To obtain the certificate of airworthiness,it is essential to conduct a full-scale aircraft static test.During such test,accurate and comprehensive wing deformation measurement is crucial for assessing its strength,st...To obtain the certificate of airworthiness,it is essential to conduct a full-scale aircraft static test.During such test,accurate and comprehensive wing deformation measurement is crucial for assessing its strength,stiffness,and bearing capability.This paper proposes a novel and cost-effective videogrammetric method using multi-camera system to achieve the non-contact,highprecision,and 3D measurement of overall static deformation for the large-scale wing structure.To overcome the difficulties of making,carrying,and employing the large 2D or 3D target for calibrating the cameras with large field of view,a flexible stereo cameras calibration method combining 1D target and epipolar geometry is proposed.The global calibration method,aided by a total station,is employed to unify the 3D data obtained from various binocular subsystems.A series of static load tests using a 10-meter-long large-scale wing have been conducted to validate the proposed system and methods.Furthermore,the proposed method was applied to the practical wing deformation measurement of both wings with a wingspan of 33.6 m in the full-size civil aircraft static test.The overall 3D profile and displacement data of the tested wing under various loads can be accurately obtained.The maximum error of distance and displacement measurement is less than 4.5 mm within the measurement range of 35 m in all load cases.These results demonstrate that the proposed method achieves effective,high-accuracy,on-site,and visualized wing deformation measurement,making it a promising approach for full-scale aircraft wing static test.展开更多
In order to achieve a high precision in three-dimensional（3D） multi-camera measurement system, an efficient multi-cameracalibration method is proposed. A stitching method of large scalecalibration targets is deduced...In order to achieve a high precision in three-dimensional（3D） multi-camera measurement system, an efficient multi-cameracalibration method is proposed. A stitching method of large scalecalibration targets is deduced, and a fundamental of multi-cameracalibration based on the large scale calibration target is provided.To avoid the shortcomings of the method, the vector differencesof reprojection error with the presence of the constraint conditionof the constant rigid body transformation is modelled, and mini-mized by the Levenberg-Marquardt （LM） method. Results of thesimulation and observation data calibration experiment show thatthe accuracy of the system calibrated by the proposed methodreaches 2 mm when measuring distance section of 20 000 mmand scale section of 7 000 mm × 7 000 mm. Consequently, theproposed method of multi-camera calibration performs better thanthe fundamental in stability. This technique offers a more uniformerror distribution for measuring large scale space.展开更多
This study analyzes the function of different muscles during arm wrestling and proposes a method to analyze the optimal forearm angle for professional arm wrestlers.We built a professional arm-wrestling platform to me...This study analyzes the function of different muscles during arm wrestling and proposes a method to analyze the optimal forearm angle for professional arm wrestlers.We built a professional arm-wrestling platform to measure the shape and deformation of the skin at the biceps brachii of a volunteer in vivo during arm wrestling.We observed the banding phenomenon of arm skin strain during muscle contraction and developed a model to evaluate the moment provided by the biceps brachii.According to this model,the strain field of the area of interest on the skin was measured,and the forearm angles most favorable and unfavorable to the work of the biceps brachii were analyzed.This study demonstrates the considerable potential of applying DIC and its extension method to the in vivo measurement of human skin and facilitates the use of the in vivo measurement of skin deformation in various sports in the future.展开更多
A new calibration algorithm for multi-camera systems using 1D calibration objects is proposed. The algorithm inte- grates the rank-4 factorization with Zhang (2004)'s method. The intrinsic parameters as well as th...A new calibration algorithm for multi-camera systems using 1D calibration objects is proposed. The algorithm inte- grates the rank-4 factorization with Zhang (2004)'s method. The intrinsic parameters as well as the extrinsic parameters are re- covered by capturing with cameras the 1D object's rotations around a fixed point. The algorithm is based on factorization of the scaled measurement matrix, the projective depth of which is estimated in an analytical equation instead of a recursive form. For more than three points on a 1D object, the approach of our algorithm is to extend the scaled measurement matrix. The obtained parameters are finally refined through the maximum likelihood inference. Simulations and experiments with real images verify that the proposed technique achieves a good trade-off between the intrinsic and extrinsic camera parameters.展开更多
Accurate and efficient methods for identifying and tracking each animal in a group are needed to study complex behaviors and social interactions.Traditional tracking methods(e.g.,marking each animal with dye or surgic...Accurate and efficient methods for identifying and tracking each animal in a group are needed to study complex behaviors and social interactions.Traditional tracking methods(e.g.,marking each animal with dye or surgically implanting microchips)can be invasive and may have an impact on the social behavior being measured.To overcome these shortcomings,video-based methods for tracking unmarked animals,such as fruit flies and zebrafish,have been developed.However,tracking individual mice in a group remains a challenging problem because of their flexible body and complicated interaction patterns.In this study,we report the development of a multi-object tracker for mice that uses the Faster region-based convolutional neural network(R-CNN)deep learning algorithm with geometric transformations in combination with multi-camera/multi-image fusion technology.The system successfully tracked every individual in groups of unmarked mice and was applied to investigate chasing behavior.The proposed system constitutes a step forward in the noninvasive tracking of individual mice engaged in social behavior.展开更多
Visual tracking has been a popular task in computer vision in recent years,especially for long-term tracking.A novel object tracking framework is proposed in this paper.For surveillance cameras with overlapping areas,...Visual tracking has been a popular task in computer vision in recent years,especially for long-term tracking.A novel object tracking framework is proposed in this paper.For surveillance cameras with overlapping areas,the target area is divided into several regions corresponding to each camera,and a simple re-matching method is used by matching the colors according to the segmented parts.For surveillance cameras without overlapping areas,a time estimation model is employed for continuously tracking objects in different fields of view(FoVs).A demonstration system for collaborative tracking in real time situation is realized finally.The experimental results show that compared with current popular algorithms,the proposed approach has good effect in accuracy and computation time for the application of continuously tracking the pedestrians.展开更多
为减少因船舶偏离航道而造成的搁浅、碰撞航标或桥墩等水上交通事故,提出了一种基于多目相机自动识别航道的桥区航行异常船舶预警方法。基于YOLOv5(You Only Look Once version 5)目标检测算法,联动变、定焦相机识别并定位航标和船舶,...为减少因船舶偏离航道而造成的搁浅、碰撞航标或桥墩等水上交通事故,提出了一种基于多目相机自动识别航道的桥区航行异常船舶预警方法。基于YOLOv5(You Only Look Once version 5)目标检测算法,联动变、定焦相机识别并定位航标和船舶,跟踪并记录船舶航迹点,计算船舶的速度和航向并推算船位。提出了一种基于视频船舶航迹点的密度聚类识别航道两侧航标的方法,实现航道自适应可视化。基于船位推算识别并预警航行状态异常的船舶。实验结果表明:航标、船舶的检测正确率分别达84.8%、90.3%,相较单一相机检测模型,正确率分别提高了32.1%、5.5%;能够自适应可视化航道并识别、预警航行异常船舶。展开更多
为提高无人车障碍物检测跟踪的精度和稳定性,首先针对YOLO v5(You only look once version 5,YOLO v5)网络存在的语义信息和候选框信息丢失的问题,引入深度可分离空洞空间金字塔结构与目标框加权融合算法完成对网络的优化;其次针对单阶...为提高无人车障碍物检测跟踪的精度和稳定性,首先针对YOLO v5(You only look once version 5,YOLO v5)网络存在的语义信息和候选框信息丢失的问题,引入深度可分离空洞空间金字塔结构与目标框加权融合算法完成对网络的优化;其次针对单阶段障碍物点云聚类精度低的问题,设计一种考虑点云距离与外轮廓连续性的两阶段障碍物点云聚类方法并完成三维包围盒的建立;最后将注意力机制引入MobileNet使网络更加聚焦于目标对象特有的视觉特征,并综合利用视觉特征和三维点云信息共同构建关联性度量指标,提高匹配精度。利用KITTI数据集对构建的障碍物目标检测、跟踪与测速算法进行仿真测试,并搭建实车平台进行真实环境试验,验证所提算法的有效性和真实环境可迁移性。展开更多
This paper seeks to determine how the overlap of several infrared beams affects the tracked position of the user, depending on the angle of incidence of light, distance to the target, distance between sensors, and the...This paper seeks to determine how the overlap of several infrared beams affects the tracked position of the user, depending on the angle of incidence of light, distance to the target, distance between sensors, and the number of capture devices used. We also try to show that under ideal conditions using several Kinect sensors increases the precision of the data collected. The results obtained can be used in the design of telerehabilitation environments in which several RGB-D cameras are needed to improve precision or increase the tracking range. A numerical analysis of the results is included and comparisons are made with the results of other studies. Finally, we describe a system that implements intelligent methods for the rehabilitation of patients based on the results of the tests carried out.展开更多
针对当前视觉同时定位与建图(Simultaneous Localization and Mapping,SLAM)面对如纹理稀疏、光照变化强烈及图像模糊等挑战性场景时,普遍存在的前端特征跟踪鲁棒性不足的问题,提出了一个鲁棒的多相机定位系统,并对关键技术进行了深入...针对当前视觉同时定位与建图(Simultaneous Localization and Mapping,SLAM)面对如纹理稀疏、光照变化强烈及图像模糊等挑战性场景时,普遍存在的前端特征跟踪鲁棒性不足的问题,提出了一个鲁棒的多相机定位系统,并对关键技术进行了深入的研究与优化。该系统设计了一种优于主流方案的前端跟踪算法,通过融合惯性测量单元(Inertial Measurement Unit,IMU)测量数据实现特征点重投影预测,并对跟踪的灰度图进行动态校正,有效提升了复杂场景下特征跟踪的成功率和稳定性。此外,该系统利用多相机观测信息构建了具备3层自适应置信度加权算法的状态估计器,并将卷积神经网络运用于交叉回环检测,有效提高了回环检测的成功率与准确率。通过一系列公开数据集的实验,验证了该多相机定位算法在精度和稳定性方面已达到最先进技术水平,且在本地环境中的测试结果也证实了该系统在实际应用中的可行性与有效性。展开更多
同时定位与地图构建(simultaneous localization and mapping,SLAM)技术在无人化装备上有着广泛的应用,可实现室内或室外自主的定位建图任务。该文首先对视觉和激光SLAM基本框架进行介绍,详细阐述前端里程计、后端优化、回环检测以及地...同时定位与地图构建(simultaneous localization and mapping,SLAM)技术在无人化装备上有着广泛的应用,可实现室内或室外自主的定位建图任务。该文首先对视觉和激光SLAM基本框架进行介绍,详细阐述前端里程计、后端优化、回环检测以及地图构建这四个模块的作用以及所采用的算法;在这之后梳理归纳视觉/激光SLAM发展历程中的经典算法并分析其优缺点以及在此之后优秀的改进方案;此外,列举当前SLAM技术在生活中的典型应用场景,展示在自动驾驶、无人化装备等领域的重要作用;最后讨论SLAM系统当前的发展趋势和研究进展,以及在未来应用中需要考虑的挑战和问题,包括多类型传感器融合、与深度学习技术的融合以及跨学科合作的关键作用。通过对SLAM技术的全面分析和讨论,为进一步推动SLAM技术的发展和应用提供深刻的理论指导和实践参考。展开更多
文摘To obtain the certificate of airworthiness,it is essential to conduct a full-scale aircraft static test.During such test,accurate and comprehensive wing deformation measurement is crucial for assessing its strength,stiffness,and bearing capability.This paper proposes a novel and cost-effective videogrammetric method using multi-camera system to achieve the non-contact,highprecision,and 3D measurement of overall static deformation for the large-scale wing structure.To overcome the difficulties of making,carrying,and employing the large 2D or 3D target for calibrating the cameras with large field of view,a flexible stereo cameras calibration method combining 1D target and epipolar geometry is proposed.The global calibration method,aided by a total station,is employed to unify the 3D data obtained from various binocular subsystems.A series of static load tests using a 10-meter-long large-scale wing have been conducted to validate the proposed system and methods.Furthermore,the proposed method was applied to the practical wing deformation measurement of both wings with a wingspan of 33.6 m in the full-size civil aircraft static test.The overall 3D profile and displacement data of the tested wing under various loads can be accurately obtained.The maximum error of distance and displacement measurement is less than 4.5 mm within the measurement range of 35 m in all load cases.These results demonstrate that the proposed method achieves effective,high-accuracy,on-site,and visualized wing deformation measurement,making it a promising approach for full-scale aircraft wing static test.
基金supported by the National Natural Science Foundation of China(61473100)
文摘In order to achieve a high precision in three-dimensional（3D） multi-camera measurement system, an efficient multi-cameracalibration method is proposed. A stitching method of large scalecalibration targets is deduced, and a fundamental of multi-cameracalibration based on the large scale calibration target is provided.To avoid the shortcomings of the method, the vector differencesof reprojection error with the presence of the constraint conditionof the constant rigid body transformation is modelled, and mini-mized by the Levenberg-Marquardt （LM） method. Results of thesimulation and observation data calibration experiment show thatthe accuracy of the system calibrated by the proposed methodreaches 2 mm when measuring distance section of 20 000 mmand scale section of 7 000 mm × 7 000 mm. Consequently, theproposed method of multi-camera calibration performs better thanthe fundamental in stability. This technique offers a more uniformerror distribution for measuring large scale space.
基金This study was supported by the National Natural Science Foun-dation of China(NSFC)(No.11902074).
文摘This study analyzes the function of different muscles during arm wrestling and proposes a method to analyze the optimal forearm angle for professional arm wrestlers.We built a professional arm-wrestling platform to measure the shape and deformation of the skin at the biceps brachii of a volunteer in vivo during arm wrestling.We observed the banding phenomenon of arm skin strain during muscle contraction and developed a model to evaluate the moment provided by the biceps brachii.According to this model,the strain field of the area of interest on the skin was measured,and the forearm angles most favorable and unfavorable to the work of the biceps brachii were analyzed.This study demonstrates the considerable potential of applying DIC and its extension method to the in vivo measurement of human skin and facilitates the use of the in vivo measurement of skin deformation in various sports in the future.
基金the National Natural Science Foundation of China (No. 60675017) the National Basic Research Program of China (No. 2006CB303103)
文摘A new calibration algorithm for multi-camera systems using 1D calibration objects is proposed. The algorithm inte- grates the rank-4 factorization with Zhang (2004)'s method. The intrinsic parameters as well as the extrinsic parameters are re- covered by capturing with cameras the 1D object's rotations around a fixed point. The algorithm is based on factorization of the scaled measurement matrix, the projective depth of which is estimated in an analytical equation instead of a recursive form. For more than three points on a 1D object, the approach of our algorithm is to extend the scaled measurement matrix. The obtained parameters are finally refined through the maximum likelihood inference. Simulations and experiments with real images verify that the proposed technique achieves a good trade-off between the intrinsic and extrinsic camera parameters.
基金supported by grants from the National Key R&D Program of China(2017YFA0105201)the National Natural Science Foundation of China(81925011,92149304,31900698,32170954,and 32100763+2 种基金the Key-Area Research and Development Program of Guangdong Province(2019B030335001)The Youth Beijing Scholars Program(015),Support Project of High-level Teachers in Beijing Municipal Universities(CIT&TCD20190334)Beijing Advanced Innovation Center for Big Data-based Precision Medicine,Capital Medical University,Beijing,China(PXM2021_014226_000026).
文摘Accurate and efficient methods for identifying and tracking each animal in a group are needed to study complex behaviors and social interactions.Traditional tracking methods(e.g.,marking each animal with dye or surgically implanting microchips)can be invasive and may have an impact on the social behavior being measured.To overcome these shortcomings,video-based methods for tracking unmarked animals,such as fruit flies and zebrafish,have been developed.However,tracking individual mice in a group remains a challenging problem because of their flexible body and complicated interaction patterns.In this study,we report the development of a multi-object tracker for mice that uses the Faster region-based convolutional neural network(R-CNN)deep learning algorithm with geometric transformations in combination with multi-camera/multi-image fusion technology.The system successfully tracked every individual in groups of unmarked mice and was applied to investigate chasing behavior.The proposed system constitutes a step forward in the noninvasive tracking of individual mice engaged in social behavior.
基金the National Natural Seiene Foundar tion of China(Nos.61671423 and 61271403)。
文摘Visual tracking has been a popular task in computer vision in recent years,especially for long-term tracking.A novel object tracking framework is proposed in this paper.For surveillance cameras with overlapping areas,the target area is divided into several regions corresponding to each camera,and a simple re-matching method is used by matching the colors according to the segmented parts.For surveillance cameras without overlapping areas,a time estimation model is employed for continuously tracking objects in different fields of view(FoVs).A demonstration system for collaborative tracking in real time situation is realized finally.The experimental results show that compared with current popular algorithms,the proposed approach has good effect in accuracy and computation time for the application of continuously tracking the pedestrians.
文摘为减少因船舶偏离航道而造成的搁浅、碰撞航标或桥墩等水上交通事故,提出了一种基于多目相机自动识别航道的桥区航行异常船舶预警方法。基于YOLOv5(You Only Look Once version 5)目标检测算法,联动变、定焦相机识别并定位航标和船舶,跟踪并记录船舶航迹点,计算船舶的速度和航向并推算船位。提出了一种基于视频船舶航迹点的密度聚类识别航道两侧航标的方法,实现航道自适应可视化。基于船位推算识别并预警航行状态异常的船舶。实验结果表明:航标、船舶的检测正确率分别达84.8%、90.3%,相较单一相机检测模型,正确率分别提高了32.1%、5.5%;能够自适应可视化航道并识别、预警航行异常船舶。
文摘为提高无人车障碍物检测跟踪的精度和稳定性,首先针对YOLO v5(You only look once version 5,YOLO v5)网络存在的语义信息和候选框信息丢失的问题,引入深度可分离空洞空间金字塔结构与目标框加权融合算法完成对网络的优化;其次针对单阶段障碍物点云聚类精度低的问题,设计一种考虑点云距离与外轮廓连续性的两阶段障碍物点云聚类方法并完成三维包围盒的建立;最后将注意力机制引入MobileNet使网络更加聚焦于目标对象特有的视觉特征,并综合利用视觉特征和三维点云信息共同构建关联性度量指标,提高匹配精度。利用KITTI数据集对构建的障碍物目标检测、跟踪与测速算法进行仿真测试,并搭建实车平台进行真实环境试验,验证所提算法的有效性和真实环境可迁移性。
基金partially supported by Spanish Ministerio de Economía y Competitividad/FEDER(Nos.TIN2012-34003 and TIN2013-47074-C2-1-R)FPU Scholarship(FPU13/03141)from the Spanish Government
文摘This paper seeks to determine how the overlap of several infrared beams affects the tracked position of the user, depending on the angle of incidence of light, distance to the target, distance between sensors, and the number of capture devices used. We also try to show that under ideal conditions using several Kinect sensors increases the precision of the data collected. The results obtained can be used in the design of telerehabilitation environments in which several RGB-D cameras are needed to improve precision or increase the tracking range. A numerical analysis of the results is included and comparisons are made with the results of other studies. Finally, we describe a system that implements intelligent methods for the rehabilitation of patients based on the results of the tests carried out.
文摘同时定位与地图构建(simultaneous localization and mapping,SLAM)技术在无人化装备上有着广泛的应用,可实现室内或室外自主的定位建图任务。该文首先对视觉和激光SLAM基本框架进行介绍,详细阐述前端里程计、后端优化、回环检测以及地图构建这四个模块的作用以及所采用的算法;在这之后梳理归纳视觉/激光SLAM发展历程中的经典算法并分析其优缺点以及在此之后优秀的改进方案;此外,列举当前SLAM技术在生活中的典型应用场景,展示在自动驾驶、无人化装备等领域的重要作用;最后讨论SLAM系统当前的发展趋势和研究进展,以及在未来应用中需要考虑的挑战和问题,包括多类型传感器融合、与深度学习技术的融合以及跨学科合作的关键作用。通过对SLAM技术的全面分析和讨论,为进一步推动SLAM技术的发展和应用提供深刻的理论指导和实践参考。
