www.photometrics.com ©2021 Teledyne Photometrics. All rights reserved.
14
CHAPTER 4
Intelligent Quantification
Embedded Image Processing Algorithms
The
Teledyne Photometrics
Retiga R Series Mono camera includes user controllable features
designed to enhance camera performance and image quality.
Defective Pixel Correction
QImaging only uses sensor manufacturers with the highest
grade CCD sensors in your
Teledyne Photometrics
Retiga
camera. However Manufacturer’s specifications do not always
meet customer needs for defect free imaging. This is
particularly true for “hot” pixels – pixels that generate dark
current at a far higher rate than average. The converse is also
true, e.g. certain pixels response to light is far lower than
average, known as “dark pixels”. In both cases, suitable
threshold can be determined for defective pixels, and their
locations stored in the camera during manufacture. When
Defective Pixel Correction is enabled, the signal in each defect
location is replaced by the average of the pixel immediately
preceding and following the defect. The
Teledyne
Photometrics
Retiga camera can store up to 2048 defects
locations.
The factory default method for determining defective “hot”
pixels is to take a pair of 10 minute dark images. After
removing the bias offset, the 1800 “hottest” pixels are
selected as potential defects. A suitable floor is chosen so that
the algorithm does not accidentally catch the tail of the
normal dark current noise distribution. If the same location is
defective in both dark images, its location is programmed into
the camera as a defect.
To determine defective dark pixels, the camera is exposed to
uniform light, filling the pixels to 50% saturation. A median
filter is applied, and the original image is subtracted from the filtered image. Any pixels that have a signal level
of 40% or more of saturation, are deemed defective. Up to 200 dark pixels are stored for correction using this
method.
Defective Pixel Correction: OFF
Defective Pixel Correction: ON