Table of Contents

30 June 2021, Volume 42 Issue 3 Previous Issue    Next Issue
---
Contents
Review
Image Processing and Computer Vision
Computer Graphics and Virtual Reality
Digital Design and Manufacture
Total to Discuss


For Selected:

Download Citations EndNote Ris BibTeX

Toggle Thumbnails

Contents

Select

Contents of the Third Issue of 2021 2021, 42(3): 0-0.  Abstract ( 152 )   PDF (225KB) ( 82 )   Related Articles | Metrics

Select

Don't forget what you set out to do Sun Jia-guang 2021, 42(3): 1-2.  Abstract ( 101 )   PDF (2321KB) ( 67 )   Related Articles | Metrics

Review

Select

Visual Turing: the next development of computer vision in the view of human-computer gaming  HUANG Kai-qi, ZHAO Xin , LI Qiao-zhe , HU Shi-yu 2021, 42(3): 339-348.  DOI: 10.11996/JG.j.2095-302X.2021030339 Abstract ( 567 )   PDF (2759KB) ( 386 )   Computer vision has gained wide attention in the research of artificial intelligence. After nearly 60 years of its development, great achievement has been made in aspect of algorithms, technologies, and applications. Over the past decade, deep learning, which is on the basis of big data and huge computation power, has further ushered computer vision in an era of large model. However, there remains a huge gap between algorithm adaptability and human beings. From the perspective of visual task evaluation (in terms of datasets, metrics, and methods), this paper summarized the development history of computer vision. In addition, a systematic analysis was conducted on the existing problems and obstacles for the development of computer vision heavily dependent on big data learning. Based on the analysis, this paper argued that the visual Turing test could be the next research direction of computer vision. Finally, the development of the visual Turing test and its potential research were discussed.  Related Articles | Metrics

Select

Digital twin and production line simulation technology  TIAN Ling, LIU Guo, LIU Si-chao 2021, 42(3): 349-358.  DOI: 10.11996/JG.j.2095-302X.2021030349 Abstract ( 421 )   PDF (957KB) ( 244 )   With the deep integration of new generation information technology and manufacturing technology, simulation technology is developing into the deep integration and efficient collaboration between information system and physical system. In this context, digital twin technology with the characteristics of virtual reality fusion has become a new research hotspot in recent years, and is the development direction of the next generation simulation technology. In order to explore the adoption of new generation information technology and speed up the digital transformation and upgrading of discrete manufacturing industry, the development context and work scenarios of simulation technology are analyzed, and the digital twin technology has been drew into the research on simulation analysis of production line. The concept and connotation of the digital twin technology are introduced, the relationship, similarities and differences between digital twin technology and traditional computer simulation technology are analyzed, and the directions of applying digital twin technology in production line are discussed. Then the construction method and realization way of digital twin model of production line are analyzed, while the research status and existing problems are summarized. Finally, the possible innovation directions of digital twin in production line simulation are given, which provide a reference for discrete manufacturing enterprises to cross the interaction gap between physical resources and digital world, establish the interaction and integration of physical space and information space of production line, and promote the digital management and intelligent production of the whole life cycle of production line, in which way to achieve the transformation of intelligent manufacturing.  Related Articles | Metrics

Select

Computational model for dynamical interaction scenarios  HUANG Jin, ZHANG Hao, TIAN Feng 2021, 42(3): 359-366.  DOI: 10.11996/JG.j.2095-302X.2021030359 Abstract ( 129 )   PDF (573KB) ( 89 )   With the development of human-computer interaction (HCI) technology, interactive systems including dynamic content, such as virtual/augmented/mixed reality, video surveillance system, etc., gradually becomes the mainstream. Computational models for dynamic interaction scenarios have attracted more and more attention. HCI researchers proposed numerous models that significantly improve our understanding of dynamical user interfaces. However, there are still many challenges in the research of computational model for dynamical interaction scenarios. The problems of complexity in dynamic interaction scenarios, lack of interpretability, weakness in descripting interaction mechanisms greatly limit the development and application of computational models for dynamical interaction scenarios. We review related work and recent advances in the computing model for dynamic interaction scenarios, summarize existing problems, and offer our predictions for the future development in this field. The purpose of this paper is to provide references for future research on computing models for dynamical interaction scenarios.  Related Articles | Metrics

Select

Research progress of user task prediction and algorithm analysis  HU Zhi-ming, LI Sheng, GAI Meng 2021, 42(3): 367-375.  DOI: 10.11996/JG.j.2095-302X.2021030367 Abstract ( 159 )   PDF (7609KB) ( 155 )   Users’ cognitive behaviors are dramatically influenced by the specific tasks assigned to them. Information on users’ tasks can be applied to many areas, such as human behavior analysis and intelligent human-computer interfaces. It can be used as the input of intelligent systems and enable the systems to automatically adjust their functions according to different tasks. User task prediction refers to the prediction of users’ tasks at hand based on the characteristics of his or her eye movements, the characteristics of scene content, and other related information. User task prediction is a popular research topic in vision research, and researchers have proposed many successful task prediction algorithms. However, the algorithms proposed in prior works mainly focus on a particular scene, and comparison and analysis are absent for these algorithms. This paper presented a review of prior works on task prediction in scenes of images, videos, and real world, and detailed existing task prediction algorithms. Based on a real-world task dataset, this paper evaluated the performances of existing algorithms and conducted the corresponding analysis and discussion. As such, this work can provide meaningful insights for future works on this important topic.  Related Articles | Metrics

Select

A review on neural radiance fields based view synthesis  CHANG Yuan, GAI Meng 2021, 42(3): 376-384.  DOI: 10.11996/JG.j.2095-302X.2021030376 Abstract ( 941 )   PDF (2981KB) ( 716 )    Image-based view synthesis techniques are widely applied to both computer graphics and computer vision. One of the key issues is how to use the information from the input image to represent a 3D model or scene. Recently, with the proposal of neural radiance fields (NeRF), a large number of research works based on this representation have further enhanced and extended the method, and achieved the expected accuracy and efficiency. This type of research can be broadly classified into two categories by purposes: the analysis and improvement of NeRF itself, and the extensions based on the NeRF framework. Methods of the first category have analyzed the theoretical properties and shortcomings of the NeRF representation and proposed some strategies for performance improvement, including the synthesis accuracy, rendering efficiency, and model generalizability. The second type of works are based on the NeRF framework and have extended the algorithm to solve more complex problems, including view synthesis using unconstrained images, view synthesis with relighting, and view synthesis for dynamic scenes. After outlining the background of the proposal of NeRF, other related works based on it were discussed and analyzed in this paper according to the classification mentioned above. Finally, the challenges and prospects were presented concerning the development of NeRF-based approaches.  Related Articles | Metrics

Select

Review on deep learning based prediction of image intrinsic properties  SHA Hao , LIU Yue 2021, 42(3): 385-397.  DOI: 10.11996/JG.j.2095-302X.2021030385 Abstract ( 210 )   PDF (6658KB) ( 328 )    The appearance of the real world primarily depends on such intrinsic properties of images as the geometry of objects in the scene, the surface material, and the direction and intensity of illumination. Predicting these intrinsic properties from two-dimensional images is a classical problem in computer vision and graphics, and is of great importance in three-dimensional image reconstruction and augmented reality applications. However, the prediction of intrinsic properties of two-dimensional images is a high-dimensional and ill-posed inverse problem, and fails to yield the desired results with traditional algorithms. In recent years, with the application of deep learning to various aspects of two-dimensional image processing, a large number of research results have predicted the intrinsic properties of images through deep learning. The algorithm framework was proposed for deep learning-based image intrinsic property prediction. Then, the progress of domestic and international research was analyzed in three areas: intrinsic image prediction based on acquiring scene reflectance and shading map, intrinsic properties prediction based on acquiring material BRDF parameters, and intrinsic properties prediction based on acquiring illumination-related information. Finally, the advantages and disadvantages of each method were summarized, and the research trends and focuses for image intrinsic property prediction were identified.  Related Articles | Metrics

Image Processing and Computer Vision

Select

A method of automatic image annotation for image-text mixed domain books ZHAO Hai-ying , GAO Zi-hui , DENG Lian , HOU Xiao-gang , LI Ning 2021, 42(3): 398-405.  DOI: 10.11996/JG.j.2095-302X.2021030398 Abstract ( 127 )   PDF (2588KB) ( 211 )   Efficient interpretation and intelligent processing of massive text and text data is a very challenging and practical work, but the accuracy of automatic labeling is highly dependent on the quality and quantity of training samples. In this paper, an image annotation method of images and text data mixed information is proposed. The method consists of three parts: adaptive image and text separation preprocessing, domain image semantic label construction and text-based image annotation algorithm. Firstly, the proposed hybrid layout recognition algorithm is used to extract the image, title and description text in the hybrid layout of images and text data. Then, the Traditional Cultural Domain Lexicon (PatternNet) is established based on semantic tags of digital clothing image. Finally, according to the characteristics of domain lexicon's tag space, a text-based image annotation algorithm is proposed to improve the large tag space. The simulation experiment is carried out on the ethnic costumes books that images and text data hybrid layout, also compared with other data sets. The experimental results verify the effectiveness of the algorithm proposed in this paper.  Related Articles | Metrics

Select

A semantic segmentation algorithm using multi-scale feature fusion with combination of superpixel segmentation  GUAN Shen-ke , LIN Xiao , ZHENG Xiao-mei , ZHU Yuan-yuan , MA Li-zhuang 2021, 42(3): 406-413.  DOI: 10.11996/JG.j.2095-302X.2021030406 Abstract ( 363 )   PDF (5858KB) ( 318 )   The advancement of deep learning has boosted the research on image semantic segmentation. At present, most effective methods for this research are based on the fully convolutional neural networks. Although the existing semantic segmentation methods can effectively segment the image as a whole, they cannot clearly identify the edge information of the overlapped objects in the image, and cannot effectively fuse the high- and low-layer feature information of the image. To address the above problems, superpixel segmentation was employed as an auxiliary optimization to optimize the segmentation results of object edges based on the fully convolutional neural network. At the same time, the design of a joint cross-stage partial multiscale feature fusion module can enable the utilization of image spatial information. In addition, a skip structure was added to the upsampling module to enhance the learning ability of the network, and two loss functions were adopted to ensure network convergence and improve network performance. The network was trained and tested on the public datasets PASCAL VOC 2012. Compared with other image semantic segmentation methods, the proposed network can improve the accuracies in pixel and segmentation, and displays strong robustness.  Related Articles | Metrics

Select

Image inpainting using non-convex and low-rank constraint  SUN Yan-min, GUO Qiang, ZHANG Cai-ming 2021, 42(3): 414-425.  DOI: 10.11996/JG.j.2095-302X.2021030414 Abstract ( 115 )   PDF (28254KB) ( 373 )   Due to transmission interference or improper storage, there exist some missing pixels in the images obtained in the real scene, which causes obstacles to the subsequent processing and analysis of the images. The key solution for missing pixels is to recover the image with low rank prior. However, since the rank function is discrete, the model that minimizes the rank is an NP-hard problem. In order to address this issue, a commonly used method is to employ an image-inpainting algorithm based on the nuclear norm. Unlike the methods based on the nuclear norm minimization, this paper proposed an image-inpainting algorithm using non-convex low-rank constraints, which replaced the traditional nuclear norm with a log function and overcame the inability of the nuclear norm to approach the rank minimization. In addition, to optimize the non-convex model, the augmented Lagrangian multiplier method was adopted to derive an alternating minimization algorithm. Experimental results demonstrate that the proposed method can deal with different missing pixel rates, and can far outperform other low-rank inpainting methods in inpainting.  Related Articles | Metrics

Select

A new automatic compression method for Chinese vector fonts  GAO Yi-chen, LIAN Zhou-hui, TANG Ying-min, XIAO Jian-guo 2021, 42(3): 426-431.  DOI: 10.11996/JG.j.2095-302X.2021030426 Abstract ( 106 )   PDF (991KB) ( 89 )   To solve the inconvenient usage of large-size Chinese vector fonts in embedded devices, this paper proposes a new automatic font compression method. Based on the idea of reusing and assembling components, different parts were first extracted from the whole glyphs using a traditional computer graphics-based method and their reusing relationships were calculated. Then, they were assembled and their positions and scales were iteratively optimized using the simulated annealing algorithm to produce the final output. Experimental results demonstrate that the proposed method can generate a compressed font whose volume is only about 20% of the original font while maintaining the font style, thus improving the availability of Chinese vector fonts in embedded devices with limited storage spaces.  Related Articles | Metrics

Select

Human pose estimation based on high-resolution net REN Hao-pan, WANG Wen-ming, WEI De-jian, GAO Yan-yan, KANG Zhi-hui, WANG Quan-yu 2021, 42(3): 432-438.  DOI: 10.11996/JG.j.2095-302X.2021030432 Abstract ( 152 )   PDF (5063KB) ( 255 )   Human pose estimation plays a vital role in human-computer interaction and behavior recognition applications, but the changing scale of feature maps poses a challenge to the relevant methods in predicting the correct human poses. In order to heighten the accuracy of pose estimation, the method for the parallel network multi-scale fusion and that for generating high-quality feature maps were combined for human pose estimation. On the basis of human detection, RefinedHRNet adopted the method for parallel network multi-scale fusion to expand the receptive field in the stage using a dilated convolution module to maintain context information. In addition, RefinedHRNet employed a deconvolution module and an up-sampling module between stages to generate high-quality feature maps. Then, the parallel network feature maps with the highest resolution (1/4 of the input image size) were utilized for pose estimation. Finally, Object Keypoint Similarity (OKS) was used to evaluate the accuracy of keypoint recognition. Experimenting on the COCO2017 test set, the pose estimation accuracy of our proposed method RefinedHRNet is 0.4% higher than the HRNet network model.  Related Articles | Metrics

Select

Action detection model fused with non-local neural network HUANG Wen-ming, YANG Mu-li, LAN Ru-shi, DENG Zhen-rong, LUO Xiao-nan 2021, 42(3): 439-445.  DOI: 10.11996/JG.j.2095-302X.2021030439 Abstract ( 92 )   PDF (603KB) ( 81 )   The convolutional neural network (CNN) has insufficient ability to understand the time domain information in video action detection. For this problem, we proposed a model based on fused non-local neural network, which combines non-local block with 3D CNN to capture global connections between video frames. Model used a two-stream architecture of 2D CNN and 3D CNN to extract the spatial and motion features of the video, respectively, which takes video single frames and video frame sequences as inputs. To further enhance contextual semantic information, an improved attention and channel fusion mechanism is used to aggregate the features of the above two networks, and finally the fused features are used for frame-level detection. We conducted experimental verification and comparison on the UCF101-24 and JHMDB data set. The results show that our method can fully integrate spatial and temporal information, and has high detection accuracy on video-based action detection tasks.  Related Articles | Metrics

Select

Research on depth prediction algorithm based on multi-task model  YAO Han, YIN Xue-feng, LI Tong, ZHANG Zhao-xuan, YANG Xin, YIN Bao-cai 2021, 42(3): 446-453.  DOI: 10.11996/JG.j.2095-302X.2021030446 Abstract ( 100 )   PDF (2078KB) ( 92 )   Image depth prediction is a hot research topic in the field of computer vision and robotics. The construction of depth image is an important prerequisite for 3D reconstruction. Traditional methods mainly conduct manual annotation based on the depth of a fixed point, or predict the depth based on binocular positioning according to the position of the camera. However, such methods are time-consuming and labor-intensive and restricted by factors such as camera position, positioning method, and distribution probability. As a result, the difficulty in guaranteeing high accuracy poses a challenge to subsequent tasks following the predicted depth map, such as 3D reconstruction. This problem can be effectively solved by introducing a deep learning method based on multi-task modules. For scene images, a multi-task model-based monocular-image depth-prediction network was proposed, which can simultaneously train and learn three tasks of depth prediction, semantic segmentation, and surface vector estimation. The network includes a common feature extraction module and a multi-task feature fusion module, which can ensure the independence of each feature while extracting common features, and guarantee the accuracy of depth prediction while improving the structure of each task.  Related Articles | Metrics

Select

AR-assisted intelligent analysis and identification system for mobile rice diseases based on HOG-SVM  XU Shi-pu, LI Lin-yi, JIA Jin-yuan, WANG Yun-sheng, LIU Chang, LIU Yong, MA Chao 2021, 42(3): 454-461.  DOI: 10.11996/JG.j.2095-302X.2021030454 Abstract ( 110 )   PDF (3544KB) ( 178 )   For the shortcomings of traditional disease recognition systems that require high altitude in a shooting environment and large numbers of samples, this research designed a set of f augmented reality (AR)-assisted recognition schemes based on histograms of oriented gradient (HOG)-support vector machine (SVM). Under the premise of a small amount of materials, this solution, which introduced AR technology in the diagnostic system for shooting assistance, outperforms other methods in terms of training time, recognition speed, and average accuracy. Taking the Android terminal as an example, an AR-assisted HOG-SVM-based mobile rice disease identification system was implemented, which can quickly identify diseases and guide users to improve the quality of photographed pictures. Through the identification of disease spots in batches of images, the results of disease spot recognition were analyzed from three aspects: disease accuracy, diseased leaf detection rate, and disease spot location accuracy. Finally, AR technology and rapid identification scheme based on HOG-SVM were obtained. This combination can generate faster training results and recognition results under the premise of small training samples. The average accuracy of this system is also higher than that of deep models such as YOLO v3, SSD 512, and Fast R-CNN. The proposed method is more practicable for disease identification on the current mobile terminal.  Related Articles | Metrics

Computer Graphics and Virtual Reality

Select

Research on adaptive grasping of virtual hands based on deep reinforcement learning  WU Yi-he , ZHANG Zhen-ning , QIU Dong , LI Wei-qing , SU Zhi-yong 2021, 42(3): 462-469.  DOI: 10.11996/JG.j.2095-302X.2021030462 Abstract ( 96 )   PDF (8063KB) ( 220 )   For the grasping of computer character animation, it is difficult to guarantee the naturalness, stability and adaptability of the generated action sequence at the same time. In other words, the natural and stable grasping controller are often limited in generalization and cannot be applied to other types of grabbing tasks. A virtual hand adaptive grasping controller was constructed based on deep reinforcement learning by introducing hand teaching data corresponding to the grasping types and by designing the reward function. Experimental results show that the designed controller can generate a grasping motion sequence with both naturalness and stability, and are also highly adaptive for different sizes and types of primitive objects in the material library.  Related Articles | Metrics

Select

Research on projector interaction method with common pen and paper GUO Jun-xiu, LIU You-quan, WANG Song-xue 2021, 42(3): 470-477.  DOI: 10.11996/JG.j.2095-302X.2021030470 Abstract ( 83 )   PDF (1078KB) ( 191 )    Projector interaction has freed the interaction space from the electronic screen, which is more eye-friendly than other display modes and can produce abundant interesting interaction effects. In addition,the interaction with pen-writing is inherently natural and efficient. With the integration of the advantages of both, a projection interactive method was proposed based on common pen and paper. Users can interact with the computer through common paper and pen, while the camera was employed to collect the handwriting images, and projector was utilized to present recognition result. Moreover, with the integration of CNN and KNN, the present method can automatically adapt to users’ handwriting style,so as to increase the recognition accuracy. Based on this, two typical applications are designed, including English-word remembering and simple mathematical calculation. The experiments and user surveys verified that the proposed interaction method is more natural, convenient and practical than other interaction methods.  Related Articles | Metrics

Select

Design and research of the hose-drogue aerial refueling hardware-in-the-loop simulation system  ZHOU Ren-xian , YANG Shang-jun , LIN Yi-jun 2021, 42(3): 478-484.  DOI: 10.11996/JG.j.2095-302X.2021030478 Abstract ( 130 )   PDF (892KB) ( 99 )   In order to simulate the relative motion process of hose-drogue aerial refueling, explorations were made on the aerodynamic influence and flight approach strategy of the tanker and the receiver at close range. This paper proposed a motion model of hose refueling parachute, considering the tail wake vortex effect and the head wave effect. The simulation results show that the proposed model can accurately simulate the relative motion process of soft aerial refueling. Based on the large-scale ball screen flight training simulator of a certain type of aircraft, a soft hardware-in-the-loop simulation training system for aerial refueling was developed, the control method of the parachute component was designed in detail, and a set of visual simulation training plug-in for air refueling was developed using OpenGL language.  Related Articles | Metrics

Select

Parallel intelligent transportation system framework enhanced by visual analysis   LIU Li-yan, ZHANG Hong-xin , CHEN Wei , DI Yi-ning , LIU Jia-xin , MAN Jia-ju 2021, 42(3): 485-491.  DOI: 10.11996/JG.j.2095-302X.2021030485 Abstract ( 85 )   PDF (2019KB) ( 73 )   With the advent of artificial intelligence 2.0 era, visual analysis methods have received more and more attention as an important human-machine coupling method. It is a powerful tool for big data analysis and a“navigator”for understanding data. It can effectively convert data in a ternary spatial structure (cyber-physical-human, CPH) into services and decision-making in a knowledge system, thereby further enhancing the intelligent level of transportation system. At the same time, a parallel intelligent transportation system that integrates artificial transportation system, computational experiment and parallel execution is proposed, which provides a new mechanism and new mode of manipulation in the field of intelligent transportation. Through the analysis of specific cases, we discuss the importance of visual analysis in the new generation of artificial intelligence, and the process transforming data or information into knowledge systems. It is proved that the seamless combination of visual analysis and parallel intelligent transportation system can better analyze large-scale traffic data, solve traffic problems more effectively, and achieve the enhancement effect of “1+1>2”. Based on this, a novel parallel intelligent transportation system framework enhanced by visual analysis is proposed.  Related Articles | Metrics

Digital Design and Manufacture

Select

Computation method of variable capacity constrained centroidal Power diagram  YAO Yu-you, ZHANG Gao-feng, XU Ben-zhu, ZHENG Li-ping 2021, 42(3): 492-500.  DOI: 10.11996/JG.j.2095-302X.2021030492 Abstract ( 147 )   PDF (595KB) ( 75 )   The Power diagram, as an extension of the Voronoi diagram, introduces “weight” to each site, and is characteristic of accurate tolerance. By imposing the capacity constraints to the ordinary Power diagram, a capacity-constrained Power diagram can be obtained, where the capacity of each site equates to the preset capacity constraint. The addition of the centroid constraints on a secondary basis can lead to the centroidal capacity-constrained Power diagram, in which the sites are located at its mass centers of the corresponding Power cells. In these Power diagrams, the capacity constraints are clear values. However, the capacity constraints are often intervals in some practical applications. To address this problem, a computation method was proposed for variable capacity-constrained centroidal Power diagram. On the one hand, the method can continuously update the weights of sites to meet the capacity constraints. On the other hand, the Lloyd’s method is applied to the relocation of the sites to its mass centers of the corresponding Power cells. The two steps interfere with each other in the optimization process to compute the centroidal Power diagram with interval capacity constraints. The experimental results demonstrate that the proposed method can stably compute the variable capacity-constrained centroidal Power diagram under different conditions with the advantages in high efficiency and adaptability.  Related Articles | Metrics

Select

Isogeometric topology optimization of blended B-spline solid model  YANG Jia-ming , ZHAO Gang, WANG Wei, GUO Ma-yi , DU Xiao-xiao 2021, 42(3): 501-510.  DOI: 10.11996/JG.j.2095-302X.2021030501 Abstract ( 142 )   PDF (5166KB) ( 126 )   For isogeometric topology optimization (ITO) methods, isogeometric analysis (IGA) is adopted for topology optimization to address the limitation of the finite element method, which can improve the efficiency and stability of the optimization. However, it is of great challenge for existing ITO methods to manage arbitrarily shaped design domains, especially in three-dimensional solid problems. Therefore, a new ITO method was proposed to handle unstructured solid models. A spline solid with complex structures was obtained from an unstructured hexahedral mesh based on the blended B-spline construction. The basis functions describing the unstructured spline solid were applied to the representation of density distribution and the calculation of IGA. Several examples proved the flexibility and robustness of the proposed method in dealing with complex structures. These results may shed light on the application of ITO in practical engineering problems.  Related Articles | Metrics

Select

Cylindrical feature hole recognition for automatic repair  WANG Chun-xiang, LIU Liu, ZHOU Guo-yong, JI Kang-hui 2021, 42(3): 511-516.  DOI: 10.11996/JG.j.2095-302X.2021030511 Abstract ( 77 )   PDF (2634KB) ( 104 )   For the point cloud model with complex profiles and many holes, the existing reverse software and repair algorithm exhibit such problems as lower efficiency and excessive human-computer interaction in hole-by-hole repairing, as well as low repair accuracy and loss of features in multi-hole repairing. Therefore, it is necessary to develop an efficient and high-precision automatic repair mode aiming for feature preserving based on identification and classification of holes in point clouds. Based on the above ideas, a hole recognition-based method was proposed for separating cylindrical feature holes from general types of holes. First, boundary points of holes were extracted and collected following the criterion of maximum angle recognition; the total number of holes was calculated with Euclidean clustering. After that, cylindrical feature holes in the model were extracted using RANSAC and the set distance threshold. According to experimental results, the method can both extract from the model multiple cylindrical feature holes of different diameters, and make estimate of geometric parameters of the cylindrical surface more accurately. In this way, the automatic repairing-oriented identification of a specific type of feature holes can be achieved.  Related Articles | Metrics

Select

Parametric surface-based additive manufacturing conformal lattice structure generation method  XIAO Wen-lei, LIN Zai-sheng, XIONG Chang-ri, WANG Shi-ping, WEI Wei, ZHAO Gang 2021, 42(3): 517-524.  DOI: 10.11996/JG.j.2095-302X.2021030517 Abstract ( 87 )   PDF (6130KB) ( 385 )   The lattice structure has become one of the most important research fields in the design and manufacture of complex structures in additive manufacturing, because of its special mechanical properties. The conventional method for model lattice structure generation is achieved by trimming the parametric modeling lattice structure grid or by performing conformal deformation of the lattice structure. Such methods are relatively inefficient. A conformal lattice structure generation method was proposed based on parametric surface for additive manufacturing, which can realize the adaptation and efficient generation of lattice structure in surface space. Firstly, a matrix-based approach was proposed to express and construct the lattice structure frame. Then, three types of enclosed spaces, which were formed by point-surface, curve-surface, and surface-surface patterns, were implemented to perform conformal transformation on the lattice frame to adapt the surface space. Finally, a grid generation and splicing method based on the lattice structure skeleton was employed to generate the mesh model of the conformal lattice structure. The parametric surface was extracted from the CATIA model based on CAA to generate the conformal lattice structures effectively, and the structures can adapt to the surface space well, which proves the potential value for the engineering implementation and application of the method.  Related Articles | Metrics

Total to Discuss

Select

Vol.42, No.3, 2021 2021, 42(3): 339-524.  Abstract ( 16 )   PDF (93683KB) ( 131 )   Related Articles | Metrics