Search

In this paper, a fast sphere-tree generation method used for collision detection called Branch On-collide Sphere-trees is proposed. Using the video card graphic processing unit (GPU), a sphere-tree is constructed in real-time inside an animation. With this method, the core memory usage is minimized because no pre-computed data is loaded at any time during simulation life cycle. With our method, real-time conservative collision detection is achieved using the GPU, core memory is managed efficiently and the error is lowered using fast-construction sphere-tree structures.

The location of an object or parts of an object in huge environments is of fundamental importance in several research areas. The problem to locate an object or parts of an object in very complex systems (environments with large number of objects) has been studied by several authors, contributing with a large number of solutions based on different hierarchical representations of the scenes. In this research report, an implementation of a Multiresolution K dimensional tree (Multiresolution Kdtree, MKtree) and its results are presented. A MKtree represents a hierarchical subdivision of the scene objects, that guarantees a minimum space overlap between node regions in large environments. The implementation is related to ship design applications where the number and distribution of objects are considered complex. This will serve for later inclusion of methods to better approximate objects with tighter bounding volumes.

In this paper, a review of models and algorithms that are conceived to solve the collision queries problem is presented. Here, the type of query in proximity and collision computations is stated. The phases to solve the collision detection, prediction and proximity queries are exposed and classified. A collection of methods that are found in the literature are explained.

The problem to represent very complex systems has been studied by several authors, obtaining solutions based on different data structures. In this paper, a K dimensional tree (Multirresolution Kdtree, MKtree) is introduced. The MKtree represents a hierarchical subdivision of the scene objects that guarantees a minimum space overlap between node regions. MKtrees are useful for collision detection and for time-critical rendering in very large environments requiring external memory storage. Examples in ship design applications are described.

Collision detection between several objects in a closed space is an interesting research topic in all those areas where objects must not penetrate one another, such as Computer Graphics, Computer Vision, Virtual Reality. In this paper, we review data structures that are useful for collision detection and several methods that treat the collision problem. The paper is focused on collision detection in complex systems with a large number of objects and specially centered in the first step algorithm (the broad stage) where relevant regions for collision are detected.

Reconstructing 3D models from their projection images has been an interesting research topic in computer vision and geometric modeling. In this paper, we review methods that generate octree models of objects from their silhouettes in images. The complexity analysis of the algorithms is developed and presented. It is concluded that the performance of the algorithms depends mostly on the number of starting images and the object shapes through their number of faces and concavities.

Memòria de l'assignatura d'Infomàtica bàsica II del curs 94-95.

Collision detection between several objects in a closed space is of fundamental importance in all those areas where objects must not penetrate one another, such as computer grafics, robotics, motion planning, computer animation, virtual reality, etc. The problem to detect collisions in very complex systems has been studied by several authors, obtaining solutions based on different data structures. In this paper, a new K dimensional tree Multiresolution Kdtree, MKtree is introduced. The MKtree represents a hierarchical subdivision of the scene objects that guarantees a minimum space overlap between node regions. MKtrees are useful for collision detection and for time--critical rendering in very large environments requiring external memory storage. Examples in ship design applications are described.

In this report, a new $ dimensional tree data structure ({emph Multiresolution Kdtree, MKtree}) is introduced. In a complex scene containing a high number of individual objects, the MKtree represents a hierarchical subdivision of the scene objects that guarantees a small space overlap between node regions. The proposed MKtree data structure succeeds in performing simultaneosly space and scene subdivision. MKtrees are useful for {it broad phase} collision and proximity detection tests and for time--critical rendering in large environments requiring external memory storage. Examples in ship design applications are presented and discussed.

In this paper, a novel sphere-tree generation method used for collision detection is proposed. Using existing consumer-level graphics cards and its programmable graphics processing units (GPU) a sphere-tree is constructed in real-time inside an animation. This guarantees that no construction or loading of a precomputated hierarchy is required. With our method, core memory is managed in an efficient manner, allocating and releasing memory space as necessary. By this, out-of-core techniques can perform better in real-time situations. The animation tests maintain an above the average performance and the collision detection is fast and efficient.

In this paper, the collision prediction between polyhedra under screw motions and a static scene using a new $ dimensional tree data structure ({emph Multiresolution Kdtree, MKtree}) is introduced. In a complex scene containing a high number of individual objects, the MKtree represents a hierarchical subdivision of the scene objects that guarantees a small space overlap between node regions. The proposed MKtree data structure succeeds in performing simultaneously space and scene subdivision. MKtrees are useful for {it broad phase} collision and proximity detection tests and for time--critical rendering in large environments requiring external memory storage. The paper proposes an efficient broad phase collision prediction algorithm. Examples in ship design applications are presented and discussed.

Reconstructing 3D models from their projection images has been an interesting research topic in computer vision and geometric modeling. In this paper, we review methods that generate octree models of objects from their silhouettes in images. The complexity analysis of the algorithms is developed and presented. It is concluded that the performance of the algorithms depends mostly on the number of starting images and the object shapes through their number of faces and concavities.

Collision detection between several objects in a closed space is an interesting research topic in all those areas where objects must not penetrate one another, such as Computer Graphics, Computer Vision, Virtual Reality. In this paper, we describe a new data structure design that is useful for collision detection in the environment of ship design applications and several algorithms that solve the collision problem in such environment. The paper is focused on collision detection in complex systems with a large number of objects.