As I understand it from reading Kirsch’s original report (PDF), the scanner used a very sensitive light-detecting tube called a photomultiplier to translate the parts of the image into black or white square pixels. If light was reflected off a scanned spot on the photo, SEAC registered a 0 (white). If no light signal was received, it’d register a 1 (black).
Kirsch has posted a brief video describing the event. He has a surprise guest too.
The National Institute of Standards and Technology Museum has an online exhibit based on a 1998 report by Kirsch.