SIGDEM File format specification

The Scaled Integer Gridded Digital Elevation Model (SIGDEM) is a binary file format containing Gridded Digital Elevation Model data.

File naming

SIGDEM files have the .sigdem file extension. The filename before the extension is the baseName (e.g. canada for canada.sigdem).

File extension: sigdem

Mime Type: image/x-revolsys-sigdem

SIGDEM files can also be wrapped in a .gz ([baseName].sigdem.gz) or .zip ([baseName].sigdem.zip) file. The baseName of the .sigdem file inside the zip file must be the same as the baseName of the zip file.

Software implementing the SIGDEM file formats SHOULD be written to directly read .gz and .zip compressed files.

Header

A SIGDEM file starts with the following 132 byte file header.

Field	Type	Description
fileId	char[6]	The file identification characters "SIGDEM".
version	short	The current version number (e.g. 1).
coordinateSystemId	int	The EPSG coordinate system id (e.g. https://epsg.io/26910 NAD83 / UTM zone 10N). Use 0 if there is no EPSG ID for that coordinate system.
offsetX	double	The offset to add to the x value.
scaleX	double	The scale factor for x coordinates (e.g. 1000 for mm precision).
offsetY	double	The offset to add to the y value.
scaleY	double	The scale factor for y coordinates (e.g. 1000 for mm precision).
offsetZ	double	The offset to add to the z value.
scaleZ	double	The scale factor for z coordinates (e.g. 1000 for mm precision).
minX	double	The minimum x coordinate of the file bounding box.
minY	double	The minimum y coordinate of the file bounding box.
minZ	double	The minimum z value of the elevations stored.
maxX	double	The maximum x coordinate of the file bounding box.
maxY	double	The maximum y coordinate of the file bounding box.
maxZ	double	The maximum z value of the elevations stored.
gridWidth	int	The number of grid cells in each row.
gridHeight	int	The number of rows in the grid.
gridCellWidth	double	The width of each cell in x units.
gridCellHeight	double	The height of each cell in y units.

NOTE: The offsetX, scaleX, offsetY, and scaleY are included for completeness in the file, but are not used for any data in the file.

NOTE: The fields in the header (e.g. minX) are the full coordinates and are not offset or scaled.

Content

The elevation data is stored immediately after the header in grid cells, ordered by row relative to the lowest left corner of the file's bounding box. The first grid cell stores the elevation for the coordinates minX,minY. The last grid cell contains the elevation for the coordinates maxX, maxY.

The following shows the coordinates of each row in the file.

minY:     minX, minX + gridCelWidth, ..., maxX
minY + 1: minX, minX + gridCelWidth, ..., maxX
  :
maxY:     minX, minX + gridCelWidth, ..., maxX

Grid cells are referenced using gridX in the range 0..gridWidth - 1 and gridY in the range 0..gridHeight - 1.

Each grid cell contains the elevation (z-coordinate) as a scaled int value. NULL values are stored using the value -2³¹

The map coordinates of each grid cell (x, y) are at the lowest left corner of the grid cell (other DEM formats use the centre).

When reading a file the z-coordinate can be calculated from the scaled int value using the following formula.

z = offsetZ + value / scaleZ

When writing a file the scaled int value can be calculated from the z-coordinate using the following formula.

value = round((z - offsetZ) * scaleZ)

The following pseudo code shows how to read the elevations from the file.

model = new DEM()
for gridY in 0..gridHeight - 1
  y = minY + gridY * gridCellHeight
  for gridX in 0..gridWidth - 1
    value = readInt()
    x = minX + gridX * gridCellWidth
    z = offsetZ + value / scaleZ
    model.setZ(x, y, z)

The grid is optimised for direct file access to the elevations for a given x, y coordinate using simple math. Assuming the x,y is in the file bounding box the following pseudo code can be used.

grixX = floor((x - minX) / gridCellWidth)
grixY = floor((y - minY) / gridCellHeight)
fileOffset = 132 + (gridY * gridWidth + gridX) * 4
value = readInt(fileOffset)
z = offsetZ + value / scaleZ

Supplementary files

For ease of distribution the use of supplementary files should be avoided. Multiple files would require wrapping all the files in a .zip container.

NOTE: Other supplementary files can be included in a distribution package but these should not be expected to be handled by software processing .sigdem files.

Projection (.prj) file

Only if there an EPSG coordinateSystemId doesn't exist for the coordinate system a ESRI WKT .prj file be provided. The .prj filename will be [baseName].prj (e.g. canada.prj for canada.sigdem).

NOTE: Implementations and users MUST not generate and ignore this file if the SIGDEM file includes a coordinateSystemId.

Data types

The following data types are used. The names are based on the Java type names. All data is stored using big-endian byte order.

Name	Byte Count	Description
char[n]	n	Array of UTF-8 characters. n is the array length. Current use is the text SIGDEM which is also the ASCII subset of UTF-8.
short	2	Big-endian 16-bit signed integer number.
int	4	Big-endian 32-bit signed integer number.
double	8	Big-endian 64-bit IEEE 754 floating point number.