Meade Image Processing Instrukcja Użytkownika Pobierz pdf (Strona 15)

clipped. It is good practice (whenever doable) to avoid collecting more electrons than that in

any one pixel, and one can do that by limiting the exposure time. However, when we want

to image a faint nebula close to a bright star, we won’t even see the nebula before the star is

already clipped. We must be satisfied with some sort of compromise.

One way to handle this situation is to take repeated shorter exposure images, and then

add them together by software. The adding procedure is quite easy. However, each separate

image introduces extra readout noise (normally ± 7-8 counts), and when we add too many

images, the readout noise adds up. It is rarely beneficial to add together more than ten

images.

At this point image depth becomes important. The count on each pixel is 0, 1, 2, 3, …, a

whole number. The maximum count in a pixel is determined by the image depth. On 8-bit

images, the digital camera standard, the counts go from zero to 255 (=2

-1). This allows

only 256 shades of gray. On such images all stars are clipped long before any deep-sky

objects stick out of the noise. The ST10 camera has 16-bit image depth, the counts can go

from zero to 65,535 (=2

-1). This allows us a very large number of shades of gray, and we

can see details that would be normally lost with a regular camera. You don’t see the

difference between an 8-bit and a 16-bit image directly, but when you increase the contrast

by software, additional detail becomes visible on 16-bit images.

Strictly speaking, the maximum count on an individual camera pixel is only 40,000, so

the image depth is only 14 bit. The camera is only 14 bits deep, but the software can handle

16-bit images. The extra 2 bits come in useful when several images are added together.

Telescope systems can rarely be kept free of dust. The large image depth brings out all

the tiny variations in contrast, so effects of dust shadows, slight pixel-to-pixel variations in

sensitivity are brought out and emphasized on the final image. Unprocessed images appear

full of “dirt”, bright and dark spots, the edges are dark. These must be removed by the

procedure of flat fielding. We take an image of the clear sky, or some other evenly

illuminated object, to obtain a flat, a supposedly even picture with nothing but the

disturbing “dirt” on it. The light frame is then divided by the flat frame, using software, and

the resulting image will be free of the spots and the varying darkness towards the edges.

The camera takes black-and-white images. Color cameras lose much light, and are not

well suited for astronomical imaging. In order to obtain color pictures, we need to place

color filters in front of the camera and take images in three different colors (usually red,

green and blue, RGB). The individual filtered images are processed separately and a color

image is composed at the end.

As all this discussion indicates, the raw images need quite a complex procedure to turn

them into pleasing pictures with as little noise and other disturbances as possible. This

image processing takes up more than half of the time and effort that goes into obtaining

pictures, but it can be done at a convenient time in a convenient place with a computer.

1 2 ... 10 11 12 13 14 15 16 17 18 19 20 ... 28 29

Brak uwag

Meade Image Processing Instrukcja Użytkownika Strona 15