最新刊期
卷 33 ， 期 4 ， 2025
- “最新研究报道，科研团队开发出三种规格的光纤布拉格光栅穿戴式多关节指套，可屏蔽温度影响，精确监测手指各关节运动角度灵敏度和运动情况，适用于不同人群。”摘要：A wearable multi-joint fingerstall with three different specifications based on Fiber Bragg Grating （FBG） was designed. It could shield the temperature effects on FBG. With good fixing coupling and anti-creep ability， a multi-joint finger motion model was established. Calibration experiments for joint angle measurements were performed by wearing the three types of fingerstalls on a finger model and the fingers of two volunteers. A comparative experiment was conducted using an Inertial Measurement Unit （IMU） sensor and the FBG sensing unit to monitor the angular velocity of the index， middle， and ring fingers. The fingerstall was used to monitor three different hand gestures， and a Backpropagation （BP） neural network was employed for predictive analysis. The sensitivity error of the index finger in comparison with theoretical analysis is less than 7.7%. The average sensitivity for the corresponding joints during the "forward stroke" and "reverse stroke" in the finger model is greater than 5.3 pm/°， and for two Volunteers， the average sensitivity for the corresponding joints during the "forward stroke" and "reverse stroke" is greater than 3.8 pm/°. Calibration experiments show that the fingerstall can accurately monitor the motion sensitivity of different finger joints and is applicable to various populations， the finger sleeve device can shield it from the influence of human body temperature. Bland-Altman analysis of the comparative experiment data indicates consistency betweenthe results， demonstrating that the FBG sensing unit can monitor the motion of the index， middle， and ring fingers effectively. Using the fingerstall to monitor gestures and applying the BP neural network for prediction analysis， the Mean Squared Errors （MSE） for two Volunteers are 0.000 04 and 0.000 002， respectively. Comprehensive experimental results show that the FBG multi-joint fingerstall can precisely monitor the joint motion angle sensitivity and movement， and the BP neural network can be used for prediction.关键词：optical fiber Bragg grating;wearable multi-joint fingerstall;joint motion sensitivity;gesture monitoring2|0|0更新时间：2025-03-31
- “在光电精确制导领域，专家建立了基于数据驱动的多模复合光电制导导引头数据层融合方法，大幅提升了制导精度和抗干扰能力，满足实际应用需求。”摘要：In order to improve the guidance accuracy and anti-jamming ability of optoelectronic precision guidance guides， a data-driven fusion method based on multimode composite optoelectronic guidance guides was established. First， the physical structure of the laser/millimeter-wave radar/infrared multimode composite precision guidance was analyzed. Then， the Lagrangian interpolation and coordinate transformation method was used to realize the multimode guidance data alignment. Finally， a multimode composite guidance information fusion method based on a novel Transformer model was proposed to realize the data fusion and interference identification of multimode composite guidance guides. In order to verify the effectiveness of the proposed method， an experimental system of the guide head hang-flying was constructed. The experimental results show that the multimode composite guidance information fusion method can realize the fusion of multimode guidance in the data layer， and compared with the mutual covariance fusion method， the root-mean-square error is reduced by more than 22%， and the probability of capture when interfered with is increased from 71% to 91%， which greatly improves the accuracy of multimode guidance and the anti-interference ability， and can satisfy the practical needs of the multimode composite optoelectronic precision guidance application.关键词：multimode composite guidance;deep learning;data fusion;target tracking2|0|0更新时间：2025-03-31
- “在小行星地形探测领域，研究人员提出了一种新型接收光学系统，成功解决了激光雷达在极端环境下的应用难题，为小行星探测提供了新方案。”摘要：To ensure that LiDAR met the requirements of temperature range and vacuum environment in the detection of extraterrestrial terrain and to solve technical problems such as detection blind spots， small detection range， and difficult accurate alignment of the illumination array and pixel array， this paper proposed a receiving optical system based on a transversal mirror， multi-curvature spherical lens， and flat glass. The common optical axis of the transmitting optical system and the receiving optical system was realized by placing a reflector with 80% reflectivity at 45°. The blind detection area is avoided by using the characteristics of different radii of curvature in different aperture regions of the multi-curvature spherical lens， and an ultra-wide detection range is realized. Through the optimization of the lens material and lens distance tolerance， the thermal insensitive characteristics of the optical system and the precise control of distortion and focal length are realized. The test results show that the receiving optical system can achieve a field of view of Ø1.91°， a super-wide working distance of 30~300 m， a working temperature range of 0~40 ℃， and can adapt to two working environments of vacuum and atmospheric pressure， and has been successfully applied to the asteroid terrain detection lidar.关键词：laser radar;receiving optical system;optical design;multi-curvature spherical lens2|0|0更新时间：2025-03-31
Modern Applied Optics
- “在数字化位移测量领域，研究者设计了基于球面光投影成像的直线位移测量系统，实现了高分辨力直线位移测量，为减小系统体积提供技术依据。”摘要：Image-based displacement measurement using image recognition algorithms is a novel digital displacement measurement technology that offers high robustness and fault tolerance. To reduce the volume of the measurement system while achieving high-resolution linear displacement measurement， a linear displacement measurement system based on spherical light projection imaging was designed. Firstly， the magnification effect of spherical light projection was utilized to design the spherical light projection imaging magnification optical path， and the pixel grayscale values captured by the image sensor were compensated based on the optical wave plane model. Subsequently， to achieve absolute position measurement and encoding element discrimination on a single code track， a pseudo-random sequence encoding method and a decoding method based on width threshold were designed. Then， to achieve sub-pixel linear displacement measurement and enhance system accuracy， a linear displacement subdivision algorithm and a multi-centroid fusion algorithm were designed. Finally， an experimental setup was designed to experimentally validate the proposed method.The measurement optical path height of the designed experimental setup is only 5 mm. The experimental results indicate that the proposed method can achieve a theoretical resolution of 5 nm within a 250 mm range， with an error range of -3.3 μm to 3.6 μm， and after compensation， the error range is reduced to -0.6 μm to 1 μm. The research conducted can provide a technical basis for high-resolution linear displacement measurement studies while reducing the system volume and achieving high-resolution linear displacement measurement.关键词：spherical light projection imaging;image-based linear displacement measurement;high robustness;high fault tolerance;high resolution2|0|0更新时间：2025-03-31
- “在视轴稳定技术领域，专家分析了IRU系统多源异质扰动抑制效果，为提高系统鲁棒性提供解决方案。”摘要：Inertial Reference Unit -based visual axis stabilization is the main technical means to overcome the angular vibration interference of moving carriers and realize high-precision aiming. The IRU system is faced with multiple sources of heterogeneous disturbances， such as mechanical resonance， sensing noise， actuator disturbance， base disturbance， etc.， and the stability of the system is affected by the different disturbances through the additive， multiplicative， or implicit way. Disturbance observer can improve the robustness of the system without affecting the system stability. In this paper， on the basis of analyzing the sources and characteristics of internal and external disturbances in IRU systems， we compared and analyzed the suppression of multi-source heterogeneous disturbances in IRU systems by three structures： the classical DOB， the noise observer， and the noise reduction disturbance observer. The transfer characteristics of the classical DOB， NOB， and NRDOB to the input command signal and the multi-source heterogeneous disturbance of the system were deduced， the design requirements of the disturbance estimator were derived， and the simulation and experimental platforms were designed to compare and analyze the sinusoidal tracking capability and the suppression capability of the three structures to the multi-source heterogeneous disturbances.The results show that the disturbance suppression ability and tracking accuracy of both DOB and NRDOB structures are better than that of the NOB structure， and the tracking error of the two structures for the 15 Hz， 0.1 mrad sinusoidal input signal under the effect of 20 Hz， 0.2 mrad base disturbance is 19.2 μrad， which is reduced by 39.43% compared with that of the NOB structure.关键词：photoelectric tracking and targeting system;inertial reference unit;disturbance observer;disturbance suppression2|0|0更新时间：2025-03-31
- “在回转体零件误差测量领域，专家利用工业CT图像，建立了三维点云分割和几何模型优化体系，为提高测量精度提供新方案。”摘要：To address the issue of internal error measurement for rotary components， this paper exploited the advantages of industrial CT， which is non-destructive and capable of detecting the internal structure， and achieved the internal error measurement of rotary components through industrial CT images. Firstly， the axis parameters of the rotary body were calculated， and these parameters and Rodrigues’s formula were incorporated into the region growing method to enhance the accuracy of 3D point cloud segmentation. Subsequently， the objective function of the geometric model was established， and initial parameters of the geometric model were obtained through least squares linear fitting. Further， the optimal parameters of the geometric model were solved using Levenberg-Marquardt nonlinear fitting. Finally， the geometric model was utilized to compute the shape and position errors of the parts. Experimental results demonstrate the superiority of the proposed point cloud segmentation method in terms of accuracy when compared to other methods. Additionally， the surface reconstruction method exhibits better stability than the RANSAC algorithm. Compared with commercial VG software， the accuracy of shape and position error calculation is higher， with an increase of 8.5% in cylindricity， 21.1% in coaxiality， and 4.1% in verticality. Moreover， the cone tolerance calculation yields results within the design specifications， which satisfying the requirements of error measurement engineering.关键词：computed tomography;point clouds segmentation;surface reconstruction;error measurement2|0|0更新时间：2025-03-31
- “在磁场测量领域，研究人员采用标量抽运-检测型原子磁力仪，有效降低三轴矢量原子磁力仪的测量盲区，使其适用于可驻停或缓慢运动的平台。实验结果表明，该矢量原子磁力仪具有动态连续测量、测量范围宽和测量盲区小的技术特征。”摘要：In order to reduce the requirement of three-axis vector atomic magnetometers for the attitude control range of the motion platform， a scalar pump-probe atomic magnetometer with a smaller dead zone was selected for the vector atomic magnetometer based on the magnetic field rotation-modulation method. The improved vector atomic magnetometer had a smaller dead zone and had the potential to be used on a stationary or slow-moving platform. Firstly， the working principle of the vector atomic magnetometer based on the magnetic field rotation-modulation method was introduced. Secondly， the distribution of dead zones of the vector atomic magnetometers was analyzed when the scalar atomic magnetometer was selected as an Mx type optical pump magnetometer or a pump-probe type atomic magnetometer. Finally， the technical specifications of the three-axis atomic magnetometer with a small dead zone were validated near the geomagnetic field. When the three-axis vector atomic magnetometer measures the vector magnetic field near 40 000 nT， the sensitivity of measuring total fields is less than 1 nT/Hz1/2（@0.1 Hz）， the sensitivity of measuring angles is less than 0.1°/Hz1/2（@0.1 Hz）， and the proportion of dead zone in spatial solid angle is less than 12%. The vector magnetometer described in this article has the technical characteristics of dynamic continuous measurement， wide range and small dead zone.关键词：Vector atomic magnetometer;pump-probe;magnetic field rotation-modulation method;dead zone2|0|0更新时间：2025-03-31
Micro/Nano Technology and Fine Mechanics
- “在无人机影像点云和三维激光点云配准领域，提出了一种新的算法，有效提高了点云数据精度和三维地形重建的精度，满足了国家规范要求。”摘要：Addressing the issues of significant density variations， uneven spatial distribution， and indistinct features in UAV image point cloud and 3D laser point cloud within vegetation-covered and multi-slope regions， this study introduced a novel algorithm that combined sampling consistency initial alignment （SAC-IA） and iterative closest point （ICP） methods to enhance key slope features in point cloud data. Initially， preprocessing was performed on both source point cloud datasets， followed by the application of the random sample consensus （RANSAC） algorithm to fit the post-preprocessing point cloud regions with weak features， thereby enhancing the surface features and establishing multiple key slopes. Subsequently， the SAC-IA and ICP algorithms were integrated to register the two-source point cloud， subsequently eliminating redundancies and overlapping points to achieve fusion. Ultimately， the asymptotic encrypted irregular triangulation network （PTIN） filtering algorithm was employed to extract ground points from the fused point cloud， while the inverse distance weighting （IDW） algorithm was utilized for 3D terrain reconstruction， resulting in the generation of a digital elevation model （DEM）. Validation using actual measurement data demonstrates that， compared to the traditional SAC-IA and ICP combined algorithm， the point cloud data accuracy represented by root mean square error value after registration of the algorithm in this paper is reduced by 3.325 m； The DEM point accuracy （represented by mean absolute error and root mean square error） reconstructed from the fused point cloud data decreased by 0.18 m and 0.14 m respectively. The DEM generated by this study's algorithm meets the national specification requirements for a 1：500 scale， and it more accurately reflects topographic details.关键词：Random Sample Consensus（RANSAC）;slope fitting;point cloud registration;Digital Elevation Model（DEM）;3D reconstruction2|0|0更新时间：2025-03-31
- “在结直肠癌病理图像半监督语义分割领域，专家提出了伪标签置信度调控方法，实现了高质量多类别分割，为结直肠癌病理图像分析提供算法支持。”摘要：In order to improve the under-utilization of low-confidence pseudo-labels， the need to optimize the accuracy of high-confidence pseudo-labels and the imbalance of pseudo-label categories in the semi-supervised semantic segmentation task of colorectal cancer pathological images， this paper proposed a pseudo-label confidence regulation method to achieve high-quality multi-class semi-supervised semantic segmentation of colorectal cancer pathological images. First， based on the semi-supervised semantic segmentation framework of the teacher-student model， we propose to embed class confidence regulation in the consistency regularization， and to enhance the certainty by removing the confusing classes in the low confidence pseudo-labels generated by the untrained teacher model， so as to increase the contribution rate of the low confidence pseudo-labels. Secondly， an operation paradigm of first screening and then refining the pseudo-tags generated by the teacher model after training is proposed. By refining the filtered high-confidence pseudo-tags based on conditional random fields， the problems of boundary ambiguity and lack of semantic information in high-confidence pseudo-tags are improved. Finally， in order to alleviate the category imbalance in pseudo-label data， an adaptive random cascade strong data enhancement method based on the classification number of pseudo-label is designed. Through the experimental verification of the self-built colorectal cancer pathological image dataset and the published multi-class pathological image dataset， the proposed method achieves 74.09% average segmentation accuracy of four categories of colorectal cancer pathological images， which is 6.43% higher than that of the benchmark network， and provides powerful algorithm support for semi-supervised semantic segmentation of colorectal cancer pathological images.关键词：colorectal cancer pathological images;semi-supervised semantic segmentation;teacher-student model;consistency regularization;conditional random fields;data augmentation2|0|0更新时间：2025-03-31
- “在遥感图像道路提取领域，专家基于SwinUnet网络提出了多尺度上下文感知网络，有效提升了提取精度和性能。”摘要：To address the issues of local feature loss and low extraction accuracy faced by deep neural networks in remote sensing image road extraction， a multi-scale context-aware network was proposed based on the SwinUnet network for remote sensing image road extraction. Firstly， a branch with a context aggregation module was designed in the encoder to enhance the extraction of contextual information and alleviate the problem of semantic ambiguity caused by occlusion. Secondly， to solve the problem of semantic information mismatch between the encoder and decoder and to improve the model's ability to extract spatial information， a spatial feature extraction module was introduced in the skip connections， replacing the direct copying of encoder features in SwinUnet. Finally， a feature compression module was designed in the down-sampling stage to reduce information loss in the encoder and enhance the network's segmentation capability. The test results on the Massachusetts road dataset show that this method achieved F1， IoU， Pr， and Re scores of 80.91%， 69.40%， 78.03%， and 65.20%， respectively. In comparison with mainstream methods such as UNet and SwinUnet， the IoU improved by 4.45% and 2.72%， respectively， demonstrating that the proposed algorithm effectively improves the accuracy and performance of remote sensing image road extraction through global modeling， context enhancement， and information matching optimization.关键词：Remote sensing;road extraction;semantic segmentation;SwinUnet4|0|0更新时间：2025-03-31
- “在图像加密领域，专家提出了双超混沌系统结合压缩感知和Fibonacci变换的彩色图像加密算法，有效保护图像安全性和完整性。”摘要：In order to solve the problems in the existing chaotic image encryption algorithms， such as the weak chaotic property and low security of low-dimensional chaotic systems， the high transmission cost and large storage space of high-dimensional chaotic systems， as well as the large occupied space and high transmission cost of traditional measurement matrices in compressed sensing technology， this paper proposed a color image encryption algorithm combining dual hyperchaotic systems with compressed sensing and Fibonacci transformation. Firstly， the Arnold algorithm was optimized through the Lorenz hyperchaotic system to improve security. Secondly， the measurement matrix of compressed sensing was improved by using the 6D hyperchaotic system to reduce space resources and transmission costs. Thirdly， scrambling was performed using the 6D hyperchaotic system with a large key space and high security to enhance security. Finally， the Fibonacci Q matrix was used for block diffusion to enhance unpredictability. In addition， the 2D projection gradient embedding decryption algorithm was introduced in image reconstruction. Compared with the traditional reconstruction algorithm， it had higher security and computational efficiency. Experimental results show that the information entropy exceeds 7.999， which is closer to the ideal value of 8， the correlation coefficient is close to 0， the rate of pixel number change and the uniform change rate are close to the ideal values of 99.609 4% and 33.463 5%. While effectively protecting the security and integrity of the image， it has strong resistance to statistical attacks and differential attacks as well as better robustness.关键词：image encryption;hyperchaotic system;Compressed sensing;fibonacci Q-matrix4|0|0更新时间：2025-03-31
- “在Micro-LED芯片检测领域，提出了基于聚类思想的显微图像拼接方法，有效提升了拼接速度和准确率，具有实际工程应用价值。”摘要：Aiming at the problem of slow algorithmic stitching and low accuracy due to the existence of a large number of similar and neatly arranged LEDs in the existing microscopic images of Micro-LED chips， this paper proposed a clustering-based Micro-LED chip micro-image stitching method. Firstly， the OTSU method was used to preprocess the images， and then the adaptive area expansion phase correlation method was used to peak the overlapping regions of two adjacent images to obtain the offset of the preliminary alignment. Subsequently， the optimized alignment strategy using image chunking and the DHash algorithm was employed to obtain the optimized offset. Finally， density-based spatial clustering of applications with noise was introduced to automatically screen out and deal with the possible erroneous offsets. The experimental results show that the average time consumption of this paper's algorithm is 64 ms， which meets the demand of real-time stitching， the stitching correct rate reaches 99.4%， the alignment accuracy is controlled within 1 pixel， and the fusion processing can achieve a perfect stitching effect subjectively， which also provides a new solution to the overall misalignment problem caused by the wrong offsets. The algorithm in this paper effectively solves the key link of microscopic image stitching in wafer-level Micro-LED chip inspection and can be promoted and applied to other machine vision automation fields with repetitive features and high precision requirements， with high practical engineering application value.关键词：Image stitching;Micro-LED chip micrographs;phase correlation;DHash;clustering idea;defect detection2|0|0更新时间：2025-03-31
- “在视觉跟踪领域，研究者提出了基于知识蒸馏的Transformer跟踪器，有效提升了跟踪精度和速度。”摘要：To achieve high-precision and real-time tracking with limited computing resources， a transformer-based visual tracker via knowledge distillation was proposed. By introducing the image dynamic correction module， our tracker fused the search image of the current frame with the predicted image based on optical flow， which could effectively deal with challenges such as fast motion and motion blur. In order to reduce model complexity， the knowledge distillation learning strategy was adopted to compress the model. By introducing homoscedastic uncertainty into the loss function， loss weights of different subtasks could be learned through our network， thereby avoiding the cumbersome and difficult manual parameter tuning. Additionally， during training for the student network， a random blurring strategy was employed to enhance model robustness. Two tracking frameworks with different complexities， named KTransT-T and KTransT， were proposed and compared with 12 algorithms on 5 public datasets. Experimental results show that KTransT-T has significant advantages in precision and success rate， while KTransT has lower model complexity and competitive tracking performance. KTransT runs at a speed of up to 158 frames per second， which can meet the requirements of real-time tracking.关键词：computer vision;object tracking;transformer;knowledge distillation;homoscedastic uncertainty2|0|0更新时间：2025-03-31
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