All About Space Astronomer Book - 2014 UK

Astronomy with your camera

used – lossless and lossy. Lossless preserves all of the original image data

while lossy uses special algorithms to remove selected information from

the original image in order to reduce data. While this loss of information

doesn’t normally make a huge visual impact on the initial reconstruction of

the image, such images are no longer viable for scientific analysis. Opening

a lossy compressed image and resaving it using a similar lossy format

increases the loss of data from the original. Over several open-save cycles,

the visual impact of this action can become noticeable.

Common Image File Formats

JPEG (extension .jpg, uses lossy compression)

Joint Expert Photographic Group format. Commonly used and adopted by

general-purpose digital cameras. Level of compression can be defined.

JPEG is limited to 8-bits-per-channel colour.

JPEG2000 (extension .jp2/.jpx, lossy and lossless compression)

An update to JPEG with greater storage flexibility. Not as common as JPEG

due to greater computing requirements to decode/encode the format.

TIFF (extension .tif, typically lossless compression but can also support no

compression and lossy compression)

The Tagged Image File Format isn’t strictly a format in its own right but

rather a transport framework for image data. The header of a TIFF file

contains information on how the data is stored. Data may be uncompressed

or compressed. The type of compression may vary too.

GIF (extension .gif, lossless format)

Graphics Interchange Format (GIF) images were once the mainstay of

the internet. Supporting 8-bit colour, GIFs were also capable of providing

animation capabilities which made them popular for website design. GIFs

are still used in astronomy for simple "flick-book"-type image animations.

PNG (extension .png, lossless)

The Portable Network Graphic format is a popular compressed format used

for internet images. It originally came into being to replace the commonly

used GIF format which became subject to licensed use.

One downside to ABGs is that when a star is over-exposed because
the well receiving its light becomes full, it is not possible to make accurate
measurements of the star’s relative brightness beyond that point. As a
consequence, sensors fitted with ABG technology are not useful for scientific
brightness measurements unless the image remains under-exposed.

Colour

Imaging chips work as greyscale devices producing images which contain
black, white and the grey tones in between. Although there are alternatives,
most colour cameras commonly produce colour results by virtue of a Bayer
colour filter matrix placed over the image sensor. This places a red, blue or
green filter over each photosite and is arranged in a repeating pattern of 2 x
2 containing two green, one red and one blue filter. The resultant greyscale
image is processed by a de-Bayering routine which analyses the relative
values that each photosite contains and compares this to its neighbours.
With this information and knowing which colour filter was covering which
photosite, the colour information of the scene can be restored.

Image Files

Once the photosite values have been read off a sensor and passed through
a de-Bayering routine, the resulting data is stored in an image file. The
photosite data that finally emerges from a colour camera has information
on the tone and colour of that part of the image. These values are what are
known as picture elements or pixels.
Image pixel data and other pertinent data are stored in an image file. The
make-up of an image file is determined by the file format and there are
many different types of these in existence. At the most basic level, an image
file will contain an image header which holds information about the image,
the camera settings used to take it and how the image data has been stored.
The bulk of the image file will be the image data itself. In order to reconstruct
the original image from the image file, a computer program that’s aware of
the file format will read the header and interpret the image data accordingly.
Image data can be stored in an uncompressed or compressed way.
Uncompressed image-data files can be quite large so compressed formats
tend to be popular, especially as digital imaging can quite easily generate
thousands of images every year. There are two types of compression

[2] Colour images are produced
from a greyscale sensor by
overlaying a filter grid known as
a Bayer matrix.

Bayer Filter Matrix

Sensor Photosites

2