I want to build an AR application on Android by opencv and opengl. I found GL_MODELVIEW can place camera and model in world coordinates like the following link I tried to load a simple model-view matrix by gl.glLoadMatrixf(newMat, 0); and it works so I want to draw 3d model on the top of my target image if i got the four corner points of the target image because model-view matrix is needed I want to know if I can obtain the camera view matrix by opencv's findhomography