There is 2 ways to image target computers. The first method you already know, but registering the target computer, then select image at the end of the process or schedule a deployment in the web ui. You uses this method if you want FOG to manage the target computer after imaging. In this method the FOG Client is used to name and connect the computer to AD.
There is another method that MSP (system rebuilders) use need to load an OS on the computer, but the computer will not be managed by FOG after imaging. Typically once a system rebuilder loads a computer that computer is put up for sale and is never seen again by the FOG server. You access this method via the Deploy Image iPXE menu.
There is another process that is in between these two methods. Where you use the Deploy Image menu to image the machine, but the system name is caculated in a post install script and the unattend.xml file in windows is used to connect the computer to AD. In this case the fog client isn’t used or needed. If the FOG Client is installed in the image it will contact the FOG server after OSD is done in windows and will be added in a pending state in the web ui. From there you will need to accept the system in the UI and then you can manage the system.
So there are a couple of ways to image a machine, none are totally hands off and some take some setup to complete.