Essentially the XBee modules form a cableless serial connection which acts as a transparent pass through of commands and data.
The shield you mention in the first reference, aids with interfacing the XBee module with other devices, as the pinouts are only 2mm, not the standard breadboardable 2.54mm.
Of course, if you were to design your own PCB with 2mm header sockets, neither breakout/shield would be required.
Keep in mind, the relatively slow throughput of the serial connection implemented by the XBee modules limits acquisition to single frames, not full motion video.
The camera employed in the first reference has a built-in UART which receives commands to control the camera and transmit the image.
Using a similar setup, an additional microcontroller is not required, only the two XBee modules and the camera with the built-in UART are essential.
BigDog