Raster Data from Other Software Vendors

Producer Field Guide

HGD_Product
Producer Field Guide
HGD_Portfolio_Suite
Producer

You can use ERDAS IMAGINE to import data created by other software vendors. This way, if another type of digital data system is currently in use, or if data is received from another system, it easily converts to ERDAS IMAGINE file format for use in ERDAS IMAGINE.

Data from other vendors may come in that specific vendor’s format, or in a standard format which can be used by several vendors. The Import and Direct Read functions handle these raster data types from other software systems:

  • GRID and GRID Stacks
  • JFIF (JPEG)
  • JPEG2000
  • MrSID
  • Sun Raster
  • TIFF and GeoTIFF

SHARED Tip Other data types might be imported using Generic Binary import option.

Vector to Raster Conversion

Vector data can also be a source of raster data by converting it to raster format.

ERDAS Ver. 7.X

ERDAS Ver. 7.X series was the predecessor of ERDAS IMAGINE software. The two basic types of ERDAS Ver. 7.X data files are indicated by file name extensions:

  • .LAN—a multiband continuous image file. The name is derived from the Landsat satellite.
  • .GIS—a single-band thematic data file in which pixels are divided into discrete categories. The name is derived from geographic information system.

.LAN and .GIS image files are stored in the same format. Image data are arranged in a BIL format and can be 4-bit, 8-bit, or 16-bit. ERDAS Ver. 7.X file structure includes:

  • a header record at the beginning of the file
  • data file values
  • a statistics or trailer file

SHARED Tip When you import a .GIS file, it becomes an image file with one thematic raster layer. When you import a .LAN file, each band becomes a continuous raster layer within the image file.

GRID and GRID Stacks

GRID is a raster geoprocessing program distributed by Environmental Systems Research Institute, Inc. (Esri) in Redlands, California. GRID is designed to complement the vector data model system, ArcInfo is a well-known vector GIS that is also distributed by Esri. The name GRID is taken from the raster data format of presenting information in a grid of cells.

The data format for GRID is a compressed tiled raster data structure. Like ArcInfo Coverages, a GRID is stored as a set of files in a directory, including files to keep the attributes of the GRID.

Each GRID represents a single layer of continuous or thematic imagery, but it is also possible to combine GRIDs files into a multilayer image. A GRID Stack (.stk) file names multiple GRIDs to be treated as a multilayer image. Starting with ArcInfo version 7.0, Esri introduced the STK format, referred to in ERDAS software as GRID Stack 7.x, which contains multiple GRIDs. GRID Stack 7.x format keeps attribute tables for the entire stack in a separate directory, in a manner similar to that of GRIDs and Coverages.

JFIF (JPEG)

JPEG is a set of compression techniques established by the Joint Photographic Experts Group (JPEG). The most commonly used form of JPEG involves Discrete Cosine Transformation (DCT), thresholding, followed by Huffman encoding. Since the output image is not exactly the same as the input image, this form of JPEG is considered to be lossy. JPEG can compresses monochrome imagery, but achieves compression ratios of 20:1 or higher with color (RGB) imagery, by taking advantage of the fact that the data being compressed is a visible image. The integrity of the source image is preserved by focusing its compression on aspects of the image that are less noticeable to the human eye. JPEG cannot be used on thematic imagery, due to the change in pixel values.

There is a lossless form of JPEG compression that uses DCT followed by nonlossy encoding, but it is not frequently used since it only yields an approximate compression ratio of 2:1. ERDAS IMAGINE only handles the lossy form of JPEG.

While JPEG compression is used by other file formats, including TIFF, the JPEG File Interchange Format (JFIF) is a standard file format used to store JPEG-compressed imagery.

Read more about JPEG standard at http://www.jpeg.org/jpeg/index.html.

JPEG2000

JPEG2000 compression technique is a form of wavelet compression defined by International Organization for Standards (ISO). JPEG2000 provides both a lossy and lossless encoding mode, with the lossless attaining only relatively low compression ratios but retaining the full accuracy of the input data. The lossy modes can attain very high compression ratios, but may possibly alter the data content. JPEG2000 is designed to retain the visual appearance of the input data as closely as possible even with high compression ratio lossy processing.

J2I files (.j2i) are separate index files generated in conjunction with certain JP2 files, typically very large files. These index files are designed specifically to lower memory usage by decompressing the block index to disk, rather than holding it in memory. In the case of ERDAS APOLLO Essentials or Image Web Server where hundreds of images may be opened at once, this allows the server to open significantly more images before memory is exhausted. For a desktop application that is viewing one image, memory usage is not a problem.

Beginning in IMAGINE 11.0.5, j2i extension files are generated only for certain types of files to speed up decoder processing time. These index files are used only by products that implement the ERDAS-Intergraph ECW Compression (ECWJP2 SDK) product.

Specify Output File Size

The concept of ECW and JPEG2000 format compression is that you are compressing to a specified quality level rather than a specified file size. Choose a level of quality that benefits your needs, and use the quality levels option to compress images within that quality range.

The goal is visual similarity in quality levels amongst multiple files, not similar file size. Using a file-sized based compressor, the resultant image quality of each file compressed will be different (depending on file size, image features and so forth), and you will notice visible quality differences amongst the output images. Visible quality differences would not be the goal for an end product such as air photos over a common area.

Currently there is no way of directly controlling the output file size when compressing images using the ECW or JPEG2000 file format and ERDAS products and SDKs. This is because the output size is affected not only by the compression algorithm used but also by the specific low‑level character of the input data. Certain kinds of images are simply easier to compress than others.

Quality Levels

JPEG and JPEG2000 quality value ranges from 1 (lowest quality, highest compression) to 100 (highest quality, lowest compression). Values between 50 and 95 are normally used. Specifying a quality value of 100 creates a much larger image and slight increase in quality compared to a quality value of 95.

Numerically Lossless Compression

When compressing to JPEG2000 format, which supports lossless compressed images, numerically lossless compression is specified by selecting a target compression ratio of 1:1. This does not correspond to the actual compression rate, which will generally be higher (between 2:1 and 2.5:1).

Read more about JPEG2000 standard at http://www.jpeg.org/jpeg2000.

MrSID

Multiresolution Seamless Image Database (MrSID, pronounced Mister Sid) is a wavelet transform-based compression algorithm designed by LizardTech, Inc. in Seattle, Washington. The novel developments in MrSID include a memory efficient implementation and automatic inclusion of pyramid layers in every data set, both of which make MrSID well-suited to provide efficient storage and retrieval of very large digital images.

The underlying wavelet-based compression methodology used in MrSID yields high compression ratios while satisfying stringent image quality requirements. The compression technique used in MrSID is lossy (that is, the compression-decompression process does not reproduce the source data pixel-for-pixel). Lossy compression is not appropriate for thematic imagery, but is essential for large continuous images since it allows much higher compression ratios than lossless methods (for example, the Lempel-Ziv-Welch, LZW, algorithm used in the GIF and TIFF image formats). At standard compression ratios, MrSID encoded imagery is visually lossless. On typical remotely sensed imagery, lossless methods provide compression ratios of perhaps 2:1, whereas MrSID provides excellent image quality at compression ratios of 30:1 or more.

Read more about MrSID compression standard at LizardTech website at http://www.lizardtech.com.

SDTS

Spatial Data Transfer Standard (SDTS) was developed by USGS to promote and facilitate the transfer of georeferenced data and its associated metadata between dissimilar computer systems without loss of fidelity. To achieve these goals, SDTS uses a flexible, self-describing method of encoding data, which has enough structure to permit interoperability.

For metadata, SDTS requires a number of statements regarding data accuracy. In addition to the standard metadata, the producer may supply detailed attribute data correlated to any image feature.

SDTS Profiles

SDTS standard is organized into profiles. Profiles identify a restricted subset of the standard needed to solve a certain problem domain. The SDTS Raster Profile and Extensions (SRPE) covers gridded raster data. This is imported as SDTS Raster.

Read more about SDTS at http://mcmcweb.er.usgs.gov/sdts.

SUN Raster

A SUN Raster file is an image captured from a monitor display. In addition to GIS, SUN Raster files can be used in desktop publishing applications or any application where a screen capture would be useful.

There are two basic ways to create a SUN Raster file on a SUN workstation:

  • use the OpenWindows Snapshot application
  • use UNIX screendump command

Both methods read the contents of a frame buffer and write the display data to a user-specified file. Depending on the display hardware and options chosen, screendump can create any of the file types listed in the following table.

File Type

Available Compression

1-bit black and white

None, RLE (run-length encoded)

8-bit color paletted (256 colors)

None, RLE

24-bit RGB true color

None, RLE

32-bit RGB true color

None, RLE

Data are stored in BIP format.

TIFF

TIFF was developed by Aldus Corp. (Seattle, Washington) in 1986 in conjunction with major scanner vendors who needed an easily portable file format for raster image data. Today, the TIFF format is a widely supported format used in video, fax transmission, medical imaging, satellite imaging, document storage and retrieval, and desktop publishing applications. In addition, GeoTIFF extensions permit TIFF files to be geocoded.

TIFF format’s main appeal is its flexibility. It handles black and white line images, as well as gray scale and color images, which can be easily transported between different operating systems and computers.

BigTIFF

BigTIFF data is now supported which breaks the 4 GB barrier for TIFF files. It is possible to create TIFF files up to 18,000 petabytes in size. BigTIFF has complete backward compatibility (when using ERDAS tools) and can be read the same way Classic TIFF files are read.

TIFF is a 32-bit offset data file format that can store image data, and metadata (stored in tags) up to 2^32. BigTIFF is the same TIFF specification but stores the data using a 64-bit offset, thus the file size expands to 2^64. It is also known as TIFF 64.

TIFF File Formats

TIFF’s great flexibility can also cause occasional problems in compatibility. This is because TIFF is really a family of file formats that is comprised of a variety of elements within the format.

The following table shows key Baseline TIFF format elements and the values for those elements supported by ERDAS IMAGINE.

SHARED Tip Any TIFF file that contains an unsupported value for one of these elements may not be compatible with ERDAS IMAGINE.

Byte Order

Intel (LSB/MSB)

Motorola (MSB/LSB)

Image Type

Black and white

Gray scale

Inverted gray scale

Color palette

RGB (3-band)

Configuration

BIP

BSQ

Bits Per Plane

1, 2, 4, 8, 16, 32, 64

Compression

None

CCITT G3 (B&W only)

CCITT G4 (B&W only)

Packbits

LZW

LZW with horizontal differencing

For Bits per Plane, all bands must contain the same number of bits (that is, 4, 4, 4 or 8, 8, 8). Multiband data with bit depths differing per band cannot be imported into ERDAS IMAGINE.

For Bits per Plane, 1 and 2 must be imported and exported as 4-bit data.

For Bits per Plane, 16, 32, and 64 must be imported and exported as 4-bit data.

For Compression, compression is supported on import and direct read/write only.

Read more about TIFF specification at Adobe Systems Inc. website http://partners.adobe.com/public/developer/tiff/index.html.

GeoTIFF

According to the GeoTIFF Format Specification, Revision 1.0, "The GeoTIFF spec defines a set of TIFF tags provided to describe all ’Cartographic’ information associated with TIFF imagery that originates from satellite imaging systems, scanned aerial photography, scanned maps, digital elevation models, or as a result of geographic analysis" (Ritter and Ruth, 1995).

GeoTIFF format separates cartographic information into two parts: georeferencing and geocoding.

Georeferencing

Georeferencing is the process of linking the raster space of an image to a model space (that is, a map system). Raster space defines how the coordinate system grid lines are placed relative to the centers of the pixels of the image. In ERDAS IMAGINE, the grid lines of the coordinate system always intersect at the center of a pixel. GeoTIFF allows the raster space to be defined either as having grid lines intersecting at the centers of the pixels (PixelIsPoint) or as having grid lines intersecting at the upper left corner of the pixels (PixelIsArea). ERDAS IMAGINE converts the georeferencing values for PixelIsArea images so that they conform to its raster space definition.

GeoTIFF allows georeferencing via a scale and an offset, a full affine transformation, or a set of tie points. ERDAS IMAGINE currently ignores GeoTIFF georeferencing in the form of multiple tie points.

Geocoding

Geocoding is the process of linking coordinates in model space to the Earth’s surface. Geocoding allows for the specification of projection, datum, ellipsoid, and so forth. ERDAS IMAGINE interprets the GeoTIFF geocoding to determine the latitude and longitude of the map coordinates for GeoTIFF images. This interpretation also allows the GeoTIFF image to be reprojected.

In GeoTIFF, the units of the map coordinates are obtained from the geocoding, not from the georeferencing. In addition, GeoTIFF defines a set of standard projected coordinate systems. The use of a standard projected coordinate system in GeoTIFF constrains the units that can be used with that standard system. Therefore, if the units used with a projection in ERDAS IMAGINE are not equal to the implied units of an equivalent GeoTIFF geocoding, ERDAS IMAGINE transforms the georeferencing to conform to the implied units so that the standard projected coordinate system code can be used. The alternative (preserving the georeferencing as is and producing a nonstandard projected coordinate system) is regarded as less interoperable.

Read more about GeoTIFF specification at http://www.remotesensing.org/geotiff/spec/geotiffhome.html.