Removal of Polynomials From Raw Aperture Flux Time Series?

While examining some data about the new discovery of Kepler-22b, I noticed a plotted data set described as a "flux time series after [the] removal of a second-order polynomial for each segment and normalizing the data of each quarter by the median" (see Figure 1 at http://kepler.nasa.gov/Mission/discoveries/kepler2... ). I'd like to better understand what this means. Any guidance, links, or direction would be greatly appreciated.

I understand a flux time series is, of course, a measure of the light arriving from the star as plotted over time but I do not understand the process of removing polynomials and normalizing the data or the value of doing these things. My assumption is that this involves some sort of reverse local regression.