GeoTIFF Web Page Table of Contents Top of Section 2

2.4 GeoTIFF File and "Key" Structure

This section describes the abstract file-format and "GeoKey" data storage mechanism used in GeoTIFF. Uses of this mechanism for implementing georeferencing and geocoding is detailed in section 2.6 and section 2.7 .

A GeoTIFF file is a TIFF 6.0 file, and inherits the file structure as described in the corresponding portion of the TIFF spec. All GeoTIFF specific information is encoded in several additional reserved TIFF tags, and contains no private Image File Directories (IFD's), binary structures or other private information invisible to standard TIFF readers.

The number and type of parameters that would be required to describe most popular projection types would, if implemented as separate TIFF tags, likely require dozens or even hundred of tags, exhausting the limited resources of the TIFF tag-space. On the other hand, a private IFD, while providing thousands of free tags, is limited in that its tag-values are invisible to non-savvy TIFF readers (which don't know that the IFD_OFFSET tag value points to a private IFD).

To avoid these problems, a GeoTIFF file stores projection parameters in a set of "Keys" which are virtually identical in function to a "Tag", but has one more level of abstraction above TIFF. Effectively, it is a sort of "Meta-Tag". A Key works with formatted tag-values of a TIFF file the way that a TIFF file deals with the raw bytes of a data file. Like a tag, a Key has an ID number ranging from 0 to 65535, but unlike TIFF tags, all key ID's are available for use in GeoTIFF parameter definitions.

The Keys in GeoTIFF (also call "GeoKeys") are all referenced from the GeoKeyDirectoryTag, which defined as follows:

      Tag = 34735 (87AF.H) 
      Type = SHORT (2-byte unsigned short)
      N = variable, >= 4
      Alias: ProjectionInfoTag, CoordSystemInfoTag
      Owner: SPOT Image, Inc.

This tag may be used to store the GeoKey Directory, which defines and references the "GeoKeys", as described below.

The tag is an array of unsigned SHORT values, which are primarily grouped into blocks of 4. The first 4 values are special, and contain GeoKey directory header information. The header values consist of the following information, in order:

  Header={KeyDirectoryVersion, KeyRevision, MinorRevision, NumberOfKeys}
     "KeyDirectoryVersion" indicates the current version of Key
     implementation, and will only change if this Tag's Key 
     structure is changed. (Similar to the TIFFVersion (42)).
     The current DirectoryVersion number is 1. This value will
     most likely never change, and may be used to ensure that
     this is a valid Key-implementation.
     "KeyRevision" indicates what revision of Key-Sets are used.
     "MinorRevision" indicates what set of Key-codes are used. The
     complete revision number is denoted <KeyRevision>.<MinorRevision>
     "NumberOfKeys" indicates how many Keys are defined by the rest
     of this Tag.

This header is immediately followed by a collection of <NumberOfKeys> KeyEntry sets, each of which is also 4-SHORTS long. Each KeyEntry is modeled on the "TIFFEntry" format of the TIFF directory header, and is of the form:

   KeyEntry = { KeyID, TIFFTagLocation, Count, Value_Offset }
     "KeyID" gives the key-ID value of the Key (identical in function
     to TIFF tag ID, but completely independent of TIFF tag-space),
     "TIFFTagLocation" indicates which TIFF tag contains the value(s)
      of the Key: if TIFFTagLocation is 0, then the value is SHORT,
      and is contained in the "Value_Offset" entry. Otherwise, the type
      (format) of the value is implied by the TIFF-Type of the tag
      containing the value.
     "Count" indicates the number of values in this key.
      "Value_Offset" Value_Offset indicates the index-
      offset *into* the TagArray indicated by TIFFTagLocation, if
      it is nonzero. If TIFFTagLocation=0, then Value_Offset
      contains the actual (SHORT) value of the Key, and
      Count=1 is implied. Note that the offset is not a byte-offset,
      but rather an index based on the natural data type of the
      specified tag array.  

Following the KeyEntry definitions, the KeyDirectory tag may also contain additional values. For example, if a Key requires multiple SHORT values, they shall be placed at the end of this tag, and the KeyEntry will set TIFFTagLocation=GeoKeyDirectoryTag, with the Value_Offset pointing to the location of the value(s).

All key-values which are not of type SHORT are to be stored in one of the following two tags, based on their format:

      Tag = 34736 (87BO.H) 
      Type = DOUBLE (IEEE Double precision)
      N = variable
      Owner: SPOT Image, Inc.

This tag is used to store all of the DOUBLE valued GeoKeys, referenced by the GeoKeyDirectoryTag. The meaning of any value of this double array is determined from the GeoKeyDirectoryTag reference pointing to it. FLOAT values should first be converted to DOUBLE and stored here.

      Tag = 34737 (87B1.H)  
      Type = ASCII
      Owner: SPOT Image, Inc.
      N = variable

This tag is used to store all of the ASCII valued GeoKeys, referenced by the GeoKeyDirectoryTag. Since keys use offsets into tags, any special comments may be placed at the beginning of this tag. For the most part, the only keys that are ASCII valued are "Citation" keys, giving documentation and references for obscure projections, datums, etc.

Note on ASCII Keys:
Special handling is required for ASCII-valued keys. While it is true that TIFF 6.0 permits multiple NULL-delimited strings within a single ASCII tag, the secondary strings might not appear in the output of naive "tiffdump" programs. For this reason, the null delimiter of each ASCII Key value shall be converted to a "|" (pipe) character before being installed back into the ASCII holding tag, so that a dump of the tag will look like this.


A baseline GeoTIFF-reader must check for and convert the final "|" pipe character of a key back into a NULL before returning it to the client software.

GeoKey Sort Order:
In the TIFF spec it is required that TIFF tags be written out to the file in tag-ID sorted order. This is done to avoid forcing software to perform N-squared sort operations when reading and writing tags.

To follow the TIFF philosophy, GeoTIFF-writers shall store the GeoKey entries in key-sorted order within the CoordSystemInfoTag.

  GeoKeyDirectoryTag=(   1,     1, 2,     6,
                      1024,     0, 1,     2,
                      1026, 34737,12,     0,
                      2048,     0, 1, 32767,
                      2049, 34737,14,    12,
                      2050,     0, 1,     6,
                      2051, 34736, 1,     0 )
  GeoAsciiParamsTag(34737)=("Custom File|My Geographic|")

The first line indicates that this is a Version 1 GeoTIFF GeoKey directory, the keys are Rev. 1.2, and there are 6 Keys defined in this tag.

The next line indicates that the first Key (ID=1024 = GTModelTypeGeoKey) has the value 2 (Geographic), explicitly placed in the entry list (since TIFFTagLocation=0). The next line indicates that the Key 1026 (the GTCitationGeoKey) is listed in the GeoAsciiParamsTag (34737) array, starting at offset 0 (the first in array), and running for 12 bytes and so has the value "Custom File" (the "|" is converted to a null delimiter at the end). Going further down the list, the Key 2051 (GeogLinearUnitSizeGeoKey) is located in the GeoDoubleParamsTag (34736), at offset 0 and has the value 1.5; the value of key 2049 (GeogCitationGeoKey) is "My Geographic".

The TIFF layer handles all the problems of data structure, platform independence, format types, etc, by specifying byte-offsets, byte-order format and count, while the Key describes its key values at the TIFF level by specifying Tag number, array-index, and count. Since all TIFF information occurs in TIFF arrays of some sort, we have a robust method for storing anything in a Key that would occur in a Tag.

With this Key-value approach, there are 65536 Keys which have all the flexibility of TIFF tag, with the added advantage that a TIFF dump will provide all the information that exists in the GeoTIFF implementation.

This GeoKey mechanism will be used extensively in section 2.7, where the numerous parameters for defining Coordinate Systems and their underlying projections are defined.