Bad data replacement

Dataflow object

Aim

Echoview feature

Fileset1: Sv V1 telegrams T1

Sv data from transducer. Echogram with missing data. Pings with missing samples are categorized with Bad data regions.

Raw variable

Sv processed data

Apply the line and bad data region exclusion settings specified on the Analysis page of the Variable Properties dialog for its input variable, changing excluded data points to 'no data' values.

Use the Processed data operator. The input variable is Fileset1: Sv V1 telegrams T1.

Processed data operator

Remove missing pings

Remove pings which meet certain criteria.

Use the Reduce pings operator. The input variable is Sv processed data. In this case, pings are removed if their samples have no data values.

Reduce pings operator

Dataflow object

Aim

Echoview feature

Fileset1: Sv V1 telegrams T1

Sv data from transducer. Echogram with missing data. Pings with missing samples are categorized with Bad data regions.

Raw variable

Sv processed data

Apply the line and bad data region exclusion settings specified on the Analysis page of the Variable Properties dialog for its input variable, changing excluded data points to 'no data' values.

Processed data operator

Missing samples created

Create compatible ping geometry for the problem pings.

Set up the required ping geometry with the Resample by number of pings operator. The ping geometry of a good Sv ping can be viewed using Shortcut menu > Show Information > Ping tab. The input variable is Sv processed data.

Resample operator settings:

• All pings in interval, Number of pings in interval = 1, Weighted Mean
• Output = Custom ranges, set Start range, Stop range and Number of datapoints.

Resample by number of pings operator

Fill the missing pings with an average of the ping before and the ping after.

The 3x3 Convolution operator is used to determine this average. The input variable is Missing samples created. The convolution matrix is:

0 0 0

1 0 1

0 0 0

The 3x3 convolution runs the average over the whole echogram (which is now slightly blurry). You only need the sample values for the problem pings.

Create a bitmap to dictate which samples to replace.

Use the Data range bitmap operator. The input variable is Missing samples created.

The operation sets a "true" value for each input value that is within a specified range, and a "false" value for each input value that is outside the specified range. "No data" values are always translated to "false".

Minimum in-range value = -999.00

Maximum in-range value = 999.00

Create an echogram by sourcing data from the original variable where available and from the averaged variable where it is not.

Use the Select operator.

The input variables are:

• Operand 1 = Missing samples created
• Operand 2 = Missing pings filled
• Operand 3 = Sv sample range

The resulting echogram contains the same number of pings as the original echogram. Problem samples are replaced by an average from either side of the problem ping.