Data Storage

Producer Field Guide

HGD_Product
Producer Field Guide
HGD_Portfolio_Suite
Producer

Image data can be stored on a variety of media—CD-ROMs or DVD-ROMs for example—but how the data are stored (for example, structure) is more important than on what they are stored.

All computer data are in binary format. The basic unit of binary data is a bit. A bit can have two possible values—0 and 1, or "off" and "on" respectively. A set of bits, however, can have many more values, depending on the number of bits used. The number of values that can be expressed by a set of bits is 2 to the power of the number of bits used.

A byte is 8 bits of data. Generally, file size and disk space are referred to by number of bytes. For example, a PC may have 2 gigabytes of RAM (random access memory), or a file may need 55,698 megabytes of disk space. 1,024 bytes = 1 kilobyte. A megabyte (Mb) is about one million bytes. A gigabyte (Gb) is about one billion bytes.

Storage Formats

Image data can be arranged in several ways on a tape or other media. The most common storage formats are:

  • BIL (band interleaved by line)
  • BSQ (band sequential)
  • BIP (band interleaved by pixel)

For a single band of data, all formats (BIL, BIP, and BSQ) are identical, as long as the data are not blocked.

BIL

In BIL (band interleaved by line) format, each record in the file contains a scan line (row) of data for one band (Slater, 1980). All bands of data for a given line are stored consecutively within the file as shown in the figure below.

Band Interleaved by Line (BIL)

data_storage_bil_diagram

Although a header and trailer file are shown in this diagram, not all BIL data contain header and trailer files.

BSQ

In BSQ (band sequential) format, each entire band is stored consecutively in the same file (Slater, 1980). This format is advantageous, in that:

  • one band can be read and viewed easily, and
  • multiple bands can be easily loaded in any order

Band Sequential (BSQ)

data_storage_bsq_diagram

Landsat TM data are stored in a type of BSQ format known as fast format. Fast format data have the following characteristics:

  • Files are not split between tapes. If a band starts on the first tape, it ends on the first tape.
  • An end-of-file (EOF) marker follows each band.
  • An end-of-volume marker marks the end of each volume (tape). An end-of-volume marker consists of three end-of-file markers.
  • There is one header file per tape.
  • There are no header records preceding the image data.
  • Regular products (not geocoded) are normally unblocked. Geocoded products are normally blocked.

ERDAS IMAGINE imports all of the header and image file information.

See Geocoded Data for more information on geocoded data.

BIP

In BIP (band interleaved by pixel) format, the values for each band are ordered within a given pixel. The pixels are arranged sequentially on the tape (Slater, 1980). The sequence for BIP format is:

Pixel 1, Band 1

Pixel 1, Band 2

Pixel 1, Band 3

...

...

...

Pixel 2, Band 1

Pixel 2, Band 2

Pixel 2, Band 3

...

...

...

Storage Media

Most raster data are available on a variety of storage media to meet the needs of users, depending on the system hardware and devices available. When ordering data, it is sometimes possible to select the type of media preferred. Some common forms of storage media are discussed in the following section:

  • CD-ROM / optical disk
  • DVD-ROM

Other types of storage media are:

  • flash memory drive
  • film, photograph, or paper
  • videotape
  • web-based services

CD-ROM

A CD-ROM is an optical read-only storage device that can be read with a CD player. Up to 644 MB can be stored on a CD-ROM. Because this device is read-only, the CD-ROM protects the data from accidentally being overwritten, erased, or changed from its original integrity.

DVD-ROM

DVD-ROM is an optical disk storage device which is read by a DVD drive in a computer. A single-sided, one-layered disk has 4.7 GB storage capacity. DVDs are available in single-sided or double-sided format, and each side can have one or two layers. Double-sided, single-layer, and single-sided, double-layer DVDs can store about 8.5 GB. Double-sided, double-layer DVDs can store about 15.9 GB.

Calculating Disk Space

To calculate the amount of disk space a raster file requires on an ERDAS IMAGINE system, use the following formula:

data_storage_calculating_disk_space_formula

Where:

y = rows

x = columns

b = number of bytes per pixel

n = number of bands

1.4 adds 30% to the file size for pyramid layers and 10% for miscellaneous adjustments, such as histograms, lookup tables, and so forth.

This output file size is approximate. See Pyramid Layers for more information.

For example, to load a 3 band, 16-bit file with 500 rows and 500 columns, about 2,100,000 bytes of disk space is needed.

[((500 x 500) x 2) x 3] x 1.4 = 2,100,000 b or 2.1 Mb

Bytes Per Pixel

The number of bytes per pixel is listed below:

4-bit data: .5

8-bit data: 1.0

16-bit data: 2.0

On the PC, disk space is shown in bytes. On the workstation, disk space is shown as kilobytes (1,024 bytes).

ERDAS IMAGINE Format (.img)

In ERDAS IMAGINE, file name extensions identify the file type. When data are imported into ERDAS IMAGINE, they are converted to the ERDAS IMAGINE file format and stored in image files. ERDAS IMAGINE image files (.img) can contain two types of raster layers:

  • thematic
  • continuous

An image file can store a combination of thematic and continuous layers, or just one type.

Image Files Store Raster Layers

data_storage_erdas_imagine_format

ERDAS Version 7.5 Data

When importing a GIS file from ERDAS Version 7.5, it becomes an image file with one thematic raster layer. When importing a LAN file, each band becomes a continuous raster layer within an image file.

Thematic Raster Layer

Thematic data are raster layers that contain qualitative, categorical information about an area. A thematic layer is contained within an image file. Thematic layers lend themselves to applications in which categories or themes are used. Thematic raster layers are used to represent data measured on a nominal or ordinal scale, such as:

  • soils
  • land use
  • land cover
  • roads
  • hydrology

Thematic raster layers are displayed as pseudo color layers.

Example of a Thematic Raster Layer

raster_layer_thematic

See Image Display for information on displaying thematic raster layers.

Continuous Raster Layer

Continuous data are raster layers that contain quantitative (measuring a characteristic on an interval or ratio scale) and related, continuous values. Continuous raster layers can be multiband (for example, Landsat TM data) or single band (for example, SPOT panchromatic data). The following types of data are examples of continuous raster layers:

  • Landsat
  • SPOT
  • digitized (scanned) aerial photograph
  • DEM
  • slope
  • temperature

Continuous raster layers can be displayed as either a grayscale raster layer or a true color raster layer.

Examples of Continuous Raster Layer - True Color and Grayscale

raster_layer_continuous_sat_dem

Tiled Data

Data in the .img format are tiled data. Tiled data are stored in tiles that can be set to any size.

SHARED Tip The default tile size for image files is 512 × 512 pixels.

Image File Contents

The image files contain the following additional information about the data:

  • data file values
  • statistics
  • lookup tables
  • map coordinates
  • map projection

SHARED Tip View this additional information using Image Metadata dialog.

Statistics

In ERDAS IMAGINE, the file statistics are generated from the data file values in the layer and incorporated into the image file. This statistical information is used to create many program defaults, and helps you make processing decisions.

Pyramid Layers

Sometimes a large image takes longer than normal to display in the Viewer. Use the pyramid layer option to display large images faster. Pyramid layers are image layers which are successively reduced by the power of 2 and resampled.

See Image Display for more information on pyramid layers. See Help for information on ERDAS IMAGINE file formats.