These look like MELF package resistors as others have said.
It's hard to tell the exact colour of the rings from the photos. To me they all appear to have five coloured rings, brown black black red (slightly wider gap) brown, although some are fitted "back to front". This would make them 10K with 1% tolerance. Other evidence supports this:
Note that some of the other components (like the small surface mount resistors next to these) don't have their orientation controlled either. Uncontrolled orientation seems a far more likely explanation than that there are two different values used here that just happen to have the same band colours but in reverse.
When measuring these in place on the board you are measuring the resistance in parallel with "something". As you measure 5.6K, these resistors must be larger than 5.6K. They can't be 10 or 12 ohms as stated in another answer, or 120 ohms (which is what they would be if I got the band colours right but the direction wrong i.e. brown red black black brown).
These are also unlikely to be 12K as stated in another answer. 12K would read brown red black red X or brown red orange X.
Finally, these are unlikely to be inductors if you are reading a resistance across them as high as 5.6K - you might see a few ohms from a small inductor but not kilo-ohms.
Following the tracks in the photos, I think this is the circuit diagram for the input. The mystery components are the two resistors at the left, R1 and R2.
I think these resistors are designed to handle voltage spikes in the input signal. MELF packages can usually handle about 1W of power and are usually good with short surges above that. A 10K resistor with 100V across it will dissipate about 1W. These are probably capable of handling short spikes up to 200V or so.
This will be why the MELFs were chosen instead of using the same ceramic 10K resistors used elsewhere on the board (marked 103) which can usually only handle about a tenth to a quarter of a watt.
A high voltage spike at the input will take two routes. Half the current will go through R1 to ground. The other half will go through R2 and start charging up the capacitor C1. The values will have been chosen so that even the longest spike expected won't be able to charge C1 to higher than about 5V which is the maximum the ATMEGA128 can handle.
It's hard to see exactly what value R3 has (they're the little black ceramic resistors between the MELFs on the top side and the little brown capacitors) but it will be there to set a maximum frequency that can pass through this input filter i.e. to cut out high frequency noise in the input signal.
R4 is just to limit the maximum current that can flow into the microcontroller in the event a spike changes the voltage at C1 quickly.
After R4 the signal goes to a via under the Atmega128 microcontroller, so it probably goes straight to one of the microcontroller's input pins.