초록 ▼

A system and method for processing radar returns to identify returns from man-made objects. In one embodiment, a plurality of radar returns is used to form a corresponding plurality of resonance maps, which are combined using an ordered statistic to form a processed resonance map. A discriminant is

A system and method for processing radar returns to identify returns from man-made objects. In one embodiment, a plurality of radar returns is used to form a corresponding plurality of resonance maps, which are combined using an ordered statistic to form a processed resonance map. A discriminant is calculated as (a) the fourth power of the ratio of (i) the power in a first rectangle about each pixel to (ii) the power in a region outside the first rectangle and inside a second, larger rectangle, or (b) zero if the ratio is less than 1. Other processing operations including thresholding operations to suppress noise and clutter are used to improve the signal to noise ratio for detecting man-made objects.

대표청구항 ▼

1. A method for identifying a target of interest, the method comprising: generating, by a radar transmitter and a radar receiver, one or more digitized radar returns;forming, by a processing unit, a processed resonance map from the one or more digitized radar returns, the processed resonance map com

1. A method for identifying a target of interest, the method comprising: generating, by a radar transmitter and a radar receiver, one or more digitized radar returns;forming, by a processing unit, a processed resonance map from the one or more digitized radar returns, the processed resonance map comprising a plurality of pixels, each pixel having a value; andforming, by the processing unit, a signature discriminant array, each element of the signature discriminant array corresponding to a pixel of the plurality of pixels, each element of the signature discriminant array being: the fourth power of the ratio of: the power within an inner rectangle, tothe difference between the power within an outer rectangle and the power within the inner rectanglewhen the power within the inner rectangle exceeds the power within the outer rectangle, andzero otherwise;the power within the inner rectangle being the sum of the squares of the pixel values in a rectangle of a first height and a first width centered on the corresponding pixel of the plurality of pixels,the power within the outer rectangle being the sum of the squares of the pixel values in a rectangle of a second height and a second width centered on the corresponding pixel of the plurality of pixels,the second height and width being greater than the first height and width, respectively, andthe first height and width and the second height and width being selected to match a target of interest. 2. The method of claim 1, wherein the one or more digitized radar returns comprise a plurality of digitized radar returns, and the forming of the processed resonance map from the one or more digitized radar returns comprises: forming a plurality of raw resonance maps, each of the raw resonance maps being formed from one of the plurality of digitized radar returns by processing the radar return with a modified covariance spectral estimation super resolution function and taking the magnitude of the result, each of the raw resonance maps comprising a plurality of pixels, each pixel having a value; andsetting the value of each pixel of the processed resonance map to be equal to the smallest value of the corresponding set of pixels in the raw resonance maps. 3. The method of claim 1, further comprising: taking the mean of the signature discriminant array to form a discriminant mean;taking the maximum of the signature discriminant array to form a discriminant maximum;dividing the signature discriminant array by the discriminant mean to form a normalized discriminant array;multiplying the processed resonance map, pixel by pixel, by the normalized discriminant array to form a first modified resonance map;multiplying a first constant by the maximum value of the first modified resonance map to form a first threshold;setting to zero all values of the first modified resonance map that are less than the first threshold, to form a second modified resonance map;applying a median filter to the second modified resonance map to form a third modified resonance map;forming a threshold array by: setting to one all values of the signature discriminant array exceeding the product of: a second constant, andthe discriminant maximum, andsetting to zero the remaining values of the signature discriminant array; andmultiplying the third modified resonance map, pixel by pixel, by the threshold array to form a fourth modified resonance map. 4. The method of claim 3, wherein the forming of the processed resonance map comprises: receiving a plurality of digitized radar returns;forming a plurality of raw resonance maps, each of the raw resonance maps formed from one of the plurality of digitized radar returns by processing the digitized radar return with a modified covariance spectral estimation super resolution function and taking the magnitude of the result, each of the raw resonance maps comprising a plurality of pixels, each pixel having a value; andsetting the value of each pixel of the processed resonance map to be equal to the smallest value of the corresponding set of pixels in the raw resonance maps. 5. The method of claim 4, wherein the first constant is between 0.5 and 0.9 and the second constant is between ¼ and ½. 6. The method of claim 4, further comprising displaying the fourth modified resonance map on a display. 7. A radar system, comprising: a radar transmitter;a radar receiver; anda processing unit configured to: receive a plurality of digitized radar returns;form a processed resonance map comprising a plurality of pixels from the plurality of digitized radar returns, each pixel having a value;form a signature discriminant array, each element of the signature discriminant array corresponding to a pixel of the plurality of pixels, each element of the signature discriminant array being: the fourth power of the ratio of: the power within an inner rectangle, to the difference between the power within an outer rectangle and the power within the inner rectangle when the power within the inner rectangle exceeds the power within the outer rectangle, and zero otherwise;the power within the inner rectangle being the sum of the squares of the pixel values in a rectangle of a first height and a first width centered on the corresponding pixel of the plurality of pixels,the power within the outer rectangle being the sum of the squares of the pixel values in a rectangle of a second height and a second width centered on the corresponding pixel of the plurality of pixels,the second height and width being greater than the first height and width, respectively, andthe first height and width and the second height and width being selected to match a target of interest. 8. The system of claim 7, wherein the forming of the processed resonance map from the plurality of digitized radar returns comprises: forming a plurality of raw resonance maps, each of the raw resonance maps formed from one of the plurality of digitized radar returns by processing the digitized radar return with a modified covariance spectral estimation super resolution function and taking the magnitude of the result, each of the raw resonance maps comprising a plurality of pixels, each pixel having a value; andsetting the value of each pixel of the processed resonance map to be equal to the smallest value of the corresponding set of pixels in the raw resonance maps. 9. The system of claim 7, wherein the processing unit is further configured to: take the mean of the signature discriminant array to form a discriminant mean;take the maximum of the signature discriminant array to form a discriminant maximum;divide the signature discriminant array by the discriminant mean to form a normalized discriminant array;multiply the processed resonance map, pixel by pixel, by the normalized discriminant array to form a first modified resonance map;multiply a first constant by the maximum value of the first modified resonance map to form a first threshold;set to zero all values of the first modified resonance map that are less than the first threshold, to form a second modified resonance map;apply a median filter to the second modified resonance map to form a third modified resonance map;form a threshold array by: setting to one all values of the signature discriminant array exceeding the product of: a second constant, andthe discriminant maximum, andsetting to zero the remaining values of the signature discriminant array; andmultiply the third modified resonance map, pixel by pixel, by the threshold array to form a fourth modified resonance map. 10. The system of claim 9, wherein the forming of the processed resonance map from the plurality of digitized radar returns comprises: forming a plurality of raw resonance maps, each of the raw resonance maps formed from one of the plurality of digitized radar returns by processing the radar return with a modified covariance spectral estimation super resolution function and taking the magnitude of the result, each of the raw resonance maps comprising a plurality of pixels, each pixel having a value; andsetting the value of each pixel of the processed resonance map to be equal to the smallest value of the corresponding set of pixels in the raw resonance maps. 11. The system of claim 10, wherein the first constant is between 0.5 and 0.9 and the second constant is between ¼ and ½. 12. The system of claim 10, further comprising a display, wherein the processing unit is further configured to display the fourth modified resonance map on the display. 13. A method for identifying a target of interest, the method comprising: receiving a first radar return from a known sample target;forming a first processed resonance map from the first radar return, the first processed resonance map comprising a plurality of pixels, each pixel having a value;forming a plurality of first signature discriminant arrays from the first processed resonance map, for a plurality of sets of inner rectangle dimensions and a plurality of sets of outer rectangle dimensions;taking the maximum of each of the plurality of first signature discriminant arrays, to form a plurality of discriminant maximums;selecting a first set of inner rectangle dimensions and a first set of outer rectangle dimensions, the first set of inner rectangle dimensions and the first set of outer rectangle dimensions corresponding to the largest-valued discriminant maximum;receiving a second radar return from a scene;forming a second processed resonance map from the second radar return, the second processed resonance map comprising a plurality of pixels, each pixel having a value; andforming a second signature discriminant array from the second processed resonance map, for the first set of inner rectangle dimensions and the first set of outer rectangle dimensions. 14. The method of claim 13, wherein the forming of a signature discriminant array from an input resonance map, for a set of input inner rectangle dimensions and a set of input outer rectangle dimensions, the input resonance map comprising a plurality of input pixels, each of the input pixels having an input pixel value, comprises: forming the signature discriminant array as an array comprising a plurality of output pixels, each of the output pixels corresponding to an input pixel and having an output pixel value, each output pixel value being: the fourth power of the ratio of: the power within an inner rectangle, to the difference between the power within an outer rectangle and the power within the inner rectangle when the power within the inner rectangle exceeds the power within the outer rectangle, and zero otherwise;the power within the inner rectangle being the sum of the squares of the input pixel values in a rectangle having as dimensions the set of input inner rectangle dimensions centered on the corresponding input pixel,the power within the outer rectangle being the sum of the squares of the input pixel values in a rectangle having as dimensions the set of input outer rectangle dimensions centered on the corresponding input pixel. 15. The method of claim 13, further comprising displaying the second signature discriminant array on a display.