Bottom Tracking

One of the basic requirements for a side scan sonar mosaic is to correct the imagery for distortion caused by the height of the sonar above the seabed, known as "water-column removal" or "slant range correction" this link has more information on the subject:  To make the correction we need to know the towfish altitude above the seabed, in some cases the sonar may measure the altitude directly, and it will be stored in the XTF file, in other cases the bottom may have been detected by the acquisition software and tracked during recording.  There is a tool in Perspective that allows the user to track (or re-track) the bottom so that the altitude may be correctly determined.

The guide can be read from the strart for the whole procedure or you may jump to each section:

Checking the Status of the XTF file.

Re-Tracking the Bottom

Other Settings

Checking the Status of the XTF file

Before we make a mosaic we need to check whether the XTF file has valid towfish altitude stored in the file.

Start Perspective and Import a Raw Data (XTF) file:

Answer No to the next question (we only want to view the file at this stage)

The file is imported and displays only the track of the vessel in the Map Window:

To view the file in the Perspective Waterfall either double click directly on the vessel's trackline in the Main window or on the actual file name in the navigation section of the tree view  (green)

    The Waterfall window opens, use the slider at the bottom of the window to scroll through the line, note that in this file the altitude is recorded as zero.  Meaning that the altitude will need to be tracked before a mosaic can be constructed.

Re-Tracking the Bottom

Click the Bottom Track Tool in the Waterfall Tool Bar:

The bottom tracking tool starts:

The window initially displays the whole line compressed, and the bottom tracking method is OFF. Select the following:

  1. Method: Amplitude
  2. Level: as required (start with about 25%)
  3. Holdoff:  as required, choose a value that does not intersect the bottom (first return) at any point (green line)
  4. Click Compute to have the program try and track the bottom:

Try different values for Level, and hit Compute to re-track.  As in this example there may be places where the auto bottom tracking does not work.

To correct the problem uncheck the Compress checkbox to expand the image to its full resolution across the screen and use the << and >> buttons to move to the area of interest, now:

  1. Select Method > Manual
  2. Right Click and draw a selection rectangle around the area to correct.
  3. Hit Clear

The bottom track values inside the selection box are now set to the last accepted value on the left side of the selection box:

Working within the previously selected window starting at the LEFT side, use the LEFT mouse button to click on the first return, the red bottom track line will re-locate to the new position:

When everything is OK hit Apply:

Re-open the file in the waterfall window to check the bottom tracking (red line).  Note that the altitude displayed refers to the horizontal magenta line across the center of the waterfall window:

NOTE:  At this point nothing has changed in the original XTF file, modifications of this type (Bottom Tracking, Navigation etc). are made ONLY in the Cache file (.XTF_idx) that Perspective creates.

Other Settings

Channel

In the Bottom Tracking window use the Channel selector to select the channel with the least noisy (blackest) water column, the auto tracking routine will work better.

Gradient Option

There is another option for tracking called Gradient that can be used.  The method described above was Amplitude and this is the most common method, however it can fail under some conditions.  The method assumes that the strongest return will be from Nadir (the point closest to and directly below the tow-fish)  However in some cases there may be other reflectors that cause problems:

In this case there are some very strong reflectors (steel pilings); the Amplitude tracking algorithm sees them as being a stronger return, it would of course be possible to track the whole line manually, however under these conditions the Gradient option may give a better result:

Because this is a completely automatic process there are no user controls, clearly it suffers from a need to "Search" for the first return, which causes an anomaly at the very start of the line.  Also it tends to track a little "deeper" than the Amplitude method. And can be affected by other changes in the image.