I wonder if you have read the GL Tutorial posted, but what you're asking is really basic.

. You might not have base knowledge on 3D maths and programing, I wonder.
However,
The very basic idea is triangles. Good GPU pumps triangles onto screen nice and fast.
The idea is vertices and indices. You pump those buffers into OpenGL buffer, call draw method and it will draw stuffs on the surface.
To draw image at an XYZ, the idea is billboarding. Billboards are just 3D rectangles ( quads ).
Ray intersection is the idea of how to get coordinate from touch event. From touch event, you only get a 2D coordinate, from which you'll get a ray. With ray you can test intersection with 3D objects.
It might take you a long time to understand and make it yourself those stuffs, or you can wait an Android 3D engine.