Read about Raster Data Loss
What is a Raster Image?
A raster is a way of storing images in a digital format. In the most simplest terms, it is coloring squares on a grid. Generally the squares(though occasionally rectangles) are called pixels.
Since a computer can't actually understand the concept of a color, we have to give it some hard data instead. Every color the human eye can see can be represented as a combination of three primary colors: red, green, and blue. So to store images on a computer, we represent each color as a set of three numbers each one corresponding to a color. When your computer recombines the red, green, and blue information on your display, the three pixels (one of each color) can reproduce all the other colors we see. This is mostly commonly known as the RGB system, and it usually consists of one byte of data (the numbers 0 through 255) for each color for a total of 3 bytes per pixel. Since we have limited combinations of colors (256 x 256 x 256), this system limits us to around 16 million different colors.
Storing Data Instead of Colors
The MODIS images available from MichiganView are available in three types. Instead of placing normal color information in the pixels, we've put in information about light that humans CAN'T see.
|Three Available Combinations|
|143||Visible light - Red, Green, and Blue are all stored like normal|
|214||Near Infrared - NIR stored in the red pixels, Red in the green, and Green in the blue|
|721||Midrange Infrared - MidIR stored in the red pixels, NIR in the green, and Red in the blue|
Because the images are storing data about other wavelengths (not just the visible ones) in the images, opening the 214 and 721 images displays a strange looking picture. For example, opening the 214 image will produce something like this:
No one spilled red paint on Michigan, it's simply that Michigan is radiating a large quantity of infrared radiation which has been stored in the pixels instead of the red color information.
Indexed or Palleted Images
There are different ways to store the image information. Some image formats forgo the entire color spectrum in favor of a smaller subset of colors used. Generally, this format uses 1 byte or 256 colors, each chosen as a representative of a subset of the true colors in an image. The advantage of this format is that this image can be stored in a much smaller size, but can also result in a substantial loss of quality.
For small pictures, storing the image as is will probably be fine. But if you start working with more or larger images, they can quickly devour your hard drive. Another option for reducing the size of an image (instead of palleting) is compression. Here's a comparison of the number of images, compressed and uncompressed, you can fit on a two gigabyte flash drive:
|22.2 Mb||MichiganView with 250m pixel||none|
|204||10Mb||MichiganView with 250m pixel|
|3495||600Kb||MichiganView with 250m pixel||JPEG|
It should be obvious that compression allows us to store either more pictures or a larger area in a much smaller space, but whats the catch? Depending on the type of compression you are using, the cost is image quality. If you zoom in on the JPEG, you will start to notice blocky splotches, called artifacts. (You can see this clearly near Garden Island, West of the bridge) On the TIF, it just gets blurry (or pixelated depending on your image viewer).
Lossy compression means that data is lost when the image is compressed, there is no way to revert from a compressed image back to the original. This also means that every time you compress an image while working on it, you lose a bit more data, over time this could destroy your image. JPEG is a lossy compression.
Lossless compression takes advantage of repetition in an image. In essence, it is a way store the same data more efficiently. TIF format uses a compression called LZW that is lossless. The 250m TIF above is 17Mb using an uncompressed format, or a savings of 7Mb without loss of image quality. However, compression ratio's vary widely image to image and may not always cut the image down enough for it to be manageable.
So far we've mentioned JPEG and TIF, but these two are only a small fraction of the formats out there. Luckily, most images are in a handful of commonly used formats:
|Format||Compression Type||Compression Ratio||Notes|
|GIF||palleted||image dependent||Supports transparency; bad when precise color is important|
|JPEG||lossy||user variable||Best for photo-like images when size is more important than quality|
|PNG||lossless||image dependent||Successor the of the GIF format; supports transparency|
|TIF||lossless||image dependent||Container format that allows storage of other data with the image|
For a far more in-depth overview of common format types, click here.