Keywords : quality, control

Quality Control

Thanks to its "Multilayer" management, SonarScope is particularly effective to conduct "Quality control" work on sonars and sounders data. One can measure the influence of one or several parameters on the bathymetry or the imagery. The following example shows the solving of a problem met on a sonar EM1002.

This is an example of Kongsberg EM1002 backscatter data from the Canadian Hydrographic Service: Horiz. Axis: beam number Vert. Axis: ping cycle

On this image, one can see two distinct artifacts:

  • Vertical stripping which seems to be beam-dependent
  • Change in reflectivity (vertical banding) which is transmit beam dependent (see also following slide)

This is the image representing the transmit beam layer. At this depth range, the EM1002 fires three acoustic bursts in three different directions. The acoustic burst are made sequentially, each with its own frequency.

Here, the mean(*) for each beam has been calculated and grouped by transmit beam number. One can clearly see the steps and differences in reflectivity associated to the transmit beams. This mean will be used to compute a bias curve for the swath.

(*) It is a « biais curve » : the mean curve minus the global average.

Here, a statistical compensation has been applied: for each pixel: the value was corrected (subtraction) using the corresponding value on the bias curve.

This compensation seems to be giving good results and we shall assume the bias curve can be used to correct the entire data set.

The CHS used these images to demonstrate to the manufacturer the sonar was not operating as expected. The boards in the transceiver have been changed, which corrects the situation.

Following this analysis a maintenance has been done on the sounder.