Communications – electrical: acoustic wave systems and devices – Echo systems – Returned signal used for control
Reexamination Certificate
2003-05-30
2004-10-05
Pihulic, Daniel (Department: 3662)
Communications, electrical: acoustic wave systems and devices
Echo systems
Returned signal used for control
C367S088000
Reexamination Certificate
active
06801474
ABSTRACT:
FIELD OF THE INVENTION
This invention relates generally to the transformation of the echoes of sonar pings which have been transmitted towards and scattered back from the seabed, so as to improve the classification of sediments on or other attributes of the seabed. In particular, this invention describes a method for sampling the echoes so as to remove the effects of water depth on echo duration, thereby enabling an improved classification of sediments of the seabed.
BACKGROUND OF THE INVENTION
Echo sounders have been used since before World War II to determine the depth of water by measuring the time it takes a sound pulse to travel to and return from the seabed. In a similar fashion, underwater objects such as mines and submarines have been successfully detected by listening for the echoes received in response to pings transmitted by a sounder. During World War 11 the name sonar (analogous to radar) was given to this technology. Further successful uses of sounder technology have included locating fish, measuring the thickness of ice in Arctic regions and oceanographic charting. Echo sounders can also be used to estimate the nature of sediments on the seabed by examining the shape and duration of the echoes of pings returning from the seabed.
The sounders in use today frequently use materials such as piezoelectric crystals (for example, quartz or tourmaline), magnetostrictive materials (for example, iron or nickel), or electrostrictive crystals (such as barium titanate) which change shape under electric or magnetic fields. Such materials can be used both to generate sound pulses (“pings”) by applying an oscillating electric or magnetic field of suitable frequency and to generate electric signals in response to sound energy received. Generally, the pings transmitted are short in duration (less than 1 ms). Each ping is a bundle of sound waves at a frequency typically between 5 kHz and 300 kHz.
The determination of the sedimentary characteristics of the seabed from data supplied by an echo sounder is generally described in “Seabed Classification—A New Layer for the Marine GIS” (J. V. Watt, Quester Tangent Corporation, Sidney, British Columbia, Canada) and in “Acoustic Seabed Classification using Echo Sounders: Operational Considerations and Strategies” a paper presented by William T. Collins and Karl P. Rhynas at the Canadian Hydrographic Conference, Victoria, British Columbia, Canada, 1998. Briefly:
(1) A vessel carrying an echo sounder travels over an area of the seabed. At points along the vessel's path, a ping is transmitted from the echo sounder towards the seabed. Each ping advances spherically from the vessel beneath the surface of the ocean. The ping's wavefront continues spreading until it strikes the seabed below the vessel. At that point, the ping insonifies a small area of the seabed and some of the sound is scattered back towards the surface. As the wavefront continues to advance, it spreads out over a circular (or possibly elliptical or annular) area of the seabed as well as down into sediments covering the seabed, causing further scattering of sound energy back towards the surface.
(2) A detector receives energy scattered back from the seabed. Usually, the detector measures the energy received over a narrow solid angle (less than 10 degrees) which, typically, is cone-shaped. The area on the seabed within the detector's field of view is known as the ping's “footprint”.
(3) The detector samples the analog signal at rates up to 5 MHz, but usually more slowly, and records amplitude values for each ping as a time series. Each time series is typically made up as follows:
(a) An initial peak caused by reflection of the ping from the apparatus of the sounder itself or from bubbles close to the sounder;
(b) An almost silent period corresponding to the round-trip passage of the ping from the sounder to the seabed (the “pre-echo”). During the pre-echo, the amplitude of the signal received by the detector is only background noise.
(c) The echo of the ping.
The starting point of the echo in the time series begins at a time corresponding to the first arrival of sound at the detector after it has been scattered back by that part of the footprint closest to the detector. The echo continues for a period of time corresponding to the wavefront's passage over the footprint and into the seabed sediments, after which time it dies out gradually and the signal returns to background noise.
(d) A second silent period occurring after the echo has died out completely and before the transmission of the next ping (the “post-echo”).
(4) The number of samples in the initial peak and the pre-echo portions of the time series represent the travel time of the ping to the seabed and back. From this number and the speed of sound in water, the depth to the seabed is calculated. The leading values are then removed by a technique known as “bottom picking”. The resulting time series contains the echo of the ping and the post-echo (the “ping time series”).
(5) In addition to the values of depth and the ping time series itself, data are recorded for each ping describing the geographical position and the orientation of the vessel at the time of transmission. These additional data make it possible to associate each ping with a unique survey point on the seabed.
The character of the seabed is estimated by comparing the shape and duration of the echo portion of the ping time series associated with each of the survey points within the survey area. The techniques used are statistical in nature and rely on the calculation of a number of measures (“features”) from each ping time series. Many of the features so calculated make use of the duration of the echo. However, the duration of a ping's echo is affected by the depth and this effect must be removed before valid comparisons can be made.
Conventional techniques to adjust for depth assume that the duration of the echo is proportional to the depth and such echoes are scaled appropriately to compensate for that effect. However, the adjustment of echo duration according to depth is an approximation which only holds well in deep water where the lengthening of the echo is primarily due to larger size of the footprint on the seabed. In shallow water, there is much less spreading and the duration of the ping itself and the penetration of the ping into the seabed sediment are very often the dominant factors. Failure to properly account for these effects introduces artifacts into the data that appear subsequently in the classification as patterns of incorrectly classified sediments.
SUMMARY OF THE INVENTION
The subject of this invention is a method and apparatus for classifying a survey area according to an attribute of the seabed.
More particularly, the classification uses echoes recorded from sonar pings transmitted towards the seabed to calculate a feature vector (typically having more than 100 elements) of statistical measures at survey points along the course of a survey vessel. The feature vectors are used to calculate a smaller number (typically 3) of principal components. The application of the principal components to the feature vector at a survey point produces a Q-vector having as many elements as principal components used. The Q-vectors at each survey point are used in a cluster analysis to optimally assign survey points to clusters. In a final step the clusters are interpreted according to the direct observation of an attribute (such as sediment type, weed cover or roughness) of the seabed at a small number of selected survey points.
The echoes received from deeper depths are usually longer than those from shorter depths. This affects the values of the features calculated and subsequently the classification made. An adjustment must be made so that echoes from different depths which are scattered back from similar conditions on the seabed produce similar feature vectors. The invention described herein provides a method for resampling echoes at a sampling frequency which takes into account the depth of the water, sediment condition
Brinks Hofer Gilson & Lione
Pihulic Daniel
Quester Tangent Corporation
LandOfFree
Resampling sonar echo time series primarily for seabed... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Resampling sonar echo time series primarily for seabed..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Resampling sonar echo time series primarily for seabed... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3267294