CHRIS PROBA

1
Dr Trevor Read Earth Observation, Navigation
Science (UK) Astrium Ltd Anchorage
Road Portsmouth PO3 5PU England Tel 44 (0)
1438 313456 Direct Tel [phone]
705375 Email trevor.read_at_astrium-space.com
CHRIS/PROBA WORKSHOP Trevor READ ESTEC 3-4 April
2003
2

PROBA PAYLOAD PROCESSING

• INTRODUCTION
• PROBA Payload On-Board Processing (PPU)
• Processor
• CHRIS Data
• PASS (PROBA)
• PROBA and Data Compression
• JPEG Compression of microcamera images
• Data Compression Activities at Astrium
• Advanced Methods for Lossless Compression of
  Hyperspectral Data

3
PROBA PAYLOAD PROCESSING

• Astrium developed and assembled the flight model
  of the Payload Processing Unit flying on the
  PROBA spacecraft, launched in October 2001.
• The programme has been supported by ESA, BNSC,
  DERA and Astrium Ltd.
• The unit uses the TSC21020F processor operating
  at 20 MHz, provides 1.28 Gbits of EDAC protected
  science data memory, and required 6 new FPGA
  designs and 7,650 lines of application code to
  fulfil all of the requirements.
• The payload supported consists of
• The Sira electro-optics Compact High Resolution
  Imaging Spectrometer (CHRIS)
• Two European Micro Cameras (HRC WAC)
• A Payload Autonomous Star Sensor (PASS)
• (Sira/Astrium - derived from the DERA WASS
  programme)
• The ESA Smart Instrumentation Point bus (SIPS),
• and Solid state gyros.
• The systems continues to operate well after 17
• months.

The PPU on PROBA 1 low cost non-redundant
4

Payload Processing PROBA CHRIS data
First view of Belgium The coast and snow The
Astrium Payload Processor (PPU) supports the
commanding, telemetry and data management for
the CHRIS multi-spectral sensor, including the
onboard storage of the image data. The PPU also
supports five other experiments, including the
Sira/Astrium star tracker.
The PROBA spacecraft and the CHRIS sensor are
provided by ESA, Verhaert and Sira
electro-optics. CHRIS Compact High Resolution
Imaging Spectrometer
5
PASS (PROBA) - Image taken 28th JAN 03Astrium
Ltd/Sira electro-optics Ltd Star Tracker
Stellar Magnitudes
7.9
ZANIAH (e VIRGO)
7.3
3.0
6.8
5.8
Initial results before optimisation of thresholds
and application of software updates- NEA better
than 10 arcsec.
5.8
6.0
Approximately 8 of full FOV (Processed image)
6
PROBA AND DATA COMPRESSION

• Data Compression combats
• Large data volumes
• Short ground contact periods
• JPEG Compression available for MicroCameras on
  PROBA
• JPEG
• A lossy compression technique
• Not suitable where all data is required
• Hyperspectral data requires a loss-less technique

7
DATA COMPRESSION ACTIVITIESat ASTRIUM PORTSMOUTH

• Various National ESA Projects
• Algorithm, Software (DSP) Hardware developments
• Camera images, Hyperspectral images, SAR data
• Lossless Techniques, BAQ, JPEG
• BNSC studies
• ESA Advanced Methods for Lossless Compression of
  Hyperspectral Data

8
Advanced Methods for Lossless Compression of
Hyperspectral Data

• An ESA TRP programme (in conjunction with MDA of
  Canada)
• To investigate lossless compression of the
  Hyperspectral Image Data.
• Hyperspectral Image can be defined according to
  the following
• Large number of bands (several hundreds)
• Spectral Bands are contiguous across the spectral
  range
• Higher spectral Resolution in the order of 10 nm.
  

9
Advanced Methods for Lossless Compression of
Hyperspectral Data

• a) Selection of suitable algorithms
• b) Characterisation of selected algorithms
  against
•         Compression Ratio
•         Compression Time
•         Complexity
•         Implementation (software and hardware)
•         System Issues
• i Sensor characteristics
• ii Error propagation
• iii Memory Requirements
• iv Packetisation
• c) Adaptation to Restricted (space) Environment
  of the selected algorithm.
• d) Implementing the selected algorithm on a DSP
  breadboard system

10
Advanced Methods for Lossless Compression of
Hyperspectral Data

•  
•    
•  

• Experimental Trials
• Initial trails in a PC environment have been
  carried out with encouraging results.
  Compression rations from 1.3 to over 3.5 have
  been achieved, depending upon the data set.
• We are now working on a target environment using
  an ADSP21020 DSP Breadboard Demonstrator System
  Results expect in May 2003