0

In Christer Ericson great book on Collision Detection, he gives an accelerated GPU method for detecting collision among convex polygons with the following algorithm.

I'm completely new to OpenGL, and my question is, given two std::vector of Points where:

struct Point{
    double x,
    double y, 
}

How can I pass this std::vector<Point> to the following function and return the result in C++? My vector is a convex polygon order CCW.

// Initialize depth buffer to far Z (1.0)
glClearDepth(1.0f);
glClear(GL_DEPTH_BUFFER_BIT);
// Disable color buffer writes
glColorMask(GL_FALSE, GL_FALSE, GL_FALSE, GL_FALSE);
// Enable depth testing
glEnable(GL_DEPTH_TEST);
// Initialize occlusion queries
Gluint query[1], numSamplesRendered;
glGenQueries(1, query);
// Set pixels to always write depth
glDepthFunc(GL_ALWAYS);
glDepthMask(GL_TRUE);
// Draw front faces of object A
glCullFace(GL_BACK);
RenderObject(A);
// Pass pixels if depth is greater than current depth value
glDepthFunc(GL_GREATER);
// Disable depth buffer updates
glDepthMask(GL_FALSE);
// Render back faces of B with occlusion testing enabled
glBeginQuery(GL_SAMPLES_PASSED, query[0]);
glCullFace(GL_FRONT);
RenderObject(B);
glEndQuery(GL_SAMPLES_PASSED);
// If occlusion test indicates no samples rendered, exit with no collision
glGetQueryObjectuiv(query[0], GL_QUERY_RESULT, &numSamplesRendered);
if (numSamplesRendered == 0) return NO_COLLISION;
// Set pixels to always write depth
glDepthFunc(GL_ALWAYS);
glDepthMask(GL_TRUE);
// Draw front faces of object B
glCullFace(GL_BACK);
RenderObject(B);
// Pass pixels if depth is greater than current depth value
glDepthFunc(GL_GREATER);
// Disable depth buffer updates
glDepthMask(GL_FALSE);
// Render back faces of A with occlusion testing enabled
glBeginQuery(GL_SAMPLES_PASSED, query[0]);
glCullFace(GL_FRONT);
RenderObject(A);
glEndQuery(GL_SAMPLES_PASSED);
// If occlusion test indicates no pixels rendered, exit with no collision
glGetQueryObjectuiv(query[0], GL_QUERY_RESULT, &numSamplesRendered);
if (numSamplesRendered == 0) return NO_COLLISION;
// Objects A and B must be intersecting
return COLLISION;
DaynaJuliana
  • 1,144
  • 1
  • 14
  • 33

1 Answers1

1

One possible way :

  1. U need to pass your std::vector<Point> to a function which will update object's (A or B or both) buffers (vertex/index).
  2. Call the function which will do the occlusion query and return result. Probably a boolean in your case
Raki
  • 329
  • 3
  • 18