Method and system for local interpolation function based sar imaging for irregular scanning geometries

A system, comprising:

• a memory storing instructions;
• one or more communication interfaces;
• and
• one or more hardware processors coupled to the memory via the one or more communication interfaces, wherein the one or more hardware processors are configured by the instructions to:
• receive (i) a first raw radar data collected at a set of sparse non-uniform spatial coordinates, and (ii) a second raw radar data collected at a set of dense uniform spatial coordinates, for each object amongst a plurality of objects in a training dataset;
• generate a noisy synthetic aperture radar (SAR) image, from the received first raw radar data for each object amongst the plurality of objects using a Non-Uniform Fast Fourier Transform (NUFFT) based Range Migration Algorithm (RMA) (NUFFT-RMA) technique;
• generate a ground truth SAR image comprising a set of coordinates from the received second raw radar data for each object amongst the plurality of objects using a Fast Fourier Transform (FFT) based RMA technique;
• and
• iteratively train a Local Interpolation Function (LIF) neural network by updating an encoder weight and a decoder weight associated with an encoder and a decoder comprised in the LIF neural network, until a termination criteria is met, to obtain a trained LIF neural network, wherein the step of iteratively training comprises:
• generating an LIF feature map from the noisy SAR image for each object amongst the plurality of objects, using the encoder, wherein the LIF feature map comprises a latent representation for each coordinate of the set of coordinates;
• generating a predicted SAR image from the LIF feature map using the decoder by querying each coordinate of the set of coordinates;
• computing an L1 loss between the predicted SAR image and the ground truth SAR image;
• and
• updating the encoder weight and the decoder weight based on the L1 loss;
• wherein the termination criteria are one of (i) an L1 loss of a previous iteration is less than or equal to an L1 loss of a current iteration, or (ii) completion of a predefined number of iterations.

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 an input/output (I/O) interface, (i) a first raw radar data collected at a set of sparse non-uniform spatial coordinates, and (ii) a second raw radar data collected at a set of dense uniform spatial coordinates for each object amongst a plurality of objects in a training dataset;
• generating, a noisy synthetic aperture radar (SAR) image from the received first raw radar data for each object amongst the plurality of objects using a Non-Uniform Fast Fourier Transform (NUFFT) based Range Migration Algorithm (RMA) (NUFFT-RMA) technique;
• generating, a ground truth SAR image comprising a set of coordinates from the received second raw radar data for each object amongst the plurality of objects using a Fast Fourier Transform (FFT) based RMA technique;
• and
• iteratively training, a Local Interpolation Function (LIF) neural network by updating an encoder weight and a decoder weight associated with an encoder and a decoder comprised in the LIF neural network, until a termination criteria is met, to obtain a trained LIF neural network, wherein the step of iteratively training comprises:
• generating an LIF feature map from the noisy SAR image for each object amongst the plurality of objects, using the encoder, wherein the LIF feature map comprises a latent representation for each coordinate of the set of coordinates;
• generating a predicted SAR image from the LIF feature map using the decoder by querying each coordinate of the set of coordinates;
• computing an L1 loss between the predicted SAR image and the ground truth SAR image;
• and
• updating the encoder weight and the decoder weight based on the L1 loss;
• wherein the termination criteria are one of (i) an L1 loss of a previous iteration is less than or equal to an L1 loss of a current iteration, or (ii) completion of a predefined number of iterations.