Well, I dont quite have openGL code, but this site has some good explanations on different types of collision detection, the pros and cons of each, what to avoid, etc. It does have some pseudo code, which shouldnt be that hard to convert. Good luck!

http://www.gamedev.net/reference/docs/refarticlelistuser.asp?catid=45

(BTW, the first few links are mainly 2d related, but the last one is 3d)

-Fuzz

quote:Original post by cxi

AnyOne know of good OPENGL code for COllision Detection i can get? I can not seem to get mine to work, and i woudl love to see some good examples. please email me dan@pjcc.com


thanks

It strikes me that the easiest way to detect collision along a path (Ie where I am now to where I am after I bang
along this here path a few notches, would be essentially
to go through ones list of polygons, and see if the line between the two points intersects, a b-tree or somethin''
might be usefull to break down the poly''s to be tested.

I''d figure that''d be the easiest.
The Code is a Living breathing entity, and will move in accordance with you, therefore, be one with the Tao of the Code..

Hi cxi,

You''re heading in the general direction, but not quite right. There are heaps of articles aroung on collision detection (although, I must admit some are pretty full-on), but to get you started I''ll quickly explain the bounding sphere techique.

Nominate a central point inside each of your objects, and figure out what size radius a sphere would need to have to (only just) completely enclose the object. This is your bounding sphere. Now, when your objects move (but before drawing the a frame, so you can change stuff like object positions etc after collision if you need to), test the distance between the centre of each object. If it is less than the sum of the radii of their respective bounding spheres, they have ''collided'' (as you may have guessed, this method is not very accurate, but it is fast, and is good as a first level of collision detection).

You can refine this in many ways. One is to not bother testing collision between static objects (as they will never collide), and make sure you don''t test for collision between two object twice (ie test all objects against one object for collision, then iterate to the next object and test for collision with the previous object... you already did that ).

This method is a good first layer. You can then refine your detection by testing for collision with a heirarchy of bounding boxes if a ''collision'' is detected using the bounding sphere (to detect if there actually was a collision).

Hope this was helpful in some small way... I''d hate to think I was wasting my time, being at work and all

-------------
squirrels are a remarkable source of protein...

Hi again cxi,

Glad I could be of some use to someone for a change

With the bounding sphere, you don''t actually draw it, its just a conceptual thing that surrounds the object. It is simply an imaginary sphere with a radius from the centre of the object to the vertex furthest from that centre. You are just using it to define a distance (i any direction) from the object centre that, if another object''s sphere comes closer than this distance, a collision may have occurred (as the object resides somewhere within this imaginary sphere).

Am I making sense? Sorry if I''ve confused you... the main point I''m trying to make is that you don''t need to draw bounding spheres/boxes/etc, they are simply (approximate)mathematical definitions of the space that your object takes up.

Let me know if you want me to clarify anything (just don''t expect too much, as I''ve only just started in this type of thing recently too)

Well, back to punching out crappy database accessing code with VB... the joy just never stops

-------------
squirrels are a remarkable source of protein...