All About Image Callibration on Maxlm DL
Recently we have been using a program called Maxim DL to perform a process called calibrating images. We use “darks” and “flats” along with the actual image of the space object in order to get the highest quality we can achieve with the tools we have. A dark calibration image is like taking a picture with the lens cap on. It captures only the "noise" made by the camera itself, like the buzzing or hissing you might hear from electronic devices. We use these kinds of images to figure out how much noise the camera adds to the real pictures we take. By subtracting this noise from our actual photos, we make them clearer and prettier to look at. Now, for a flat image, think about taking a photo of a blank wall, but the light isn't even across the whole picture. For example, one corner might look brighter than the other. A flat calibration image is like taking a photo to see where the light isn't quite right. We can then use this photo to level out and fix the lighting in our real pictures, making sure they look even from one side to the other. When creating colorized images of space objects, we use a red filter photo, a blue filter photo, a green filter photo, and a clear/luminance filter photo. Each of the color photos has numerous amounts of flats, darks, and actual images. For the space objects the we calibrated in this class, I would say there was an average of about 50 images per color image that required multiple rounds of calibration. When we have a bunch of dark or flat calibration images, we need to combine them to make one good calibration picture. We usually choose the "median" combine mode because it helps us deal with any weird or out-of-place spots in our pictures. For example, if there is a spec of dust on the camera, calibrating with the median will not make the spec extremely noticeable, in fact it may just remove it all together. Think about one of your friends blinking while taking a group photo. If we use the "average" combine mode, their blink might make everyone's face look strange. But with the "median" mode, we ignore that blink and just focus on everyone's eyes being open, making for a better final picture. The median is like focusing on the positive, good looking parts of the picture rather than the things that may be out of place.
