Chesapeake Times, Vol 8 | January 2022
When we have a new release here at Chesapeake, the software goes through several different types of tests. Whether through automated (Ranorex) or manual testing, we look to ensure the update works as expected. This all works well, until we need to test our real time servers. Simulators and playback only work up to a point, and it isn’t until we find ourselves on a boat with a real system that can we validate the driver and the updates to the program.
Recently, I spent two days with Mike and Nick at Seahorse Geomatics in Portland OR, to run through our multibeam collection using their NORBIT system. I set out a few goals to validate the driver, from “does it work” to “can I break it with a lot of data”. One test was to have the multibeam pinging at 50Hz with 1024 beams at full swath, with all ancillary data and collection of Velodyne LiDAR data too. The driver and software performed as expected, and with 44GB of data collected, the next test was to validate the data itself.

To ensure our collection was correct, we set up a test with 3 systems logging simultaneously, all listening in to the Norbit GUI: SonarWiz, Norbit DCT and Qinsy QPS. This was done on a single computer with a solid state hard drive. I imagine any spinning disk would have trouble keeping up with the amount of data being logged.

Our survey plan included two calibrations – a bore sight test for the LiDAR and a Patch Test for the multibeam, along with a Performance Test of the system. We took note of the high current in the Columbia River, and did the test relatively quickly, to reduce any influence from the dynamic movement of the sand waves. Reciprocal lines had less than 45 second between passes.
The survey area was on the Columbia River, just north of the Portland Airport:

Our survey plan included two calibrations – a bore sight test for the LiDAR and a Patch Test for the multibeam, along with a Performance Test of the system. We took note of the high current in the Columbia River, and did the test relatively quickly, to reduce any influence from the dynamic movement of the sand waves. Reciprocal lines had less than 45 second between passes.
Data collected in SonarWiz were the CDF files, XTF files and S7K files. The S7K file option in the server provided a bit of redundancy for the data collection, and could easily be compared to the files collected with the other software.
Data editing was done in the Swath Editor (single line) and Area Based Editor (multiple lines) to flag any bad soundings.


Editing in the Area Editor with 5 lines
Once cleaned, we performed a Beam Performance Test on the data. Subtle depth issues that were not easily seen in the color image could be computed by this test.
Area used for Patch and Performance Tests. Main scheme lines were run east-west, with the crossing (check) line in the north-south direction.

Beam performance test results showed the entire system would pass Special Order Performance for the full swath. It is interesting to note that less than 1000 points – out of almost 300,000 – would fail to meet the criteria, or less than 0.3% of all data.

The results computed by SonarWiz collection and processing was compared to the data collected and processed in the other software. Overall, the agreement matched well, with differences less than a few centimeters. This comparison – an independent check on the data – proved the system data collection through data processing was correct, and the updated server and software passed the validation test.
The Multibeam Server will be part of the next release, SonarWiz 7.9.
– Harold Orlinsky, General Manager