loss （TBL） dynamically adjusts weighting factors to maintain training equilibrium， further improving performance. On a self-constructed dataset， the proposed method achieves an average Kullback-Leibler divergence （AKLD） of 0.106 9， demonstrating substantial improvements over existing approaches. Moreover， PSNR and SSIM obtained from models trained with synthetic noisy images closely approximate those achieved with real data. Experimental results indicate that PDNN effectively characterizes the low-light noise distribution of TDI CCDs， providing practical value for enhancing the visual quality of low-light remote-sensing imagery.关键词：TDI CCD;physics-guided;neural network;noise decoupling;task balanced loss11|4|0更新时间：2026-02-14
- “水下鱼类监测领域迎来新突破，相关专家基于YOLOv8n框架构建了FDN-YOLO模型，通过创新模块与损失函数设计，有效提升检测性能，为海洋生态保护与资源管理提供有力技术支撑。”摘要：Efficient monitoring of underwater fish species is essential for marine ecosystem conservation， biodiversity assessment， and the sustainable management of aquatic resources. To address reductions in detection robustness and efficiency under complex underwater conditions， a dynamic feature enhancement model， termed Fish Detection Network YOLO （FDN-YOLO）， is proposed based on the YOLOv8n framework. First， a Multi-scale Deformable Receptive Field （MDRF） module is incorporated to adaptively regulate the effective receptive field， thereby improving backbone representations of fish targets with diverse shapes and scales. Second， a lightweight down-sampling module， Lite Space-to-Depth Depthwise Separable （Lite SPD-DS）， is designed to preserve fine-grained spatial cues during subsampling while maintaining low computational cost. Third， an Adaptive IoU-aware Varifocal Loss （AIVF Loss） is introduced by integrating adaptive IoU weighting with Varifocal Loss to strengthen the learning of high-quality localization samples and mitigate training bias caused by class and sample imbalance. Experiments on the TF-DET dataset show that FDN-YOLO increases mAP50 and mAP50∶95 by 2.8% and 2.1%， respectively， while reducing parameters and computational complexity by 13.3% and 16.0%. Additional comparative and generalization experiments further confirm that FDN-YOLO achieves a favorable trade-off among accuracy， efficiency， and robustness， highlighting its potential for ecological monitoring and data-driven marine resource management.关键词：underwater detection;dynamic enhancement;deep learning;convolutional neural network;YOLOv8n84|145|0更新时间：2026-02-14
- “相关专家针对大面积破损图像修复难题，提出了一种基于结构先验与多尺度残差聚合的修复方法，有效解决了纹理模糊、结构扭曲与边界伪影等问题，为图像修复领域带来了新的突破。”摘要：To address texture blurring， structural distortion， and boundary artifacts in the restoration of images with large-area damage （e.g.， faces， street scenes， and buildings）， a large-area damaged image restoration network， termed SPMRA-Net， is proposed. First， a Dynamic Adaptive Multi-scale （DAM） module is designed to enhance contextual modeling through dilated convolutions and a residual architecture. Subsequently， a bottleneck module consisting of a Transformer branch and a consistent semantic attention branch is constructed； global semantic information from the Transformer branch is fused with local texture information from the self-attention branch via a cross-attention mechanism， effectively suppressing structural distortion in severely corrupted regions. Moreover， to mitigate the semantic gap introduced by naive concatenation in conventional skip connections， an Adaptive Multi-scale Aggregation （AMSA） module is introduced to strengthen interactions between deep and shallow features and to preserve boundary continuity in restored images. Finally， a dual-discriminator framework is employed to assess consistency between restored results and original images， thereby improving the perceptual realism of generated outputs. Experiments demonstrate that， on the CelebA dataset， the proposed method improves PSNR by 3.06 dB and SSIM by 0.087， while reducing LPIPS by 0.078. Subjective evaluations on the Paris StreetView and Places2 datasets， together with the aforementioned objective metrics， consistently outperform comparative methods. These results indicate notable gains in both structural consistency and perceptual quality， validating the effectiveness of the proposed approach for large-area damaged image restoration.关键词：deep learning;image inpainting;generative adversarial network;Multi-scale residuals;Structural prior;transformer85|142|0更新时间：2026-02-14
- “专家提出基于自适应强跟踪滤波的手术器械定位与引导方法，显著提升光学手术导航系统的稳定性与可靠性，为高精度医疗导航提供理论基础。”摘要：Optical surgical navigation systems used in complex minimally invasive procedures are vulnerable to multisource disturbances， including surgeons’ physiological tremor， device vibration， and imaging noise. Such disturbances can induce localization jitter and trajectory discontinuities. To mitigate insufficient system robustness caused by the combined effects of disturbance-induced noise and system-model uncertainty， an adaptive strong-tracking-filtering-based method is proposed for surgical instrument localization and guidance， thereby enhancing the stability and reliability of optical surgical navigation. In the position domain， a strong tracking Kalman filter with error-statistics-based adaptive adjustment of the noise covariance matrices is employed to improve suppression of abrupt noise and drift. In the attitude domain， a quaternion unscented Kalman filter is adopted to avoid linearization errors while preserving quaternion normalization. To further enhance smoothing performance， moving-average preprocessing is applied to the position input to integrate low-frequency stability with high-frequency disturbance attenuation. Datasets from eight distinct scenarios were collected for experimental validation， and the proposed method was comparatively evaluated against existing approaches using seven performance metrics. The results indicate that the proposed method achieves superior overall performance. For attitude tracking， the angular root-mean-square error is reduced by 50%-79% relative to the reference methods. For position tracking， the overall trend residual is further reduced by 5%-20% compared with the standard Kalman filter， yielding an improved balance between high-frequency noise suppression and dynamic response stability. These results demonstrate significant gains in micro-disturbance suppression， attitude stability， and noise adaptability， providing a theoretical basis for high-precision medical navigation.关键词：optical surgical navigation;positioning of surgical instrument;Kalman filter;noise suppression8|4|0更新时间：2026-02-14
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