New Bottom Detection Algorithm
A new bottom-detection algorithm is used in gautoedit. It has three
parameters:
The first is the size of the amplitude bump. It is essentially the
same as the amplitude bump used in the waterfall editing to flag a possible
bottom signal. The default (25) is good for bottom detection in deeper
bins. You may need to make it smaller (eg 15) to detect the bottom in
shallower bins
The second parameter indicates how much smoothing is done (in the vertical)
prior to looking for the bump.The default (2) is good for deeper bins:
if you are trying to detect the bottom in shallower water, you may need to
do less smoothing (eg 1)
The third parameter indicates how large a window to use to detect the
amplitude bump. The default value (5) is good for deeper bins: you may
need to look closer (eg 2) if you're working near the surface.
Summary: the defaults values are good for deeper bins. Because of
the geometry, you might try scaling all three values together like this (as
a reasonable starting point):
bin number
|
AMP THRESHOLD
|
smoothing
|
window
|
1-2
|
10
|
0
|
2
|
5-15
|
15
|
1
|
3
|
20-40
|
25
|
2
|
5
|
The effect on UABS (ocean u) is shown below with different parameters:
....with no flagging:
... with AMP=40, smoothing=2, window=5:
... with the defaults (AMP=25, smoothing=2, window=5):
.... and with AMP=15, smoothing=1, window=3):
The full panel plots are available below with different sets of parameters:
AMP=40, smoothing=3, window=5
AMP=40, smoothing=2, window=5
AMP=40, smoothing=2, window=2
AMP=40, smoothing=1, window=5
AMP=40, smoothing=0, window=5
AMP=25, smoothing=2, window=5
AMP=15, smoothing=1, window=3