[Moon Double Comb]

Yes, it's due to observation, but that isn't necesarrily the same as photons interacting with a particle. It's because it's fundamentally impossible to gain information without a collapse of the wave-function. Interacting photons are just a tiny bit of the whole story.

In technical language: Every observable is represented by an Hermitian operator (one of the axioms of QM). Now the measurement postulate says that when you measure (observe), you will get an eigenvalue of the operator. What happens is that the wavefunction collapses to the eigenfunction of the operator.

Now if two operators don't commute you can derive the uncertainty principle. There's not only the momentum/position uncertainty, but for every two observables that don't commute there is an uncertainty relation.

So while the uncertainty principle is due to observation it can be derived from the postulates of Quantum Mechanics. No interacting photons are needed to derive this relation. It is because:
-Every particle can be represented by a vector |psi> in a Hilbert space;
-If the particle is in state |psi>, measurement of a variable will yield one of the eigenvalues and the state of the system will change from |psi> to the eigenvector corresponding with the measured eigenvalue.

I know this is all technical, but I wanted to say that from the postulates of QM, the uncertainty principle can be derived. Now, photons interacting with a particle might be an easy way to visualize it, but it's oversimplyfied and does not tell the whole story at all.

But, to make it absolutely clear, the uncertainty principle is indeed due to measurement.