The BEAM-DIMAP Data Format

Introduction

The Sentinel Toolbox and the scientific data processors use a simple data input/output format, which makes it easy to import SNAP data in other imaging applications. The format is called DIMAP and has been developed by SPOT-Image, France. The SNAP software uses a special DIMAP profile called BEAM-DIMAP.

The BEAM-DIMAP is the standard I/O product format for the Sentinel Toolbox and the Scientific Tools. The format was chosen because it has a very simple and easy-to-understand structure. It was originally developed for the BEAM toolbox and development platform,hence the name.

Overview

A data product stored in this format is composed of

• a single product header file with the suffix .dim in XML format containing the product meta-data
• an additional directory with the same name plus the suffix .data
  
  • containing ENVI^®-compatible images for each band.
  • containing a directory tie_point_grids which contains ENVI^®-compatible images for each tie-point grid
  • containing a directory vector_datawhich contains the geometries in a proprietary format.

The following diagram shows the structure:

The geo-coding of data products in satellite co-ordinates is stored in so called tie-point grid datasets. Tie-point datasets are stored in the data directory and have exactly the same format as the geophysical bands of a product.

Because of its simplicity, the product components can be accessed with nearly every image processing system or programming language. The product's metadata can directly be viewed in a text editor or an XML viewer. The following contains a closer look at the components of the product format.

Important Note: In the SNAP software, pixel values always refer to the upper left corner of the pixel. Map co-ordinates also typically refer to the upper left corner of the pixel. Pixel co-ordinates are always zero based, the pixel at X=0,Y=0 refers to the upper left pixel of an image and the upper left corner of that pixel.

XML Product Header

XML stands for eXtensible Markup Language and is a mark-up language much like the better known HTML.

The most significant difference is that HTML is about displaying information, XML is about describing information. XML can be stored in plain text files and provides a data structuring scheme composed of elements and attributes. An element is enclosed by tags, but the tags are not predefined in XML. In order to use XML as a data storage, an application or standard must define its own tags. In this case, XML is very similar to HDF because the interpretation of the content is left to the application or, respectively, to the user.

The product header for a BEAM-DIMAP data product will contain two types of information

• meta-data describing the image data contained in the product
• the geophysical bands and tie point grids of the product as references to the actual image data files, since the image data is not stored in XML format

The reason why the image data is not directly stored in XML is simple: The product header is plain text, and an image stored in text format requires about 4 to 10 times more space than the simplest binary format and therefore requires considerable processing overhead when reading to and writing from such files.

Image Data Format

One geophysical band in the data product is represented by a single image. The image data - the data product's geophysical samples - are stored in flat binary files in big endian order. What makes this format compatible with the simple ENVI image format is that an extra image header for each image is also stored: the ENVI header.

ENVI Header File Format

The header files also have plain text format and comprise key-values pairs describing storage layout of the raw data in terms of raster width, height and the sample data type. In addition to ENVI, multiple other imaging applications are capable of importing image files having flat binary format. Here is an example of a header file:

    ENVI
    samples = 1100
    lines = 561
    bands = 1
    header offset = 0
    file type = ENVI Standard
    data type = 4
    interleave = bsq
    byte order = 1

An ENVI header file starts with the text string ENVI to be recognized by ENVI as a native file header. Keywords within the file are used to indicate critical file information. The following keywords are used by the BEAM-DIMAP format: