Given a 2.4Ghz pure sine wave and a couple signal strength detectors couldn't one iteratively modify the frequency dependent values for the various walls?
Not for any practical measurements. This would work in say an anechoic environment – save for the wall under test – but you'd end up sampling multipath constructive/destrictuve interference instead of the wall reflections.