Method and system for autofocus radar imaging
AssigneeMitsubishi Electric Research Laboratories, Inc.
InventorDehong Liu
Life of this patent
Estimating the term…
Abstract
An image of a region of interest (ROI) is generated by a radar system including a set of one or more antennas. The radar system has unknown position perturbations. Pulses are transmitted, as a source signal, to the ROI using the set of antennas at different positions and echoes are received, as a reflected signal, by the set of antennas at the different positions. The reflected signal is deconvolved with the source signal to produce deconvolved data. The deconvolved data are compensated according a coherence between the reflected signal to produce compensated data. Then, a procedure is applied to the compensated data to produce reconstructed data, which are used to reconstruct auto focused images.
Claims9
A method for generating an image of a region of interest (ROI) using a radar system, wherein the radar system includes a set of one or more antennas, and the radar system has unknown position perturbations, comprising steps of:
- transmitting pulses, as a source signal, to the ROI using the set of antennas at different positions;
- receiving echoes, as reflected signals, by the set of antennas at the different positions;
- and further comprising computer implemented steps of:
- deconvolving the reflected signals with the source signal to produce deconvolved data;
- compensating phase changes of the deconvolved data according a coherence between the reflected signals to produce compensated data;
- generating reconstructed data from the compensated data;
- and
- reconstructing images using the reconstructed data, wherein the images are autofocused, wherein the deconvolved data are initialized as a residual signal, wherein the images include pixels corresponding to location in the ROI, wherein the computer implemented steps further comprise, until a termination condition is reached, iterative steps of:
- estimating the target location, as non-zero support across the images, each of which corresponding to one of the antennas;
- updating projection matrices based on coherence analysis and group sparsity;
- reconstructing the images with the updated projection matrices;
- and
- determining the residual signal for a next iteration by removing the target signal from the collected data.
Dependencies are inferred from the claim text.