Once you’ve unraveled the data you can enthusiastically start with the cleaning process. You may choose to load up all the data with:

swathed merged/sub_dir/*merged

Now you’ll be cleaning line by line, but interpreting the data is easier if you already have a grid of the (raw) data. So it’s best first to make a DTM and then examine the data and it’s surround topography.

I) getBounds

When you make a mapsheet it requires the boundaries of you data set, use getBounds to get this information. The program creates files with a .100ping_bounds extension. These are used later in the pipeline.

getBounds merged/sub_dir/filename.merged

II) make_blank

You can compare the creation of a mapsheet with pulling a sheet from the drawer, trimming of the edges to get the desired dimensions and drawing gridlines op it. The result, a .blank file, is still empty but we’ll fill it later.

make_blank dtm_name –boundsof merged/sub_dir/*merged

-boundsof

this option tells the program to use the boundaries of the merged data, and use the same ellipsoid and projection. You have to direct to the merged files.
The program will ask you for the pixel (grid) size. Without the boundsof option you need need to specify the ellipsoid, projection and bounding points of the data set.

III) tor4

Convert the mapsheet (.blank-file) to a .r4-file. The r4 will ultimately be the actual dtm with the data. tor4 creates two other weighting files which are used by the filling process.

tor4 dtm_name

IV) weigh_grid

Now you can fill the r4 with the bathy data:

weigh_grid –beamwidth x.x –custom_weight weight_file dtm_name merged/sub_dir/*merged

-beamwidth

takes the beam dimension into account

-custom_weight

requires a weight_file for weight of the complete swath

dtm_name

the r4-file without extension

merge/sub_dir/*merged

direct to the merged files that contain the bathy data

Creating a beam weights file:

makeWts - grid

The program will ask for the number of beams and the weights. Weights run from 0 to 1.0. It outputs a ascii file MOS_weights which you may want to rename to a more meaningful name, e.g. Reson8125_weights.
A bug in the program screws up the weight of the final beam, you’ll have to fix that manually with a text editor as kwrite of vi.

V) swathed

View the data in swathed with your new grid:

swathed –locate dtm_name merged/sub_dir/*merged

You can delete points by simply drawing a box around them, but you can also use some filters. Jonnies page has a complete description of all the buttons in swathed and the filters. Here follows a short description on two frequently used filters, the median and fluffer.

Median
A median filter is ideal to remove spikes cause it works on the assumption that solitary points far from the median are outliers. Apply the median filter to each data package (with spacebar you scroll trough the data 80 pings at the time) by hitting ‘m’ in the swath or longitudinal profile windows. The median filter has two adjustable parameters, the box size and a cutoff.
Within a box it flags a point when it’s further than a percentage of water depth away from the median. Define the box size with ‘<’ and ‘>’, and the cutoff with ‘%’.
You’ll make the filter more tolerant with a large boxsize and cutoff.

Fluffer
The fluffer flags all the soundings that lie outside a multiple of the standard deviation. Fluff the data with ‘f’, set the number of standard deviations with ‘F1’ ‘F2’ ‘F3’. For example to flag all the data destructively use the fluffer with 3 times the standard deviation (‘F3’).

---

next 3) Final product
back 1) Unraveling

Home