Buried object scanning sonar for AUVs

Abstract

Abstrocf-A 252 channel FM sonar is developed to generate images of objects buried in sediments using reflection tomography. An omnidirectional source, transmitting FM pulses over the band of 2 to 12 IrHz, illuminates buried targets.The backscattered signals are measured with 252 hydrophones and processed with a digital matched filter. Coherent nearfield focusing generates a 3D map of acoustic intensity for each transmission event. As the sonar approaches and passes buried targets, the processor generates 3D matrices of acoustic intensity, overlapping in space and referenced to a coordinate system fixed to the seabed. The eo-located pixels are added incoherently to generate a multi-aspect image of the target. The change in vehicle position between transmissions is measured with a DVL(dopp1er velocity log) and IMU (inertial measurement unit). The resulting reflection tomographic images provide target shape information nsefnl for target classification. The sonar can be mounted on small AUVs by replacing the 252 channel, 1.5 m diameter array with one meter long line arrays mounted as wings. The receiver aperture is generated synthetically using the near field focusing processor which performs time delay focusing based on the positions of the hydrophones in the line array and vehicle motion data. A comparison between synthetic swath and tomographic images of a rigid spherical target in water shows the improvement provided by reflection tomography.

A Multichannel Canonical Correlation Analysis Feature Extraction with Application to Buried Underwater Target Classification

“…Target backscattering is measured by the 252 hydrophones (or a subset) and is then processed by a CPU to generate a final composite image. BOSS processing involves a sequence of operations that includes chirp sonar processing, 3-D near field focusing using data from a real or synthetic receiving aperture, transmission beam steering that allows target illumination at a broad range of aspect angles, real-time 3-D image construction of buried objects showing burial depth, shape, and orientation and location of buried objects, and real-time buried object detection [9].…”

Buried Underwater Target Classification using the New BOSS and Canonical Correlation Decomposition Feature Extraction

“…This was accomplished by maintaining a constant transmitter and receiver grazing angle as the BOSS vehicle was towed over the target field. In these studies, it was found that different grazing angles provided unique shape information about the target, and hence it was concluded that a wide variety of grazing angles should be used during image construction to capture the shape of the buried object [12].…”

“…Target backscattering is measured by the 252 hydrophones (or a subset of these) and is then processed by a CPU to generate a final composite image. BOSS processing involves a sequence of operations that includes chirp sonar processing, 3D nearfield focusing using data from a real or synthetic receiving aperture, transmission beam steering that allows target illumination at a broad range of aspect angles, real time 3D image construction of buried objects showing burial depth, shape, and orientation and location of buried objects, and real-time buried object detection [12].…”

“…The data used for this study was generated by the Buried Object Scanning Sonar (BOSS) [12]. Section II describes the new disk BOSS system as well as the data collection process.…”

Synthetic Aperture Processing of Buried Object Scanning Sonar Data

“…Previous publications on BOSS processing discuss strip-map [1] and multi-aspect processing [2,3] based on time-delay focusing of data from a large aperture hydrophone array. In these approaches the resolution of target imagery is limited by the dimensions of the hydrophone array.…”