To drive home the scientific principles that serve as the bases for interpreting infrared satellite imagery, please take a little time to explore this nifty interactive tool (you can resize the window to best fit your screen). Click on the link to open the tool and keep it open as you read further. Just so that you have your bearings, please note that the controls for cloud thickness, cloud-top height and surface temperature on the right cross section of the atmosphere. After you set the controls in the right window, the left window will show you the corresponding infrared image that corresponds to your settings in the right window. Please keep in mind that the window on your left displays a simulation of what the real infrared image would look like.
Let's try a couple of examples. First, set the ground temperature to an arbitrary 20 degrees Celsius. Next, set the cloud depth to "thick". Okay, drag the cloud up and down (thus decreasing and increasing the temperature of the cloud top). Note that, as you drag thick the cloud upward into the increasingly colder upper troposphere, the appearance of the overcast region gets progressively brighter and whiter on the simulated infrared image (conversely, dragging the cloud downward into the warmer lower troposphere causes the overcast region looks increasingly gray on the simulated IR image).
Okay, ready to try a second example? Keep the ground-temperature setting at 20 degrees Celsius. Next, move the cloud (still on the thickest setting) so that its base lines up fairly closely with minus five degrees Celsius (the cloud has a low, warm base). Then change the thickness of the cloud to "thin" (the top of the now "skinny" cloud stays at the same altitude as it did on the thick setting). Note that the appearance turns darker. That's because radiation from the warmer earth's surface likely leaks through the thin cloud, making its cloud top appear warmer than it actually is. Are you getting the hang of it now?
Time for one last example. First, set the ground temperature to minus ten degrees Celsius (14 degrees Fahrenheit, which would correspond to a temperature on a cold, winter night). Next, set the thickness of the cloud to medium -- halfway between thick and thin. Drag the "medium cloud" so that its top lines up with minus 10 degrees. What do you observe on the simulated infrared image? You're right -- except for the labels, you really can't detect the presence of the low clouds because the temperature of their tops is approximately the same as the temperature of the ground. That's one of the limitations of infrared imagery -- it cannot be used to effectively detect low clouds at night.
Please experiment more with this interactive tool. Hopefully, it will clear up any and all questions pertaining to infrared imagery.