3D objects are defined in local space and are then all brought into one world space system. To facilitate projection, culling, and other operations, the objects are then transformed to view space, where the camera is positioned at the origin and looking down the positive z-axis. Once in view space, the objects are projected to the projection window. The viewport transformation transforms the geometry on the projection window to the viewport. Finally, the rasterization stage computes the individual pixel colors of the final 2D image.
The 3D object is defined in the model space and is introduced together in a world space system. Simple projection, selection, and other operations, these object models are then converted to visual space, and the visual space begins from the position of the camera and extends along the positive direction of the Z axis. Once visual space is established, these objects are projected into projected forms. The conversion of the perspective is achieved by the geometry in the projection form. Finally, the rasterization steps are completed in the final 2D image independent pixel color.
So, summarizes the modeling of objects -> World Space -> Specific operation (projection, selection, etc.) -> Visual space -> rasterization.
