FoldedGI: A highly parallel algorithm for interference detection by folding a geometry image into a 1D buffer
Introduction
Collision detection & interference detection [30] is a fundamental operation of computer animation, robotics, engineering simulation and virtual reality. Given two solid polyhedral objects, collision detection cares about if and when they are about to, or have already, come into contact with each other while interference detection focuses on identifying interferences between involved components based on proximity queries [6]. In this paper, we are more interested the latter topic.
Typical interference detection methods are bounding volume hierarchies (BVH) [6], [12] and binary space partitioning (BSP) trees [25], [42]. Bounding volumes are used to improve the efficiency of interference detection by using simple volumes to enclose more complex shapes. Commonly used bounding volumes include bounding spheres [18], bounding ellipsoids [34], bounding cylinders [16], bounding boxes [1], discrete oriented polytopes [27], convex hulls [2] and so on. By contrast, the BSP trees [25], often obtained by a recursive partitioning scheme, are able to report which cube is possibly occupied simultaneously by the two objects of interest.
With the rapid development of graphics hardware, more and more researchers concentrate on using a Graphics Processing Unit (GPU) to gain an overall speedup [3]. However, implementing the above mentioned hierarchical interference detection techniques on GPU is non-trivial. Therefore, a highly parallel interference detection scheme is desirable. On this side, Liu et al. [32] proposed a GPU based algorithm that performs massively parallel sweep-and-prune and mitigates the great density of intervals along an optimal direction, but their algorithm still requires a spatial subdivision step where the memory cost climbs sharply if hierarchical subdivision is replaced by uniform subdivision (for purpose of full parallelization). Furthermore, their algorithm has to depend on a set of user specified parameters.
In this paper, we propose a highly parallel framework for detecting the penetration status and reporting the intersection contours in real time. Our algorithm begins with a quad patch decomposition step. Then we parameterize each patch and pack them into a single geometry image [24] that has a completely regular structure, typically an n × n image. Our proposed parallel interference detection technique is inspired by two important observations: 1) the interference detection problem can be converted into finding common colors contained in the geometry images, and 2) the RGB space can be mapped onto a 1D buffer by . The 1D buffers encode the necessary geometry information and serve as the abstract representation for the interference detection purpose. All key operations required can be parallelized and thus are easy to implement via GPU.
The remaining of the paper is organized as follows. Section 2 reviews the related work. In Section 3, we present a highly parallel interference detection framework that is suitable for GPU implementation. In Section 4, we use a collection of experimental results to demonstrate the high performance and some typical applications, and compare this novel algorithm with existing interference detection approaches, and Section 6 draws the conclusion and proposes future research topics.
Section snippets
Related work
Interference detection & collision detection.In almost all forms of computer animation and physical simulation, interference detection & collision detection is a vital task. It is also one of the most computationally expensive, and therefore a frequent impediment to efficient implementation of real-time graphics applications. There is a large body of literature, and we refer to Lin and Gottschalk [30] and Kockara et al. [28] for a comprehensive overview of the subject. One way to improve the
Foldedgi
In this section, we give the details about the novel parallel algorithm for interference detection. Sections 3.1 and 3.2 briefly introduce the motivation and algorithm overview respectively, while the next subsections describe the three key technical steps, including mapping the RGB space to a 1D buffer, reporting the detection result and backtracing intersection contours.
Experimental results
We test the interference detection algorithm on the computer with the following configuration: Intel e7400 CPU, NVIDIA Geforce 605 GPU, and Win7 operating system. The programming language is CUDA 7.5.
Preprocessing time.The preprocessing phase includes mesh parametrization and rendering geometry images, where we used Floater’s mean value coordinates [17] to parameterize the input mesh into a unit square. The preprocessing time cost can be seen in the following table, where the geometry image is
Limitations
In spite of some nice properties, our algorithm, in its current form, still has several disadvantages. First, the discussion in this paper is confined to interference detection. For the collision detection problem, we have to update the 1D buffers frequently, which may be inefficient. Second, our method has to rely on a parameterization step, which makes it hard to handle non-manifold meshes. Last but not least, the coarse-to-fine surface decomposition strategy belongs to post assessment, and
Conclusions and future work
We developed a highly parallel algorithm for interference detection based on folding a geometry image into a 1D buffer. We also demonstrate some experimental results on dynamic interference detection, penetration depth computation and boolean operations, which shows that the parameter-free approach is highly time- and space-efficient.
In the future, we will investigate better strategy on how to decompose the input surface to further reduce the distortion of parameterization. At the same time, we
Acknowledgments
The work described in this paper was partially supported by the key project of NSFC (61332015), the National Natural Science Foundation of China (61772016, 61571251 and 61772318), the Public Technical Application Research Project of Zhejiang (2015C34004), Ministry of Education (RG26/17), NSF of NingBo(2016A610041), and grants from K.C. Wong Magna Fund in Ningbo University.
References (45)
- et al.
Obbtree: a hierarchical structure for rapid interference detection
Proceedings of the 23rd Annual Conference on Computer Graphics and Interactive Techniques
(1996) - et al.
Interactive generalized penetration depth computation for rigid and articulated models using object norm
ACM Trans. Graph. (TOG)
(2014) Binary Space Partitions: Recent Developments
Combinatorial and Computational Geometry
(2005)Collision detection based on bounding box for NC machining simulation
Phys Procedia
(2012)GPU accelerated convex hull computation
Comput. Graph.
(2012)- et al.
Recode: an image-based collision detection algorithm
The Sixth Pacific Conference on Computer Graphics and Applications, 1998.
(1998) - et al.
Untangling cloth
ACM Trans. Graph.
(2003) - et al.
The minimal bounding volume hierarchy
Proc. of Vision, Modeling, and Visualization
(2010) - et al.
Gi-collide: collision detection with geometry images
Proceedings of the 21st Spring conference on Computer Graphics
(2005) - et al.
Optimal Binary Space Partitions in the Plane
Computing and Combinatorics
(2010)
Geometry videos: a new representation for 3d animations
Proceedings of the 2003 ACM SIGGRAPH/Eurographics Symposium on Computer Animation
Robust treatment of collisions, contact and friction for cloth animation
Proceedings of the 29th Annual Conference on Computer Graphics and Interactive Techniques
Efficient geometrically exact continuous collision detection
ACM Trans. Graph.
Quantized global parametrization
ACM Trans. Graph.
Efficient collision detection using a dual bounding volume hierarchy
Proceedings of the 5th International Conference on Advances in Geometric Modeling and Processing
Spectral geometry image: image based 3d models for digital broadcasting applications
IEEE Trans. Broadcast.
Hardware accelerated construction of sah-based bounding volume hierarchies for interactive ray tracing
Proceedings of the ACM SIGGRAPH Symposium on Interactive 3D Graphics and Games
Greedy optimal homotopy and homology generators
Proceedings of the Sixteenth Annual ACM-SIAM Symposium on Discrete Algorithms
Interactive robot trajectory planning and simulation using augmented reality
Robot. Comput.-Integr. Manuf.
Mean value coordinates
Comput. Aided Geom. Des.
Interactive collision detection between deformable models using chromatic decomposition
ACM Trans. Graph.
Fast and reliable collision culling using graphics hardware
ACM Symposium on Virtual Reality Software and Technology
Cited by (2)
Wire arc additive manufacturing from the perspective of remanufacturing: A review of data processing
2023, Journal of Manufacturing ProcessesResearch on adaptive adjustment of welding torch pose in wire and arc additive remanufacturing of hot-forging dies
2022, International Journal of Advanced Manufacturing Technology