Dalsa CALIBIR 640 GigE, Руководство пользователя

Страница: 38 / 105

38 • Features and Configuration Options Calibir 640 GigE IR Camera User's Manual
Adaptive Contrast Enhancement
The native output of the Calibir sensor is 14-bits. Adaptive contrast enhancement (range
compression) is implemented when the pixel output is 8-bits. Adaptive contrast enhancement
automatically adjusts the range in the 14-bit image to map to 8-bits to output high quality, well-
contrasted images, regardless of the temperature differences found in the scene.
Note: For analog output, the camera output pixel format is always 8-bit, therefore
adaptive contrast enhancement is always implemented; for the GigE Vision digital output,
the camera output pixel format can be set to either 14 or 8-bits (14-bit output is
calibrated but does not use range compression).
The 14-bit image is analyzed to determine the average pixel value. Using this average value, the
Range Compression Contrast setting determines the number of standard deviations around this
average value that are mapped to 8-bit values. Lower values provide less contrast (less stretching
of values) while higher values stretch a smaller range in the original 14-bit image. Recommended
values for most applications is between 100 and 150. When contrast is set to the maximum value
(255), the image is effectively binarized (pixels above the average value in the 14-bit image set to
255 and those below the average value set to 0).
When Range Compression Contrast Limit Mode is enabled, it limits the action of the contrast
compression to a certain amount. The higher this value (between 0-255) the less contrast the
scene reaches for a given temperature difference. This is useful for very low contrast scenes where
the range of values in the 14-bit image is very small and the contrast stretching would otherwise
introduce noise and other image artifacts due to the small amount of data present. Recommended
value for most applications is between 60 and 100.
Range Compression Constrast
(higher values tighten this range from
the average pixel value)
Range Compression Constrast Limit Value
(higher values widen this limit)
All lesser values mapped to
0 (black) in 8-bit output
All greater values mapped to
255 (white) in 8-bit output
Figure 16: Range Compression Contrast