After basic image processing, the next step is to determine the bubble size and construct a bubble distribution from the resultant time series of bubble sizes (and locations).
Extraction is necessary if the captured image contains both A and B fields,
|
Raw Image (A&B Fields)
|
and arises from the way in which video cameras (not digital) read pixel elements in an interlaced manner. In extraction, each missing line is replaced by the interpolated value from neighboring lines. After extraction, the image becomes...
|
Extracted Image
|
|
Background Field
|
Result after Subtraction
|
Background subtraction can improve the image by removing variations in intensity across the image. (NOTE It is always better to remove these intensity gradients before digitizing as they represent information loss). The background can be obtained from a frame without the bubble in the field of view, or created by averaging together a large number of frames. After subtracting the backgroung, the image to the immediate left is obtained. |
| Often the captured image is not centered in intensity so that the maximum number of bits are used to span the intensities of interest. This can be corrected by adjusting the brightness and contrast. The contrast may be adjusted linearly, or logarithmically, also called adjusting the gamma. The log down adjustment increases contrast in the bright intensities, making thresholding easier, at the sacrifice of the darker portions of the bubble. These are preserved better by the log up adjustment. The most pleasing to look at is with the linear adjustment. | Log Down Adjust |
Log Up Adjust
|
Linear Adjust |
| It is possible, although not necessarily meaningful to acquire information from the image on the subpixel level through interpolation. The procedure involves interpolating the picture to a greater resolution and if the method used is non-linear, such as a cubic spline method, also incorporates some noise reduction too. | Linear Interpolation / BiCubic Interpolation.
|
