I do not think that this is beyond the capabilities of the Arduino if he uses digital detectors. However you are right. Unless the desired time resolution is known it is unclear if this is a simple or a hard task.
I am pretty sure it isn't; back when I went to High Tech Institute (I know, cheesy name) here in Phoenix, in 1991-92 - we built a pattern recognizer system using 64 phototransistors and a multiplexing system not unlike the LED grid projects you see people doing with the Arduino - just rather than driving the columns/rows of LEDs, we were switching 8 columns at a time per row to read the values via the parallel port of an Amiga 1000 (on which we had to write a 68000-based assembler code to read the port, and output whether the "mask" character we placed on the array matched a character in memory - crude form of OCR).
That Amiga was only running at 7 MHz. We did have more memory at our disposal, though. Even so, I don't think this project is out of the realm of possibility of working using an Arduino; maybe this weekend I can dig out my old papers and schematic of that system and post it here, if it would help.
On a different note, one guy for his "final project", with the help of an instructor - built and set up (breadboard style - rat's nest city!) a laser rangefinder using a linear array of LEDs as photodetector elements. The laser was an HeNe tube-based laser the school had (this was before cheap LED laser diode modules - such modules were available, but well outside our's or the school's price range); the rest of the system would detect the bounce of the laser off a mirror, with it's distance changing the angle the reflected beam back to the LEDs. Based on which LED was "lit" by the laser, a complex multiplexing and counter circuit would determine how to light up a pair of 7-segment LED arrays to form an output number, which was the "distance". No computer, microcontroller, or anything like that involved at all - just a bunch of breadboards, wires, and TTL logic. The guy got an "A" of course.