There are several ways to read relative humidity.
Some sensors uses a capacitance method (condenser). The dielectric is the air, and the capacitance value changes with the humidity of the air. Or some similar design (it's not written in stone).
Another method is a sensor constructed with an inorganic and higroscopic (correct word?) substance. It basically reacts to water vapor. The expansion/contraction of the substance is measured by a moving nucleous inside a bobinated coil (with no friction) which the alteration of fase of the signal through the coil when the nucleous moves , gives the relative humidity. Pretty simple actually.
Number three is even better. A small box with constant air movement , with ambient temperature reading. Inside , a laser is bounced on a mirror surface , which, on the spot the laser hits it, heats up and creates condensation. Then , the distortion of the light caused by the condensation is measured by a optical sensor . There is a relationship on the temperature of the ambient air and the distortion caused by the condensation , gives the relative humidity.
Take your pick
.
Anyway, I was thinking (i do that sometimes). You (BladeRunner) could get a Matrix Orbital (reviewed by pH) and use it's three temperature probes (pretty good ones) . One for the cold air (rad) , one for the hot air (chimney) and one for ambient temps. It's also possible to control 3 exits with the MO (you just need one really), so you could connect them to a simple relay circuit , and the MO has a nice sofware that you can "program behaviors" , and control the heating of the chimney. Sounds nice and (somewhat) simple. Better than my uP ghetto approach.