Segment organization

PCIDSK is a data structure for storing digital images and related data, such as lookup tables (LUT), spectral signatures, ground control points (GCP), and other data types. Each PCIDSK database is a separately named disk file. You can have any number of PCIDSK files, limited only by the disk capacity of the computer. PCIDSK files can be shared on various computer platforms.

Segments are the parts of a PCIDSK database which store data related to the imagery in the database. Unlike image channels, disk space is not allocated for segments at the time the database is created; rather, disk space is allocated dynamically whenever a segment-generating program is run. A database can store up to 1,024 segments, provided you have enough disk space.

The following types of information are stored as segments:

Bitmaps [Type 101:BIT]

Bitmap segments are raster layers and have the same pixel and line dimensions as the image data on the PCIDSK file. Each pixel is 1-bit deep with a value of 0 or 1. This means that the gray value of a bitmap pixel can be either zero (pixel is "off") or 1 (pixel is "on"). The georeferencing that is associated with the raster layers is assumed to apply to bitmap layers. One example of using bitmap segments is to mask images when creating training areas during classification.

Vectors [Type 116:VEC]

Vector layers or segments store a set of related vectors and can be points, polylines (arcs), whole polygons, topological polygons, or additional database records. They are used to represent features, such as river networks, road networks, and political divisions.

Vector objects are called shapes. A shape has a unique numeric identifier (ShapeID) greater than or equal to zero. Each shape has an associated record of attributes stored in the layer as part of the shape. Each shape has a list of vertices that have an x, y, and z value. Vertices are coordinates in the vector georeferencing system.

Signatures [Type 121:SIG]

Signature segments store statistical data that describe the spectral behavior of a particular image feature (object). Signature segments are used as input during minimum distance, parallelepiped, and maximum-likelihood classification.

Text [Type 140:TEX]

Georeferencing [Type 150:GEO]

Georeferencing segments store mathematical transforms that map all pixel locations in the database image to a georeferenced coordinate system, such as UTM.

Orbit [Type 160:ORB]

Orbit segments store satellite ephemeras data, used in the orthorectification process. This information is read from the original data source, or from a text file.

Lookup table [Type 170:LUT]

Lookup-table (LUT) segments store numerical tables that map image gray levels to new gray levels.

Pseudocolor table [Type 171:PCT]

Pseudocolor (PCT) segments store numerical tables that map image digital number (DN) values to a specific color. Colors are defined by an intensity value (between 0 and 255) for each of a red, green, and blue component.

Binary [Type 180:BIN]

Binary segments are created and used by PCI programs. These contain internal-only information, such as orthorectification models and neural networks.

Array [Type 181:ARR]

Array segments store an array of numbers. They are created and used by PCI programs. They contain information such as SAR-gain offsets and gain-scaling tables.

System [Type 182:SYS]

The system segment is used to store binary information that you should not normally access directly. System segments are not displayed by utility programs and are used to store information such as metadata, overviews, and tiled-image data.

Ground control points [Type 214:GCP]

Ground control point (GCP) segments store pairs of x and y coordinate positions that locate a point in an image and some other frame of reference, such as a vector segment, a map, another image, or keyboard input of coordinates for image correction, registration, and mosaicking.

Each segment has a numeric-type code. For example, GCP segments are type 214. This numeric code helps you locate a particular segment in a database. When listing segments with the program ASL, you can create a listing by segment type.

Pseudocolor tables

A PCT segment contains an array of 256 colors and assigns color values to 8-bit images. A PCT always contains exactly 256 entries. File formats, such as TIFF, which may have color tables with less than 256 colors, assign a value of zero to unused colors. There is no support for alpha channels in the table.

Lookup tables

The LUT segment consists of 256 8-bit values, between 0 and 255. It is used to apply enhancements to raster data. It can also be used to encode thematic-class mapping.

Ground control points

GCP segments are used to associate projection coordinates with locations on an uncorrected raw image.

GCPs can also be used to relate locations in any two georeferencing systems: in raw-image coordinates and in a georeferencing-projection system. The definitions of the georeferencing systems are kept as 16-character map unit strings.

Projection parameters are not stored. Only projections that are fully defined by the map-unit string can be used; for example, UTM, Long/Lat, or METRE. Complex projections, such as Transverse Mercator, cannot be used.

Each control point can have an elevation associated with the location in an image. An elevation-unit string is kept for each system with values in meters or feet.

GCPs have the following associated values.