Evaluation of the Accuracy of the Dark Frame Subtraction Method in CCD Image Processing

Abstract

This document evaluates the reliability of the dark-frame subtraction method for the detection of very dim targets in astronomical images. This method is frequently used for removing the image background gradient (a thermal artefact) in CCD images. This report demonstrates that this method may not be suitable for the detection of objects with very low signal-to-noise ratio. The analysis of two series of 500 dark frames, acquired at two different CCD temperatures, showed that dark frames are not reproducible with enough accuracy. The subtraction of two dark frames, assumed to be acquired at the same temperature, always leaves a residual background comparable or superior to the noise level. It is suspected that the temperature recorded into the image header is the cryo-cooler temperature and not directly the CCD temperature. The temperature oscillates and there is always a small temperature difference between the CCD ship and the cryo-cooler. However, it was found that the image mean intensity is tightly linked to the background gradient amplitude in each dark frame. The subtraction of dark frames with equal mean intensity, instead of equal recorded temperature, gives good results. Unfortunately, it is not obvious to evaluate the mean background intensity when the image contains signals, while the CCD temperature is always available in the image header. In the case where very faint objects have to be detected, the simple dark frame subtraction method should be replaced by more reliable algorithms (but generally longer to compute) able to separate the signal from the image background.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 01, 2007
Accession Number
ADA479357

Entities

People

  • Mario Lelievre
  • Martin P. Levesque

Organizations

  • DRDC Valcartier

Tags

Communities of Interest

  • Ground and Sea Platforms
  • Space

DTIC Thesaurus Topics

  • Accuracy
  • Acquisition
  • Algorithms
  • Amplitude
  • Cameras
  • Classification
  • Detection
  • Detectors
  • Electronic Circuits
  • Image Processing
  • Intensity
  • Low Temperature
  • National Security
  • Residuals
  • Security
  • Standards
  • Test And Evaluation

Readers

  • Image Processing and Computer Vision.
  • Mathematics or Statistics