Method and system for wideband beam forming based target localization

A system, comprising:

• one or more hardware processors;
• a communication interface;
• and
• a memory storing a plurality of instructions, wherein the plurality of instructions cause the one or more hardware processors to:
• receive, via the communication interface, an input data comprising a) cursor coordinates indicating a selected location in a X/Y cartesian plane, of a pipe being monitored, b) one or more N-channel signals from an N-microphone circular array of window length in seconds, c) geometry of a target, d) a selected resolution of an acoustic imaging viewport, e) a frequency range of interest, and f) a band distance;
• convert the cursor coordinates to associated polar coordinate, based on the geometry of the target and the selected resolution of acoustic imaging, wherein the polar coordinates represent location in angles from a positive x-axis and a positive z-axis, and a distance from origin;
• perform a wideband beamforming based on the geometry of the target to generate a composite acoustic map and a set of single frequency acoustic maps, by:
• shortening an N-channel audio frame from the N-microphone circular array from the start of a frame to generate the updated value window length, to reduce the computation time of the wideband beamforming;
• increasing the narrow bandwidth of each of the one or more narrow band pass filters centered at frequencies of the set of single frequency acoustic maps, to decrease the number of narrow band pass filters without changing a frequency range of the wideband beamforming;
• and
• reducing resolution of the acoustic imaging viewport along horizontal and vertical axes, to reduce the computation time of the wideband beamforming, retaining intelligibility of the acoustic map in the viewport, wherein the updated value window length, the decreased the number of narrow band pass filters, and the reduced resolution of the acoustic imaging viewport causes the wideband beamforming;
• identify a microphone from among a set of microphones in a microphone array, nearest to an angle of interest derived from generated polar coordinates, wherein the identified microphone that is nearest to the angle of interest captures maximum of a sound from an event of interest in the angle of interest;
• process a single channel audio frame from the identified microphone that is nearest to the angle of interest, using a light weight Delay Multiply and Sum (DMAS) algorithm, to generate a spatially filtered single channel audio frame;
• fuse the spatially filtered single channel audio frame with the set of single frequency acoustic maps, by frequency domain convolution of spatially filtered single channel audio frame with one or more interpolated and zero padded weightage coefficients from the single frequency acoustic maps at the cursor coordinates, to generate a spatial-spectral filtered single channel audio frame;
• and
• perform a target localization of the event of interest in the angle of interest, by fusing the spatial-spectral filtered single channel audio frame with the composite acoustic map.

One or more non-transitory machine-readable information storage mediums comprising one or more instructions which when executed by one or more hardware processors cause:

• receiving, via the communication interface, an input data comprising a) cursor coordinates indicating a selected location in a X/Y cartesian plane, of a pipe being monitored, b) one or more N-channel signals from an N-microphone circular array of window length in seconds, c) geometry of a target, d) a selected resolution of an acoustic imaging viewport, e) a frequency range of interest, and f) a band distance;
• converting the cursor coordinates to associated polar coordinate, based on the geometry of the target and the selected resolution of acoustic imaging, wherein the polar coordinates represent location in angles from a positive x-axis and a positive z-axis, and a distance from origin;
• performing a wideband beamforming based on the geometry of the target to generate a composite acoustic map and a set of single frequency acoustic maps, by:
• shortening an N-channel audio frame from the N-microphone circular array from the start of a frame to generate the updated value window length, to reduce the computation time of the wideband beamforming;
• increasing the narrow bandwidth of each of the one or more narrow band pass filters centered at frequencies of the set of single frequency acoustic maps, to decrease the number of narrow band pass filters without changing a frequency range of the wideband beamforming;
• and
• reducing resolution of the acoustic imaging viewport along horizontal and vertical axes, to reduce the computation time of the wideband beamforming, retaining intelligibility of the acoustic map in the viewport, wherein the updated value window length, the decreased the number of narrow band pass filters, and the reduced resolution of the acoustic imaging viewport causes the wideband beamforming;
• identifying a microphone from among a set of microphones in a microphone array, nearest to an angle of interest derived from generated polar coordinates, wherein the identified microphone that is nearest to the angle of interest captures maximum of a sound from an event of interest in the angle of interest;
• processing a single channel audio frame from the identified microphone that is nearest to the angle of interest, using a light weight Delay Multiply and Sum (DMAS) algorithm, to generate a spatially filtered single channel audio frame;
• fusing the spatially filtered single channel audio frame with the set of single frequency acoustic maps, by frequency domain convolution of spatially filtered single channel audio frame with one or more interpolated and zero padded weightage coefficients from the single frequency acoustic maps at the cursor coordinates, to generate a spatial-spectral filtered single channel audio frame;
• and
• performing a target localization of the event of interest in the angle of interest, by fusing the spatial-spectral filtered single channel audio frame with the composite acoustic map.