4.6.7 Mean Profile Filter

 
Basic Description: This filter removes stripes that do not run the entire length of the survey.  In many cases it does a better job than balancing the mean of entire transects (as is done with a "Zero Mean Traverse" operation).
 
Technical Details: The filter first smoothes out the lines to emphasize them using a one dimensional low pass filter in the direction of the stripes. Then, these stripes are isolated using a high pass filter in the opposite direction (across stripe size). Next, the resulting image, which shows mostly stripes and no other anomalies, is subtracted from the original.  Along stripe size is the size of the stripes that you want to get rid of. The across stripe size is the high pass filter that is done in the second part of this operation.  Beware, this filter also removes anomalies you want to keep if they are oriented with the survey direction and occupy single transects. 
 
Instructions: After choosing a selection method and making your selection on the survey, enter parameters in the following order (go to each section for details):
 
4.6.7 Mean Profile Filter
1

Stripe Length

1. Stripe Length
Set to the average length (in samples or pixels) of the stripes that you wish to remove.
2

Across Stripe Size

2. Across Stripe Size
For removing stripes that occur along single lines, use 3. Increase the number for fatter lines.
3

Stripe Angle

3. Stripe Angle
Enter 0 for stripes that go north-south (up and down), or 90 for horizontal stripes. Other angles can be used if there are stripes occurring at different angles, although 2D Fourier is generally better for removing periodic linear features such as plow marks.
 
Last Updated June 15, 2011