Suppressing reverberation by multipath separation for improved buried object detection

Abstract

This paper addresses the problem of using a vertical aperture to suppress the interference caused to SAS by surface bounce path reverberation components decorrelated by the rough sea surface. Using at-sea gathered SACLANTCEN vertical array acoustic data, we characterize the reverberation vertical properties and investigate the separation of the direct and surface bounce path reverberation and target echo components. We propose two implementations of the PCI method of interference cancellation and demonstrate with actual data that they are effective in separating the reverberation in the vertical domain. I. BACKGROUNDThe detection of objects buried in the seabed by sonar has scores of applications, including mine hunting, cable and pipeline surveying, and archaeological research. This is a challenging problem. The physics of sound penetration into the seabed is a complicated phenomena that is governed by the frequency and the angle of the impinging acoustic field relative to the bottom interface. For grazing angles above the critical angle, Snell's theory predicts that sound will refract into sediment albeit with attenuation [l]. Since sediment attenuation [2] increases as a function of frequency, present day high frequency mine hunting sonars operating at hundreds of lcHz are ineffectual for detecting buried mines. However, the situation is better at low frequencies, e.g., in the 2-16 lrHz frequency band. Recently, investigators at SACLANTCEN 113 and elsewhere have found that there is substantial low frequency sound penetration into the sediment at angles above the critical angle and also at subcritical angles. At above critical angles, a SACLANTCEN study [l] found that buried mines were easily detectable in the 2-16 kHz frequency band, even down to a 50 cm burial depth.

Rapid acquisition method of sea vessel radiating transient noise effective time based on the power-law

“…The equation's basic idea can be expressed as: When there is signal, the size of the 2 data points distributed evenly over three Fourier transform, in the case of 1 H established, accurate probability distribution function depends on the transient signal itself. Due to the method is sensitive to the characteristics of transient noise such as beginning time, signal waveform structure and duration time, so, foreign scholars widely use it into underwater acoustic transient signal discrimination.…”

“…For this, a large number of studies have been made in domestic and overseas, and the detection methods of narrow-band [1], Power-law [2], local instantaneous energy density [3] and higher-order statistics [4][5] etc are put forward, these methods have their own scope of application and advantages, most are limited to the maximum peak transient noise time for testing, and unable to provide the beginning and end time. For this reason, this article is based on the power-law algorithm for transient signal waveform of the structure of the "sensitivity", combining the methods of line-spectrum extraction technology, wavelet transform, and second order partial derivative function, to extract the beginning and end noise, theoretical simulation and experimental study show that the algorithm can accurately obtain transient state noise time effectively, get ship scale and characteristics of the transient state noise spectrum, and have good engineering application value.…”

Measurement and assessment method for ship radiating transient noise

“…In order to achieve the goal, we first need to detect transient noise and obtain the generation and end time. In view of this problem, a lot of research work has been carried out at home and abroad, and many related detection are proposed, including narrow band [1],Power-law [2], partial instantaneous energy density [3] ,high-order static [4][5]and so on. Most of these detection methods are limited to the maximum peak value detection, and it is difficult to achieve generation and end time of transient noise.…”

Blind separation of signal and reverberation by minimizing the circular variance of the phase

“…One signal is directly backscattered from the bottom to the receiver and another arrives to the receiver via a surface bounce path, with both arrivals modeled as plane waves. The signal magnitudes in (1 1) are set to a, = 1 and the interference and noise are modeled as complex Gaussian distributed, w i t h 4 = 1 and 0 ' = .01 in (2) and uncorrelated ping-to-ping. The ping-to-ping signal phase errors A&, are modeled as independent and identically distributed zero-mean Gaussian random variables.…”

“…SAS is a high resolution acoustic imaging technique that utilizes the motiou of the sonar platform to synthesize a larger acoustic aperture by coherently combining retums from multiple pings [l]. A major impediment to SAS imaging in shallow water (especially when the range R is greater than the water depth L, that is, R >> L) is interference from strong bottom reverberation components that propagate into the receiving array via surface-bounce multipaths (see figure 1) [2,3].…”