METHOD AND SYSTEM FOR FORMING VERY LOW NOISE IMAGERY USING PIXEL CLASSIFICATION

• US 20110012778A1
• Filed: 09/14/2010
• Published: 01/20/2011
• Est. Priority Date: 12/10/2008
• Status: Active Grant

• Current Assignee: The United States of America As Represented By The Secretary of Agriculture
• Original Assignee: United States Secretary of the Army
• Inventors: SICHINA, JEFFREY P., NGUYEN, LAM HUY
• Granted Patent: US 8,193,967 B2
• Time in Patent Office: Days
• US Class Current: 342/25.A
• CPC Class Codes: G01S 13/9017 with time domain processing of the SAR signals in azimuth G01S13/9005 takes precedence
• CPC Class Codes: G01S 13/9089 SAR having an irregular aperture
• CPC Class Codes: G01S 7/414 Discriminating targets with respect to background clutter

Abstract

A method and system for generating images from projection data comprising inputting from at least one data receiving element first values representing correlated positional and recorded data; each of said first values forming a point in an array of k data points; forming an image by processing the projection data utilizing a pixel characterization imaging subsystem that combines the positional and recorded data to form the SAR imagery utilizing one of a back-projection algorithm or range migration algorithm; integrating positional and recorded data from many aperture positions, comprising: forming the complete aperture A₀ for SAR image formation comprising collecting the return radar data, the coordinates of the receiver, and the coordinates of the transmitter for each position k along the aperture of N positions; forming an imaging grid comprising M image pixels wherein each pixel P_i in the imaging grid is located at coordinate (x_P(i),y_P(i), z_P(i)); selecting and removing a substantial number of aperture positions to form a sparse aperture A_i; repeating the selecting and removing step for L iterations for each A_i; classifying each pixel in the image into either target class based on the statistical distribution of its amplitude across L iterations (1≦i≦L); whereby if an image pixel is classified so as to be associated with a physical object, its value is computed from its statistics; otherwise, the pixel is assumed to come from a non-physical object and is given the value of zero.

Claims

1. A method for generating images from projection data comprising:

inputting from at least one data receiving element first values representing correlated positional and recorded data;

each of said first values forming a point in an array of N data points;

forming an image by processing the projection data utilizing a pixel characterization imaging subsystem that combines the positional and recorded data to form the SAR imagery utilizing one of a back-projection algorithm or range migration algorithm;

integrating positional and recorded data from many aperture positions, comprising;

forming the complete aperture A₀ for SAR image formation comprising collecting the return radar data, the coordinates of the receiver, and the coordinates of the transmitter for each position k along the aperture of N positions;

forming an imaging grid comprising M image pixels wherein each pixel P_i in the imaging grid is located at coordinate (x_P(i),y_P(i), z_P(i));

selecting and removing a substantial number of aperture positions to form a sparse aperture A_i;

repeating the selecting and removing step for L iterations for each A_i;

classifying each pixel in the image into either target class based on the statistical distribution of its amplitude across L iterations (1≦

i≦

L);

 

whereby if an image pixel is classified so as to be associated with a physical object, its value is computed from its statistics;

otherwise, the pixel is assumed to come from a non-physical object and is given the value of zero.

2. The method of claim 1 wherein the number of data points removed to form the sparse aperture A_i is in the range of approximately twenty to fifty percent.

3. The method of claim 1 wherein the step of selecting and removing aperture positions comprises arbitrarily removing approximately twenty to fifty percent of the data points from the array each time from substantially random locations within the array, such that each time the step is performed, substantially the same number of data points at substantially a different set of locations is removed.

4. The method of claim 1 wherein the step of selecting and removing aperture positions comprises arbitrarily removing a predetermined percentage of data points each time from random locations within the array, such that each time the step is performed, substantially the same number of data points at substantially a different set of locations is removed.

5. The method of claim 1, further comprising the step of processing return radar signals using a signal processing subsystem, the signal processing system comprising at least one signal processing algorithms performed to suppress unwanted noise from the returned radar data.

6. The method of claim 1 wherein the first point in the array represents the positional information and the recorded data obtained during the first increment;

the k^th data point in the array represents the combined positional information and recorded data obtained by the at least one data receiving element during the k^th increment;

where k is a number equal or greater than 1;

and wherein during the formation of the sparse apertures A_i approximately twenty to fifty percent of the data points are removed from the array randomly.

7. The method of claim 1 wherein the steps of selecting and removing a substantial number of aperture positions to form a sparse aperture Ai and repeating the randomly selecting and removing step for L iterations for each Ai comprises:

forming a preliminary bit-mapped image using the reduced array of data points;

forming a second array by removing approximately twenty to fifty percent of the data points from the array;

forming a second preliminary bit-mapped image using the second array of data points;

analyzing statistical distribution of each pixel across all iterations to classify the pixel into target class or noise class based upon the statistical analysis of the pixels;

whereby the values of each pixel are examined across all iterations and if the normalized fluctuation is greater than a threshold value, the pixel is assigned a value of zero and otherwise the pixel is assigned a value which is one of the maximum, average or mean value across the iterations.

9. The method of claim 7 wherein for each j^th pixel in the image grid, the decision statistic is the standard deviation of the amplitudes of the pixel P_ij across L iterations, and this standard deviation normalized by the mean value of the amplitudes

8. A method for generating images from projection data comprising:

a) inputting data from a scanned area having first values representing correlated positional and recorded data;

each of said first values forming a point in an array of k data points;

b) forming an aperture A₀ consisting of N elements, each element comprising radar receiving position information (x_R(k),y_R(k), z_R(k)) 1≦

k≦

N, the radar transmitting information (x_T(k),y_T(k), z_T(k)), and the data record s_k(t) that was measured at the location;

c) forming the imaging grid comprising M image pixels wherein each pixel P_i in the imaging grid is located at coordinate (x_P(i),y_P(i), z_P(i)) using a backprojection or range migration algorithm;

d) generating the backprojection value a j^th pixel computed by using the equation

10. The method of claim 8 wherein the step of generating a sparse aperture A_i having K positions from the complete aperture A₀ having N positions comprises removing twenty to fifty percent of the data points from the array to form a reduced array and wherein the same predetermined percentage of data points are removed each time with the selection of the data points subject to removal being computer generated.

11. The method of claim 10 wherein each time the step of removing approximately twenty to fifty percent of the data points from the array comprises removing a predetermined percentage of data points each time from random locations within the array, such that each time the step is performed, substantially the same number of data points at substantially a different set of locations is removed.

12. The method of claim 8 wherein the step of inputting data from a scanned area comprises scanning an area using at least one data receiving element to obtain first values representing combined positional and recorded data;

each of said first values forming a point in an array of k data points.

13. The method of claim 11 wherein the array of k data points represents radar data obtained by an array of k radar receivers, each row of data being derived from a separate radar receiver, and wherein columns may be formed in the array comprising data from a different time interval in which data is received by each of the k radar receivers.

14. The method of claim 8 wherein the steps (f) and (g) are repeated over a series of iterations until the desired image resolution is obtained.

15. A system for generating images from projection data comprising:

at least one processor for processing image information;

the at least one processor having an input for inputting scanned data having first values representing correlated positional and recorded data;

each of said first values forming a point in an array of N data points;

the at least one processor operating to perform the following steps;

(h) forming an aperture A₀ consisting of N elements, each element comprising radar receiving position information (x_R(k),y_R(k), z_R(k)) 1≦

k≦

N, radar transmitting information (x_T(k),y_T(k), z_T(k)), and the data record s_k(t) that was measured at the location;

(i) forming the imaging grid comprising M image pixels wherein each pixel P_i in the imaging grid is located at coordinate (x_P(i),y_P(i), z_P(i)) using one of backprojection, range migration algorithm, or polar format;

(j) generating the value a j^th pixel computed by

P_0j=F(w_0k,k,k,j) where 1≦

k≦

N and 1≦

j≦

M

and the baseline image I₀=P_0j;

(k) assigning the value of weighting factors w_0k to be 1 for A₀, w_0k defining which aperture positions contribute to the formed image or do not contribute;

(l) generating a sparse aperture A_i having K positions from the complete aperture A₀ having N positions where A_i, 1≦

i≦

L;

where L is the number of iterations, using the equation

P_ij=F(w_ik,k,j) where 1≦

k≦

N and 1≦

j≦

M

to form the image from the sparse apertures A_i, and where the value of w_ik is either 0 or 1 to define which aperture positions contribute to the formed image, and where there are K elements of w_ik having the value of 1, and (N−

K) elements of w_ik having the value of 0;

which (N−

K) and K locations inside the array A_i of are randomized for each iteration;

(m) forming the image I_i using data from the sparse aperture A_i where the backprojection value of the jth pixel using the sparse aperture A is found by computing the magnitude E_i using

E_i=|Hilbert(I_i)|,

where I_i is defined as I_i=P_ij;

(n) repeating the steps of (e) and (f) for L iterations;

whereby the value of each pixel is examined across L iterations to make a decision and classify whether or not the pixel belongs to a physical object to thereby remove unwanted noise in the generation of an image using electromagnetic signals.

16. The system of claim 15 further comprising forming an image by processing the inputted data utilizing a pixel characterization imaging subsystem that combines the positional and recorded data to form SAR imagery utilizing one of a back-projection algorithm or range migration algorithm;

17. The system of claim 15 wherein the data is inputted from a plurality of data receiving elements that scan substantially the same area and three dimensional images are obtained using signal data from each receiving element that includes the magnitude of the signal, the angle from which the signal was derived, the time of receipt, and the positional information of the receiving element.

18. The system of claim 17 wherein a target can be located by inputting data from two receiving elements, determining the respective angles from signals reflected from the target obtained by both receiving elements, and determining the target location based upon the angles at which the signals are reflected, the time at which the signal is received, and the positional information of the receiving elements.

19. The system of claim 17 wherein the step of generating a sparse aperture Ai comprises using the equation:

20. The system of claim 17 wherein the step of generating the value a j^th pixel is computed by using the equation