sentinel1decoder 1.1.0

A python decoder for ESA Sentinel-1 Level0 files

sentinel1decoder

A Python decoder for Sentinel-1 Level 0 files. The Level 0 format consists of the raw space packets downlinked from the Sentinel-1 spacecraft. This package decodes these packets and produces the raw I/Q sensor output from the SAR instrument, which can then be further processed to focus a SAR image.

An example Jupyter notebook demonstrating the process of decoding Level 0 data and forming an image is available on GitHub here or on nbviewer.org here.

This code is based on an implementation in C by jmfriedt, which can be found here.

Performance

Version 1.1.0 features a complete rewrite of the core decoding functions in Rust, delivering dramatic performance improvements:

• ~80x faster decoding: Typical burst decoding time reduced to ~30 seconds
• Multithreaded batch processing: Multiple packets are decoded in parallel for optimal performance
• Optimized algorithms: Improved FDBAQ decoding with lookup tables and efficient bit manipulation

The API remains fully backwards compatible - upgrade to enjoy the performance boost without changing your code.

Installation

This package is available on PyPI.

In a terminal window:

pip install sentinel1decoder

This package requires Python 3.8 or higher. NumPy and Pandas are also required.

Usage

High-level API

The Level0File class provides a high-level interface that automatically breaks the file into "bursts" (consecutive packets with constant swath number and number of quads) for easy handling. For more fine-grained control, see the Low-level API section below.

Import the package:

import sentinel1decoder

Initialize a Level0File object:

l0file = sentinel1decoder.Level0File(filename)

This class contains a DataFrame with the packet metadata:

l0file.packet_metadata

A DataFrame containing the ephemeris:

l0file.ephemeris

The metadata is indexed by burst as well as packet number. Metadata on individual bursts can be accessed via:

l0file.get_burst_metadata(burst)

The I/Q array for each burst can be generated via:

iq_data = l0file.get_burst_data(burst)

Returns a NumPy array of complex64 with shape (slow_time, fast_time) or (echo_num, sample_num). Each row represents one radar echo return; each column represents a sample within that echo.

By default, this method will attempt to load cached data from an .npy file if available (created using save_burst_data). You can disable this behavior by passing try_load_from_file=False.

This data can be cached in an .npy file using:

l0file.save_burst_data(burst)

Low-level API

The individual decoding functions can also be used directly:

Initialize a Level0Decoder object:

decoder = sentinel1decoder.Level0Decoder(filename)

Generate a Pandas DataFrame containing the header information associated with the Sentinel-1 downlink packets:

df = decoder.decode_metadata()

Decode the satellite ephemeris data from the packet headers:

ephemeris = sentinel1decoder.utilities.read_subcommed_data(df)

Extract the data payload from specific packets. This method takes a Pandas DataFrame as input and only decodes packets whose headers are present in the input DataFrame. This allows you to select which packets to decode, rather than always decoding the full file.

The method uses multithreaded batch processing for optimal performance. You can control the batch size (default: 256 packets) using the optional batch_size parameter.

For example, to decode only the first 100 packets:

selection = df.iloc[0:100]
iq_array = decoder.decode_packets(selection)

Or with a custom batch size:

iq_array = decoder.decode_packets(selection, batch_size=512)

Returns a NumPy array of complex64 with shape (slow_time, fast_time) or (echo_num, sample_num). Each row represents one radar echo return; each column represents a sample within that echo.