Reading SAR data: a new general reader and a proposed new standard

All recent scientific SAR data is formatted in compliance, more or less, with the CEOS standard developed recently. If this standard were followed meticulously by each organization that produces SAR data, a single program could read all of the data. However, there are small but significant differences in how each sensor's data is formatted and this has led to the need for a different program to read the data for each sensor, and even for each format from the same sensor. This makes using the data difficult for most users, as the first step is to read it in to some kind of image processing system, such as ERDAS, PCI, GRASS, KHOROS, etc. The solution to this problem has been for data providers and commercial vendors to write tape readers for their customers. Unfortunately, many of these readers do not read all the ancillary data and do not allow easy porting for use with other image processing systems. A new general reader has been developed that solves all these problems. It uses simple text "format" files to define the data that is to be read in: variable name and type, location in file, and format. The reader writes a new file in a new format, as described later. A standard reader now reads from this new file, allowing the user to write interface code to read the data (ancillary and image) into the user's image processing system. The new proposed format addresses the following issues: (1) a single file per image, with a single record-length per file; (2) Ancillary data in ASCII: KEYWORD=value; comment; (3) Self-documenting via comments and strict syntax; (4) Self-reading, so that a program can inspect input file and generate code that can read in all data in input file; (5) Addition of new parameters is simple, and users can generate readers for the new format with the program in 4. This code is written and works now.