This paper theoretically analyzes and researches the coordinate frames of a 3D vision scanning system, establishes the mathematic model of a system scanning process, derives the relationship between the general non-or...This paper theoretically analyzes and researches the coordinate frames of a 3D vision scanning system, establishes the mathematic model of a system scanning process, derives the relationship between the general non-orthonormal sensor coordinate system and the machine coordinate system and the coordinate transformation matrix of the extrinsic calibration for the system.展开更多
Surface quality monitoring of manufacturing products is critical for manufacturing industries to ensure product quality and production efficiency.With the rapid development of 3D scanning technology,high-density 3D po...Surface quality monitoring of manufacturing products is critical for manufacturing industries to ensure product quality and production efficiency.With the rapid development of 3D scanning technology,high-density 3D point cloud data can be generated by 3D scanners in complex manufacturing systems.However,due to the challenges of complex surface modeling and various types,it lacks effective surface anomaly detection methods that can meet the practical requirements regarding detection accuracy and speed.This survey aims to review the surface anomaly detection methodology of manufacturing products based on 3D machine vision.Specifically,the machine learning methodologies will be systematically reviewed for 3D point cloud data modeling and anomaly detection.Related public data sets for this research are also summarized.Finally,the future research directions are pointed out.展开更多
自然语言描述驱动的目标跟踪是指通过自然语言描述引导视觉目标跟踪,通过融合文本描述和图像视觉信息,使机器能够“像人类一样”感知和理解真实的三维世界.随着深度学习的发展,自然语言描述驱动的视觉目标跟踪领域不断涌现新的方法.但...自然语言描述驱动的目标跟踪是指通过自然语言描述引导视觉目标跟踪,通过融合文本描述和图像视觉信息,使机器能够“像人类一样”感知和理解真实的三维世界.随着深度学习的发展,自然语言描述驱动的视觉目标跟踪领域不断涌现新的方法.但现有方法大多局限于二维空间,未能充分利用三维空间的位姿信息,因此无法像人类一样自然地进行三维感知;而传统三维目标跟踪任务又依赖于昂贵的传感器,并且数据采集和处理存在局限性,这使得三维目标跟踪变得更加复杂.针对上述挑战,本文提出了单目视角下自然语言描述驱动的三维目标跟踪(Natural Language-driven Object Tracking in 3D,NLOT3D)新任务,并构建了对应的数据集NLOT3D-SPD.此外,本文还设计了一个端到端的NLOT3D-TR(Natural Language-driven Object Tracking in 3D based on Transformer)模型,该模型融合了视觉与文本的跨模态特征,在NLOT3D-SPD数据集上取得了优异的实验结果.本文为NLOT3D任务提供了全面的基准测试,并进行了对比实验与消融研究,为三维目标跟踪领域的进一步发展提供了支持.展开更多
文摘This paper theoretically analyzes and researches the coordinate frames of a 3D vision scanning system, establishes the mathematic model of a system scanning process, derives the relationship between the general non-orthonormal sensor coordinate system and the machine coordinate system and the coordinate transformation matrix of the extrinsic calibration for the system.
基金upported by the National Natural Science Foundation of China under(Grant Nos.72371219,72001139,52372308 and 72371217)Guangdong Basic and Applied Basic Research Foundation(Grant No.2023A1515011656)+3 种基金Guangzhou Funding Program(Grant No.2025A04J5288)Guangzhou-HKUST(GZ)Joint Funding Program(Grant Nos.2023A03J0651 and 2024A03J0680)Guangzhou Industrial Informatic and Intelligence Key Laboratory(No.2024A03J0628)Nansha Key Area Science and Technology(Project Nos.2023ZD003,and Project No.2021JC02X191).
文摘Surface quality monitoring of manufacturing products is critical for manufacturing industries to ensure product quality and production efficiency.With the rapid development of 3D scanning technology,high-density 3D point cloud data can be generated by 3D scanners in complex manufacturing systems.However,due to the challenges of complex surface modeling and various types,it lacks effective surface anomaly detection methods that can meet the practical requirements regarding detection accuracy and speed.This survey aims to review the surface anomaly detection methodology of manufacturing products based on 3D machine vision.Specifically,the machine learning methodologies will be systematically reviewed for 3D point cloud data modeling and anomaly detection.Related public data sets for this research are also summarized.Finally,the future research directions are pointed out.
文摘自然语言描述驱动的目标跟踪是指通过自然语言描述引导视觉目标跟踪,通过融合文本描述和图像视觉信息,使机器能够“像人类一样”感知和理解真实的三维世界.随着深度学习的发展,自然语言描述驱动的视觉目标跟踪领域不断涌现新的方法.但现有方法大多局限于二维空间,未能充分利用三维空间的位姿信息,因此无法像人类一样自然地进行三维感知;而传统三维目标跟踪任务又依赖于昂贵的传感器,并且数据采集和处理存在局限性,这使得三维目标跟踪变得更加复杂.针对上述挑战,本文提出了单目视角下自然语言描述驱动的三维目标跟踪(Natural Language-driven Object Tracking in 3D,NLOT3D)新任务,并构建了对应的数据集NLOT3D-SPD.此外,本文还设计了一个端到端的NLOT3D-TR(Natural Language-driven Object Tracking in 3D based on Transformer)模型,该模型融合了视觉与文本的跨模态特征,在NLOT3D-SPD数据集上取得了优异的实验结果.本文为NLOT3D任务提供了全面的基准测试,并进行了对比实验与消融研究,为三维目标跟踪领域的进一步发展提供了支持.
