The image below is a composite of a SpectroCam-10 image at 11.7 microns wavelength (left) taken on February 28, and an HST visible-light image (right), taken almost exactly two martian rotations earlier. The infrared data have been divided by a thermal model to remove the large diurnal heating effect which causes the afternoon (right) side of Mars to be much brighter than the morning side. The polar cap, which appears bright in the visible, is dark in the infrared because it is colder than the equatorial regions. The three bright spots which form an equilateral triangle on the right side of the visible-light image are clouds near the summits of martian volcanoes. They are also visible as dark spots on the infrared image because they are high in the atmosphere, and thus colder than features at the surface.
More very cool (no pun intended, really!) HST images of Mars are available from STScI
Some trivia: The resolution of the HST image is at least 0.1 arcsecond. At 11.7 microns, SpectroCam-10's limiting angular resolution is set by the 5 m diameter of the Hale telescope primary mirror. To achieve an infrared resolution equal to that of HST, we would need a telescope with a primary mirror at least 25 m (over 80 feet) in diameter!