Raster versus image
Raster and image are two terms that are often interchanged.
An image is a generic two-dimensional pictorial representation. It is not dependent on a wavelength or remote-sensing device, such as a satellite, aerial camera, or terrain sensor. An image is displayed on your screen or printed. You view images which represent rasters.
A raster is the data model that describes how an image is stored. A raster defines the pixels (cells) in rows and columns, the number of bands, and the bit depth that compose the image. When you view a raster you are viewing an image of that raster data.
All images are rasters, but not all rasters are considered images. For example, a digital elevation model (DEM) is a cell-based raster dataset, but is typically not considered an image. Other types of rasters that are not considered images per se are magnetic data, interferogram, bathymetric data and other grid-based data sets.
You might also hear rasters being referred to as cell-based datasets, but this is not typically used within ArcGIS documentation.
Cells and pixels
Pixel is often used synonymously with cell. Both cell and pixel refer to the smallest unit of information in raster data. Pixel is an abbreviation for picture element and is often used when describing imagery, whereas cell is often used when describing raster data.
Cells and pixels have a dimension and value. They represent information such as temperature, soil types, elevation, and real-world features, such as parks, lakes, and buildings.
Resolution, scale, and cell size
Resolution, scale, and cell (pixel) size can all refer to how large a feature is in raster data. For example, there are four types of resolution:
Rasters can have one or more bands, or channels, that are designated by a specific wavelength range. Multiband rasters are often referred to as multispectral images, and rasters with up to hundreds of bands are often referred to as hyperspectral images. Multispectral, superspectral, and hyperspectral sensors usually feature bands with non-overlapping wavelength intervals to support reliable analytical image processing. A single-band raster dataset represents a single phenomenon, such as elevation, or only one wavelength range within the electromagnetic spectrum, such as a black-and-white aerial photograph.
Raster format versus raster type
A raster format defines how pixels are stored in a file, such as the number of rows and columns, number of bands, actual pixel values, and other raster format-specific parameters. The raster type helps identify metadata, such as georeferencing, acquisition date, and sensor type, along with a raster format.
Raster products are designed to make adding imagery from specific sensors or data providers to your map easier because each raster product has a unique set of enhancements and band combinations to provide an optimal view of your data. Raster products appear in the Project pane, and utilize the metadata files associated with specific vendor products. It is the information in the metadata files that is used to generate the raster products, such as multispectral or pan-sharpened imagery from satellites such as Landsat.
Raster datasets can be displayed in your map in many different ways. How a raster dataset is displayed depends on what type of data it contains and what you want to show. Some rasters have a predefined color scheme—a color map—that is automatically used to display them. For those that don't, you can adjust the default as needed.
Functions enable you to define and apply on-the-fly processing to one or more rasters, but this processing is not permanently applied to the rasters; it is applied dynamically as the rasters are accessed.
Methods of storage: Data models
A raster dataset is any valid raster format organized into one or more bands. Each band consists of an array of pixels (cells), and each pixel has a value. A raster dataset has at least one band. More than one raster dataset can be spatially appended (mosaicked) together into a larger, single, continuous raster dataset.
A mosaic dataset is a collection of raster datasets (e.g. images) stored in a catalog and viewed or accessed as a single mosaicked image or individual images (rasters). These collections can be extremely large both in total file size and number of raster datasets. The raster data is added according to its raster type—which identifies metadata, such as georeferencing, acquisition date, and sensor type, along with a raster format. The raster datasets in a mosaic dataset can remain in their native format on disk or, if required, be loaded into the geodatabase. The metadata is managed within the raster record as well as attributes in the attribute table. Storing metadata as attributes enables parameters such as sensor orientation data to be managed easily and used to process each raster within the mosaic appropriately. Metadata is also leveraged to enable fast queries for image selection.
Using mosaic in ArcGIS
There are several ways in which the term mosaic is used in ArcGIS: