{
  "id": "geotiff",
  "title": "GeoTIFF",
  "lang": "en",
  "source_lang": "en",
  "summary": "GeoTIFF is a metadata standard for embedding geographic referencing information within TIFF raster files, enabling imagery and gridded data to be located in a coordinate reference system.",
  "category": "computer-science",
  "subcategory": "software",
  "type": "concept",
  "aliases": [
    "Georeferenced TIFF"
  ],
  "generated_at": "2026-11-06",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/geotiff/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/geotiff/",
      "json": "https://aiwiki.page/zh/geotiff.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "GeoTIFF is a standard for encoding geographic referencing information within files based on the Tagged Image File Format (TIFF). It allows a raster image or numerical grid to carry the information needed to relate its pixels to locations in a geographic or projected coordinate system. GeoTIFF combines TIFF’s image-storage capabilities with geographic metadata, making it a widely used exchange and storage format for satellite imagery, aerial photography, elevation models, and other geospatial raster data.\n\nA GeoTIFF usually has a .tif or .tiff filename extension. The extension alone does not distinguish it from an ordinary TIFF: the defining feature is its geographic tags and their interpretation."
    },
    {
      "heading": "Purpose and data model",
      "text": "A raster represents information as a rectangular grid of cells. Its values might describe visible colors, measured radiation, elevation, temperature, or land-cover classes. Without additional information, the grid’s row and column numbers do not indicate where it lies on Earth.\n\nGeoTIFF addresses two related questions:\n\n- Where is the grid? A transformation relates raster coordinates to coordinates in a geographic or projected space.\n- What do those coordinates mean? A coordinate reference system (CRS) specifies the reference framework, including relevant units, axes, and geodetic information.\n\nThese functions are distinct. A transformation can assign coordinate values to pixels, but those values cannot be interpreted reliably without knowing their reference system. Conversely, identifying a CRS does not establish the position and spacing of a particular image.\n\nGeoTIFF does not prescribe the scientific meaning of every stored value. A complete dataset may also need band descriptions, measurement units, acquisition dates, quality flags, or processing history that are supplied through additional metadata conventions."
    },
    {
      "heading": "File structure and geographic tags",
      "text": "TIFF storage\n\nTIFF organizes image information using image file directories, whose entries describe properties such as dimensions, sample types, compression, and the locations of image data. GeoTIFF uses this structure rather than replacing it.\n\nDepending on the TIFF encoding and software support, a GeoTIFF can contain:\n\n- Multiple bands, such as red, green, blue, and near-infrared measurements.\n- Integer or floating-point samples.\n- Image data arranged in strips or rectangular tiles.\n- Compressed or uncompressed pixel values.\n- Additional images, often used for reduced-resolution overviews.\n\nAn ordinary TIFF reader may display the pixels while ignoring the geographic metadata. This compatibility does not mean that every TIFF application supports every GeoTIFF sample type, compression method, or file size.\n\nGeoreferencing tags\n\nThe principal GeoTIFF tags include:\n\nTag | Purpose\nModelTiepointTag | Associates raster coordinates with coordinates in model space.\nModelPixelScaleTag | Specifies raster spacing along model axes for suitable axis-aligned transformations.\nModelTransformationTag | Encodes a transformation matrix for more general raster-to-model mappings.\nGeoKeyDirectoryTag | Stores the directory of geographic keys and references to their values.\nGeoDoubleParamsTag | Stores floating-point values referenced by geographic keys.\nGeoAsciiParamsTag | Stores text values referenced by geographic keys.\n\nThe transformation tags describe the relationship between the raster and model coordinates. The geographic keys, or GeoKeys, describe how those model coordinates should be interpreted.\n\nGeoKeys provide a compact mechanism for recording information such as the model type, CRS identifiers, units, and coordinate-system parameters. A key can contain a short numeric value directly or refer to values stored elsewhere in the GeoTIFF metadata."
    },
    {
      "heading": "Coordinate reference systems",
      "text": "GeoTIFF can describe geographic coordinate systems, in which coordinates commonly represent longitude and latitude, and projected coordinate systems, in which positions are expressed on a plane through a map projection. It also provides mechanisms for vertical reference information.\n\nMany datasets identify their CRS using codes from the EPSG Geodetic Parameter Dataset. A recognized identifier can convey a standardized definition without requiring every component to be described independently. GeoTIFF also supports user-defined parameters, although unusual or incompletely specified definitions can reduce interoperability.\n\nCorrect interpretation can depend on:\n\n- The geodetic datum or reference frame.\n- The projection and its parameters.\n- Angular and linear units.\n- Coordinate ordering and the applicable encoding conventions.\n- Any vertical reference system associated with height values.\n\nMerely assigning a different CRS identifier does not transform the data. Reprojection requires converting positions between coordinate systems and usually resampling the raster onto a new grid. Changing metadata without performing the corresponding transformation can place an otherwise valid image in the wrong location."
    },
    {
      "heading": "Raster-to-coordinate transformation",
      "text": "For a common north-up raster, the transformation can be written as\n\n\\[X = X_0 + i\\,s_x, \\qquad Y = Y_0 - j\\,s_y,\\]\n\nwhere \\(i\\) and \\(j\\) are raster coordinates, \\(X_0\\) and \\(Y_0\\) specify the model coordinates of the raster origin, and \\(s_x\\) and \\(s_y\\) are positive pixel spacings. The minus sign reflects the common arrangement in which image rows increase downward while projected northing increases upward.\n\nA more general two-dimensional affine transformation has the form\n\n\\[X = a_0 + a_1 i + a_2 j, \\qquad Y = b_0 + b_1 i + b_2 j.\\]\n\nThis permits rotation and shear as well as translation and scaling. GeoTIFF’s transformation-matrix mechanism can express such mappings.\n\nThe mapping from raster space to model space must not be confused with the map projection itself. The former places the grid in its coordinate system; the latter relates projected coordinates to a geographic reference framework.\n\nPixel areas and pixel points\n\nThe GTRasterTypeGeoKey distinguishes two interpretations:\n\n- PixelIsArea: each raster value represents an area or cell. In the GeoTIFF raster-coordinate convention, integer coordinates identify cell boundaries, so the center of the first cell is at \\((0.5, 0.5)\\).\n- PixelIsPoint: each raster value represents a sample at a point. The first sample is located at raster coordinate \\((0,0)\\).\n\nThis distinction matters when aligning datasets. Software that confuses cell corners with sample locations can introduce a half-pixel displacement. Libraries may expose a common corner-based transformation convention even when reading point-sampled data, so an application’s internal representation need not reproduce the stored coordinates literally."
    },
    {
      "heading": "Compression, large files, and efficient access",
      "text": "GeoTIFF inherits TIFF’s support for data compression. Common choices include lossless compression methods such as LZW and Deflate, and, for suitable imagery, lossy JPEG compression. The appropriate encoding depends on the data: a method acceptable for visual imagery may alter numerical measurements or categorical values.\n\nClassic TIFF uses 32-bit offsets, which constrain file addressing to roughly 4 GiB. BigTIFF extends the format with 64-bit offsets and related structural changes. GeoTIFF metadata can be stored in BigTIFF files, but readers must support that TIFF variant.\n\nTiling and reduced-resolution overviews can greatly improve access to large rasters. Tiles allow software to read a spatial subset without decoding entire image rows. Overviews allow small-scale display without reading every full-resolution pixel. Neither feature changes the dataset’s basic geographic meaning."
    },
    {
      "heading": "Cloud Optimized GeoTIFF",
      "text": "A Cloud Optimized GeoTIFF (COG) is a GeoTIFF organized to support efficient remote access. It combines a suitable internal TIFF layout, tiling, and overview arrangements with byte-range retrieval, commonly through HTTP range requests.\n\nA client can obtain metadata and selected image blocks without downloading the entire file. This makes COG useful for large imagery collections hosted on web servers or object storage.\n\nAn ordinary GeoTIFF does not automatically have this access efficiency. Conversely, a COG does not require a proprietary image encoding: it remains a TIFF-based geospatial raster file readable by compatible software. Performance also depends on server support, network latency, tile sizes, compression, and the client’s access strategy."
    },
    {
      "heading": "History and standardization",
      "text": "GeoTIFF originated in the early 1990s through efforts to standardize geographic information embedded in TIFF images. It became widely implemented before its adoption as a formal standard by the Open Geospatial Consortium (OGC).\n\nThe OGC published GeoTIFF 1.1 in 2019. This version retained the established encoding approach while clarifying requirements and updating the treatment of geodetic definitions and references to the EPSG dataset.\n\nThe format’s adoption has been supported by open-source software libraries and tools, including libgeotiff and GDAL, as well as commercial mapping, remote-sensing, and image-processing applications."
    },
    {
      "heading": "Applications and limitations",
      "text": "GeoTIFF is commonly used in geographic information systems and remote sensing for orthorectified imagery, digital elevation models, environmental grids, thematic maps, and outputs from image processing. It is particularly suited to exchanging a raster together with its basic georeferencing in a single file.\n\nImportant limitations remain:\n\n- Georeferencing does not guarantee accuracy. A correctly encoded file can still contain incorrect coordinates, an unsuitable CRS, or positional errors in its source data.\n- Missing-data conventions are not entirely uniform. Widely used no-data tags, masks, and application-specific metadata may rely on conventions beyond the core GeoTIFF standard.\n- Scientific descriptions may need additional structures. Complex time series, multidimensional variables, and detailed measurement metadata are often more naturally represented in formats designed for those purposes.\n- Software support varies. Uncommon compression methods, large-file variants, specialized CRS definitions, or auxiliary metadata may not survive every conversion.\n- Raster and vector needs differ. GeoTIFF principally stores raster data; formats such as GeoPackage address other packaging needs, including vector features and tiled raster datasets.\n\nAn external world file can supply a raster’s affine placement, but does not by itself define a complete CRS. GeoTIFF’s principal advantage over that arrangement is its ability to store both placement and coordinate-system information inside the image file."
    }
  ],
  "links": [
    "pixel",
    "metadata",
    "earth",
    "coordinate-reference-system",
    "map-projection",
    "epsg-geodetic-parameter-dataset",
    "interoperability",
    "data-compression",
    "lossless-data-compression",
    "jpeg",
    "hypertext-transfer-protocol",
    "open-geospatial-consortium",
    "open-source-software",
    "geographic-information-system",
    "remote-sensing",
    "digital-elevation-model",
    "image-processing",
    "geopackage"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}