As part of the OERA Pathway Program, SOAR is conducting research to help evaluate the effectiveness of multibeam imaging sonars for monitoring marine animals in tidal channels.  This information will help inform DFO, tidal energy developers, and other stakeholders in the design and implementation of effective monitoring systems for tidal energy projects in the Bay of Fundy, and beyond.

Multibeam imaging sonars are essentially the closest technology to echolocation used by marine animals including dolphins and porpoises.  The sonars send out pings (from multiple beams) and process the time and intensity of the returns (backscatter) as images of the subsea environment.

Two imaging sonars are being tested:

• Tritech Gemini with 720 kHz frequency, 50 m range, 120 degree field of view, and 20 degree vertical beam spread.
• Teledyne BlueView with 2250 kHz frequency, 10 m range, 130 degrees field of view, and 20 degree vertical beam spread.

These 2 sonars are world class technologies, and were selected to test the ability of the BlueView to identify and track targets at close ranges (less than 10m) and the Gemini to detect and track targets at relatively further ranges (out to 50m).

The sonars were mounted on a frame with autonomous power supply and data storage (Figure 1), then deployed on the seabed in Grand Passage (approx. 25m depth) for a flood tide.  The targets were suspended beneath research vessel Puffin (Figure 2), which drifted though the ensonified area of the sonars (Figure 3).  The targets are man made, including 1 lb lead fishing weight, fist sized basalt rock in a lobster bait bag, and the DOT V-Wing (Figure 4).

Algorithms are in development for automated data analysis, but they are not yet validated or reliable.  They also require site (data set) specific training.  So, we are basing the evaluation on manual analysis of the data files including YOUR INPUT!   Throughout this analysis you are not being tested.  The evaluation is on the ability of these sonars to detect and track the targets, as determined by a group of human observers.

NOTE: As shown on Figure 3 the ensonified area of the BlueView is very small.  As such, it was difficult to get targets within it.  We had a few successes, but not sufficient for statistical analysis of manual review.  Only data from the Tritech Gemini are included in the following instructions and analysis. 

THANK YOU FOR PARTICIPATING IN DEVELOPMENT OF EFFECTIVE ENVIRONMENTAL MONITORING SYSTEMS FOR TIDAL TURBINES!

Figure 4: Targets, including 2) 1 lb lead fishing weight, 3) fist sized basalt rock in a lobster bait bag, and 4) DOT V-Wing. 

Notes

• Target 1) 1 inch tungsten carbide sphere was used in vessel mount sonar project but was not readily distinguishable from target 2
• The squid jig was used in initial testing with a 1 lb lead weight inside.  It exhibited interesting behaviour, but for the purpose of this experiment the increased drag made it difficult to handle while being very acoustically similar to target 2.
• As such, both were removed from further analysis.

• Data are provided below in video format, which you can pause and replay as you wish. The data playback speed is set to twice as fast as real time to increase efficiency.
• All videos should be viewed full screen at 1920 x 1080 resolution.
• You will watch a training data set for the Tritech Gemini, where the target type is provided and the target is marked with a red circle that follows/tracks it through each example. Following the training data, you then watch sets of files in the test data sets and record your observations in the spreadsheet provided.
• After each “file” included in the video, hit pause and enter your observations into the spreadsheet.
• Your observations will be quantitative, but subjective. You will provide:
  • Yes/no as to if there is a target included in the data file
    • If yes, your best interpretation of the target type in comparison to those in the training data set along with a ranking of 1 (low) to 5 (high) on your certainty.
  • The range over which you viewed the target (min and max)
  • Scoring on ability to detect and track the target from 1 (low) to 5 (high)
  • Notes on the track of the target. For example, does it follow a straight line, stop at points, etc.  Anything you think useful for describing the trajectory of the target.

To start, please have a look at the spreadsheet then watch this video showing the frame deployment in Grand Passage.

Ready to go?  If so, please proceed to the Tritech Gemini training data set:

Then proceed to these test files for the Tritech Gemini: