Importing and Georeferencing ALOS CEOS Format
Map and path oriented products
The native data format for Geoscience Australia's ALOS products is CEOS. This format is a mixed ascii and binary format which may not be well supported in your image processing package.
The ALOS CEOS products include ER Mapper (.ers) and ENVI (.hdr) headers with each band. This enables direct read of this format in many image processing packages. For novice users it is recommended that you chose the GeoTIFF Format Specifications in your processing options when you order products.
If your image processing package does not have native support for ER Mapper or ENVI header files you will need to import the ALOS CEOS format data. To do this you will require:
- header offset
- lines and pixels
- registration x and y
- registration easting and northing
- pixel size
- rotation if applicable i.e. not required for map oriented products
- byte size / data type (8 or 16 bit)
- map projection
This information can be retrieved from the ER Mapper text header file located in the headers directory of scene01.
There are two ways to import and georeference your image in an image processing package:
- Direct Read (you will be prompted for information to enable the image processing software to interpret the CEOS format and visualise the data)
- Import (by using a generic binary import you can enter parameters to enable import and writing of the CEOS format to the native format of the image processing package).
Below is an example of the contents of an ER Mapper header file with the fields required to geocode your image in bold. This information can be used to import and georeference your CEOS dataset.
HeaderOffset = 8592
DataSetType = ERStorage
DataType = Raster
ByteOrder = LSBFirst
Datum = "WGS84"
Projection = "SUTM50"
CoordinateType = EN
CellType = Unsigned8BitInteger
NullCellValue = 0
Xdimension = 10.0000000
Ydimension = 10.000000
NrOfLines = 8324
NrOfCellsPerLine = 8592
Eastings = 308633.94
Northings = 6698664.8414
RegistrationCellX = 4280.5
RegistrationCellY = 4162.5000000
NrOfBands = 1
Value = ""
Below is a generic procedure which outlines how to import CEOS format ALOS data into an image processing package.
- From the data import menu, select 'Import Generic Binary' or similar import function. Select the image you would like to import
- Set the 'data format' to BIL or BSQ (ALOS image data is in individual bands so there is no difference here)
- Set the data type to Unsigned 8 Bit Integer (CellType = Unsigned8BitInteger)
- Set the 'data file header' or 'offset' to 8592 (HeaderOffset = 8592)
- Set the number of 'rows' or 'lines' to 8324 (NrOfLines = 8324)
- Set the number of 'columns' or 'pixels' to 8592 (NrOfCellsPerLine = 8592)
- Import the data and visualise
Some image processing packages may prompt you for georeferencing information at this stage. Alternatively you should be able to add this information once the image data is written to the native format of the image processing package
- Navigate to the 'edit image information' function
- Update the image information by setting the 'registration pixel x' to 4280.5 (RegistrationCellX = 4280.5)
- Set the 'registration pixel y' to 4612.5 (RegistrationCellY = 4162.5)
- If the product you have is path oriented, you will need to enter a 'rotation'
Some image processing packages will only let you enter the upper left x and y. This can be calculated by multiplying the cell dimension in metres by the number of cells. This value can then be subtracted from the header eastings value to provide the upper left easting and added to the header northing to obtain the upper left northing. You should be aware also that not all image processing packages reference pixels in the same manner. For example, some packages use 0,0 as the centre of the upper left pixel, while other refer to the same pixel as 1,1
- Enter the 'Easting' value 308633.94 (Eastings = 308633.94)
- Enter the 'Northing' value 6698664.8414 (Northings = 6698664.8414)
- Enter the 'pixel size x' as 10 (Xdimension = 10.0000000)
- Enter the 'pixel size y' as 10 (Ydimension = 10.0000000)
You will also need to update the map information for the imported image
- Navigate to the 'update projection information' function
- Enter the 'projection' as UTM, 'zone' as 50 and 'hemisphere' as South (Projection = "SUTM50")
- Enter the 'datum' as WGS84 (Datum = "WGS84")
You should now be able to visualise the image and confirm the georeferencing of the image against a confidently positioned reference.
You will need to repeat this procedure for each image within the dataset.
Topic contact: email@example.com Last updated: February 10, 2012