Sign in to follow this  
reachsky

Dividing 3D complex objects to detect collision

Recommended Posts

reachsky    122
Hi all! I was working on collision avoidance between 3D complex objects. I got the very first step working - collision avoidance between bounding boxes. (except when objects are rotated, then my bounding box is way too big) but then I got stuck. According to all forums and research articles, now I need to divide my objects into smaller pieces - triangulate them. but then some other articles say that I need to divide them into tetrahedrons. What should really happen? after I detect collision, I should be able to identify areas for collision and merge objects. What should really happen after bounding boxes report to collide? and could anyone share some useful links for algorithms (preferably with pseudo code and running time) of how to get this done. because as far as I understand java 3d would not provide me any tools for this, right? Thanks!

Share this post


Link to post
Share on other sites
Grain    500
It’s not exactly a trivial matter. You need to look up OBBs, Octrees, and primitive tests. OBBs (oriented bounding box) rotate with the object and thus won’t become bigger when it’s rotated. Octrees are used divide up your geometry and provide an efficient method detecting possible collisions. On top of that you will also need primitive tests (triangle on triangle, edge on edge, point on triangle, ect.. ) to determine if a possible collision is an actual collision..

Share this post


Link to post
Share on other sites
Guest Anonymous Poster   
Guest Anonymous Poster
Quote:
Original post by Grain
It’s not exactly a trivial matter. You need to look up OBBs, Octrees, and primitive tests. OBBs (oriented bounding box) rotate with the object and thus won’t become bigger when it’s rotated. Octrees are used divide up your geometry and provide an efficient method detecting possible collisions. On top of that you will also need primitive tests (triangle on triangle, edge on edge, point on triangle, ect.. ) to determine if a possible collision is an actual collision..


Or if you don't have the time to write everything. Then you can use bounding spheres for corse grained collision detection and then use simplified triangle meshes with triangle-triangle collision checks for the fine grained tests. The simplified meshes could be made from the original high polygon model, and animated together.

Viktor

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this