2D versus 3D Sonar Processing
(WARNING - Technical Jargon ahead!)
FarSounder
has approached sonar system design fundamentally different than other sonar
technologies. The solution FarSounder’s technology provides is able to deal with
multi-path and water depth issues, allow for fast refreshes, and create large
field of view single pings, all in a small cost effective package. FarSounder’s
approach does not hurt marine life and can be used as a marine mammal mitigation
device.
FarSounder’s single ping, 3D sonar approach enables the system to overcome
traditional forward-looking sonar limitations such as Multipath Interference,
Shallow Water Operation, Roll/Tilt Compensation, Surface Effects and Ship
Motion.

Figure 1: 2D
processing of pier wall in shallow water.

Figure 2: 3D processing of same data set. Pier wall colored red for shallow.

Figure 3: 3D image rotated to show depth perspective. |
When used
as a navigation device in shallow water, other 2D technologies often fall short
of being effective navigation tools. These systems are very good at finding the
range and bearing to a single target in deep water. However, by definition,
there are targets at every range and bearing in shallow water. Those 2D systems
cannot distinguish between the safe sea floor and the dangerous in-water
obstacle through depth measurement. They must rely on visual cues to guess
whether or not the obstacles are located in the water column or on the sea
floor. The need for visual cues forces those 2D systems to have very high
horizontal angular resolution. Unfortunately, this causes the system to require
many listening channels and requires a footprint for the receive array or lens
aperture that is very long in terms or wavelengths. This long aperture
requirement drives the system to a higher frequency which in turn excludes these
systems from long range capabilities.
To
illustrate difference between these two approaches consider the following
figures.
In figure
1, real data of a pier wall bulkhead collected with a FarSounder 3D sonar system
is processed as a simple 2D horizontal slice. In this display, color is mapped
to signal strength where red is “loud” and blue is “quiet”. The angular
resolution is roughly equal to the existing off-the-shelf Forward-Looking Sonars
(FLS) -- also called Collision Avoidance Sonar Systems (CASS) or Obstacle
Avoidance Sonars (OAS) -- manufactured by various companies. There is clearly a
feature caused by the pier wall crossing the field-of-view at about 45° (noted
with the dotted line). Yet there is also another feature or “bright” spot noted
by the oval. As a navigation tool, this 2D display is not very effective.
In figure
2, this same ping has been processed with FarSounder’s 3D processing using
patent pending Target Model technology. In this image, color is mapped to depth.
The view is a look-down orthographic projection, yet targets are actually
plotted in 3D space. Note the clear presence of the pier wall (shallow) in red
and the sea floor (deep) in blue.
In figure
3, the same data is plotted in 3D space with a rotated perspective view. Note
how clear it is to see the navigation hazards in 3D with this processing
technique.
Without 3D
capability, forward-looking sonars are unable to easily compensate for roll and
tilt without large amounts of expensive hardware. Even in these cases, 2D roll
and tilt compensation is marginal at best. FarSounder’s technology is capable of
compensating for roll and tilt entirely in software with a simple, inexpensive
roll and tilt sensor.
FarSounder’s navigation technology can be adapted for use in target
classification. This adaptation would make the system useful for swimmer
detection, organic mine detection ahead of a vessel, and fisheries research.
Many of the traditional 1D and 2D sonar classification techniques can also be
applied to FarSounder 3D technology. Also, the depth information is an entirely
new feature domain which can add real value to improving the false alarm rates
often associated with current approaches. FarSounder’s technology is appropriate
for operating on spectral, 3D geometrical, 3D spatial and temporal features.
Already, FarSounder’s Target Model technology processes on basic 3D geometrical
and special features to perform simple classification between the sea floor and
in-water targets. Additionally, FarSounder has signed a Cooperative Research and
Development Agreement (CRADA) with the Naval Undersea Warfare Center (NUWC) in
Newport, RI specifically with one of their spectral classification groups.