This article explains how to implement mouse selection in OpenGL scene, as well as moving, rotating and zooming OpenGL scene and sterilization in MDI applications. The project uses MFC Template classes to hold object primitives. There are several methods how you can implement selection in OpenGL scene. Here is a short description.
Selection method is OpenGL mechanism and it works in the following way. Draw your scene into the framebuffer, and then you enter selection mode and redraw the scene. When you're in selection mode, the contents of the framebuffer doesn't change until you exit selection mode. When you exit selection mode, OpenGL returns a list of the primitives that intersect the viewing volume. A selection hit is caused by intersection between primitive and viewing volume. The list of primitives is actually returned as an array of integers. You construct the name stack by loading names onto it as you issue primitive drawing commands while in selection mode. Thus, when the list of names is returned, you can use it to determine which primitives might have been selected on the screen by the user.
Feedback is similar to selection in that once you're in either mode, no pixels are produced and the screen is frozen. Drawing does not occur; instead, information about primitives that would have been rendered is sent back to the application. The key difference between selection and feedback modes is what information is sent back. In feedback mode, information about transformed primitives is sent back to an array of floating-point values. The values sent back to the feedback array consist of tokens that specify what type of primitive (point, line, polygon, image, or bitmap) has been processed and transformed, followed by vertex, color, or other data for that primitive.
Yet another method involves shooting a pick ray through the mouse location and testing for intersections with the currently displayed objects. OpenGL doesn't contain any function for ray intersections. To generate pick ray, you need OpenGL matrix transformation, and some knowledge about linear algebra.
The method which I implemented for mouse selection involves projection of small area from the object (primitive) to the screen and testing the mouse click. The code below shows how you can project the object point to the screen.
m_spheres->winy=m_spheres->viewport - (GLint) y - 1;
Every shape object contains projected point and rectangle. When you want to select shape, use the
LButtonDown message handler and test its point in rectangle. The code below shows the test for
void CMyView::OnLButtonDown(UINT nFlags, CPoint point)
m_LeftButtonDown = TRUE;
m_LeftDownPos = point;
CMyDoc* pDoc = (CMyDoc*)GetDocument();
CTypedPtrList<CObList,CSphere*>& sphereList =pDoc->m_SphereList;
POSITION pos = sphereList.GetHeadPosition();
while (pos != NULL)
CSphere* spheres = sphereList.GetNext(pos);
The problem of the implementation is how much can be the projected area of the object. If you set projected area too big, it will be possible that two different areas can be merged. If projected area is too small it's hard to pick it with the mouse.
- OpenGL Programming Guide The Official Guide to Learning OpenGL, Version 1.1
- Open GL Super Bible (Publisher: Macmillan Computer Publishing) Author(s): Waite group Press ISBN: 1571690735 Publication date: 08/01/96