Two-dimensional image data reducer and classifier
First Claim
1. A method for obtaining target vehicle identification values representing a target outline configuration within a detector'"'"'s field of view, comprising the steps of:
- a. digitizing the target image into a two dimensional array of pixels;
b. summing target image pixels along one scan dimension to provide a first image vector b;
c. summing target image pixels along the other orthogonal dimension to provide a second image vector c;
d. shifting image vector b normal to the scan direction such that the maximum component is at a fixed reference point;
e. Fourier transforming the shifted image vector in step d to obtain F;
f. obtaining the real part of F by summing F and its conjugate F* and dividing the resultant vector by 2;
g. obtaining the imaginary part of F by subtracting F* from F and dividing the resultant vector by 2;
h. obtaining a numerical value representing the symmetric part of F in step e by integrating over the square of the real part of F;
i. obtaining a numerical value representing the asymmetric part of F by integrating over the product of the imaginary part of F and the conjugate of the imaginary part of F;
j. obtaining the identification value of image vector b by dividing the numerical value in step h by the sum of the values in steps h and i;
k. repeating steps d through j, using the second image vector c; and
l. in a look-up table predefined in memory, looking up the vehicle corresponding to the identification values obtained by following steps a through k.
1 Assignment
0 Petitions
Accused Products
Abstract
This invention relates to surveillance and counter-surveillance classificon of military vehicles using one-dimensional analysis of the target images. A two-dimensional image is digitized into a n by n pixel matrix which is summed along each row and column to produce a pair of n-component vectors which are invariant under image translation or rotation. A one-dimensional Fourier transform can be obtained from either or both of these two vectors. The image vectors are useful in the identification of U.S. and threat vehicle targets, or in the detection of target image motion.
51 Citations
1 Claim
-
1. A method for obtaining target vehicle identification values representing a target outline configuration within a detector'"'"'s field of view, comprising the steps of:
-
a. digitizing the target image into a two dimensional array of pixels; b. summing target image pixels along one scan dimension to provide a first image vector b; c. summing target image pixels along the other orthogonal dimension to provide a second image vector c; d. shifting image vector b normal to the scan direction such that the maximum component is at a fixed reference point; e. Fourier transforming the shifted image vector in step d to obtain F; f. obtaining the real part of F by summing F and its conjugate F* and dividing the resultant vector by 2; g. obtaining the imaginary part of F by subtracting F* from F and dividing the resultant vector by 2; h. obtaining a numerical value representing the symmetric part of F in step e by integrating over the square of the real part of F; i. obtaining a numerical value representing the asymmetric part of F by integrating over the product of the imaginary part of F and the conjugate of the imaginary part of F; j. obtaining the identification value of image vector b by dividing the numerical value in step h by the sum of the values in steps h and i; k. repeating steps d through j, using the second image vector c; and l. in a look-up table predefined in memory, looking up the vehicle corresponding to the identification values obtained by following steps a through k.
-
Specification