CN104656067A - Frequency modulation slop estimation method based on picture contrast - Google Patents

Abstract

The invention belongs to an estimation method and particularly relates to a frequency modulation slope estimation method based on picture contrast. The method comprises the following steps: step 1: pretreatment: calculating initial frequency modulation slopes Ka1 and Ka2; step 2, calculating frequency modulation slope, obtaining a frequency modulation slope Ka3 by calculating through a golden cut method, using the frequency modulation slope to carry out distance direction compression and orientation direction compression to original SAR data, so as to obtain a two-dimensional image, and simultaneously calculating the contrast grade C3 of the two-dimensional image; step 3, curve fitting: carrying out secondary fitting of a parabola by using three corresponding points (Ka1, C1), (Ka2, C2) and (Ka3, C3), so as to obtain a matched curve; step 4, slop calculation; step 5, imaging. The estimation method has the following effects: according to the estimation method, a method of parabolas is used to evaluate Doppler frequency modulation slope according to the relation between the picture contrast and the orientation direction frequency modulation slope, repeated iteration during a traditional algorithm is avoided, and the computational efficiency is improved greatly.

Description

A kind of chirp rate method of estimation based on picture contrast

Technical field

The invention belongs to method of estimation, be specifically related to a kind of chirp rate method of estimation based on picture contrast.

Background technology

Synthetic-aperture radar (Synthetic Aperture Radar, SAR) is a kind of high-resolution imaging radar.In actual applications, due to the impact of atmospheric turbulence, carrier of radar exists along heading and the disturbance perpendicular to heading, thus produces kinematic error.The doppler phase error that kinematic error is brought can destroy the coherence of radar echo signal, causing the matching treatment of Azimuth Compression in imaging processing to occur mismatch, causing the orientation of image to defocusing, even cannot imaging time serious.Therefore, in synthetic aperture radar image-forming and application thereof, need to estimate chirp rate with compensation of phase error.According to the introduction of domestic and international pertinent literature, conventional chirp rate method of estimation is picture contrast optimal algorithm.Picture contrast can be regarded as the departure degree of each picture element brightness value and brightness of image average in image, by the concept of variance in theory of probability, picture contrast function is the standard deviation of orientation to brightness of image and the ratio of mean value, and the estimation of chirp rate can be equivalent to the optimizing process that solves contrast value optimum solution.And in this optimizing process, along with sharply strengthening the raising calculated amount of error compensation accuracy requirement, algorithm need iterate, cause that algorithm calculated amount is large, speed of convergence is slow, efficiency is low.

Summary of the invention

This object is for prior art defect, provides a kind of chirp rate method of estimation based on picture contrast.

The present invention is achieved in that a kind of chirp rate method of estimation based on picture contrast, it is characterized in that, comprises the steps:

Step one: pre-service

Initial chirp rate is calculated according to radar inertial guidance data wherein, V represents the speed of radar carrier, and R represents radar range-to-go, λ represents radar wavelength, and utilize these two chirp rates to carry out distance respectively to compression and orientation to compression to original SAR data, obtain two dimensional image, calculate the contrast of gained two dimensional image simultaneously respectively $C1=\frac{\sqrt{A\left\{{[I^2(n,\mathrm{Ka}1)-A[I^2(n,\mathrm{Ka}1)\left]\right]}^2\right\}}}{A\left[I^2(n,\mathrm{Ka}1)\right]},C2=\frac{\sqrt{A\left\{{[I^2(n,\mathrm{Ka}2)-A[I^2(n,\mathrm{Ka}2)\left]\right]}^2\right\}}}{A\left[I^2(n,\mathrm{Ka}2)\right]}$ Be wherein I (n, Ka), and (n=1 ..., N) orientation to the amplitude of one dimensional image, A() representation space is averaging computing;

Step 2: calculate chirp rate

Calculate chirp rate Ka3=0.618 (Ka2-Ka1)+Ka1 by Fibonacci method, and utilize this chirp rate to carry out distance to compression and orientation to compression to original SAR data, obtain two dimensional image, calculate the contrast of gained two dimensional image simultaneously $C3=\frac{\sqrt{A\left\{{[I^2(n,\mathrm{Ka}3)-A[I^2(n,\mathrm{Ka}3)\left]\right]}^2\right\}}}{A\left[I^2(n,\mathrm{Ka}3)\right]};$

Step 3: curve

Utilize above-mentioned three corresponding point (Ka1, C1), (Ka2, the C2) and (Ka3, C3) calculated to carry out second-degree parabola matching, obtain matched curve,

Step 4: slope calculations

The maximal value of the quafric curve that digital simulation obtains and corresponding chirp rate value Ka0, this chirp rate value is exactly the estimated value of chirp rate,

Step 5: imaging

Utilize Ka0 to SAR raw data carry out distance to compression and orientation to compression, obtain diameter radar image.

As above based on a chirp rate method of estimation for picture contrast, wherein, to compression, described distance refers to that namely distance is to Fourier transform and inverse transformation, to compression, orientation refers to that orientation is to Fourier transform and inverse transformation.

As above based on a chirp rate method of estimation for picture contrast, wherein, the curve of step 3 adopts formula as follows to carry out:

$C=\frac{(\mathrm{Ka}-{\mathrm{Ka}}_2)(\mathrm{Ka}-{\mathrm{Ka}}_3)}{({\mathrm{Ka}}_1-{\mathrm{Ka}}_2)({\mathrm{Ka}}_1-{\mathrm{Ka}}_3)}{C}_1\text{+}\frac{(\mathrm{Ka}-{\mathrm{Ka}}_1)(\mathrm{Ka}-{\mathrm{Ka}}_3)}{({\mathrm{Ka}}_2-{\mathrm{Ka}}_1)({\mathrm{Ka}}_2-{\mathrm{Ka}}_3)}{C}_2+\frac{(\mathrm{Ka}-{\mathrm{Ka}}_1)(\mathrm{Ka}-{\mathrm{Ka}}_2)}{({\mathrm{Ka}}_3-{\mathrm{Ka}}_1)({\mathrm{Ka}}_3-{\mathrm{Ka}}_2)}{C}_3$

Wherein Ka is independent variable, and C is dependent variable.

As above based on a chirp rate method of estimation for picture contrast, wherein, Ka0 adopts following formula to calculate:

${\mathrm{Ka}}_0=\frac{1}{2}\frac{C_1({\mathrm{Ka}}_2^2-{\mathrm{Ka}}_3^2)+C_2({\mathrm{Ka}}_3^2-{\mathrm{Ka}}_1^2)+C_3({\mathrm{Ka}}_1^2-{\mathrm{Ka}}_2^2)}{C_1({\mathrm{Ka}}_2-{\mathrm{Ka}}_3)+C_2({\mathrm{Ka}}_3-{\mathrm{Ka}}_1)+C_3({\mathrm{Ka}}_1-{\mathrm{Ka}}_2)}.$

Use effect of the present invention to be: the method be according to picture contrast and orientation to the relation of chirp rate, use parabolic method to carry out estimating Doppler chirp rate, avoid iterating in traditional algorithm, greatly can improve counting yield.

Embodiment

Based on a chirp rate method of estimation for picture contrast, comprise the steps:

Step one: pre-service

Initial chirp rate is calculated according to radar inertial guidance data wherein, V represents the speed of radar carrier, and R represents radar range-to-go, and λ represents radar wavelength.And utilize these two chirp rates to original SAR data carry out respectively distance to compression (namely distance is to Fourier transform and inverse transformation) and orientation to compress (namely orientation is to Fourier transform and inverse transformation), obtain two dimensional image, calculate the contrast of gained two dimensional image simultaneously respectively $C1=\frac{\sqrt{A\left\{{[I^2(n,\mathrm{Ka}1)-A[I^2(n,\mathrm{Ka}1)\left]\right]}^2\right\}}}{A\left[I^2(n,\mathrm{Ka}1)\right]},C2=\frac{\sqrt{A\left\{{[I^2(n,\mathrm{Ka}2)-A[I^2(n,\mathrm{Ka}2)\left]\right]}^2\right\}}}{A\left[I^2(n,\mathrm{Ka}2)\right]}$ Be wherein I (n, Ka), and (n=1 ..., N) orientation to the amplitude of one dimensional image, A() representation space is averaging computing;

Step 2: calculate chirp rate

Chirp rate Ka3=0.618 (Ka2-Ka1)+Ka1 is calculated by Fibonacci method, and utilize this chirp rate to original SAR data carry out distance to compression (namely distance is to Fourier transform and inverse transformation) and orientation to compress (namely orientation is to Fourier transform and inverse transformation), obtain two dimensional image, calculate the contrast of gained two dimensional image simultaneously $C3=\frac{\sqrt{A\left\{{[I^2(n,\mathrm{Ka}3)-A[I^2(n,\mathrm{Ka}3)\left]\right]}^2\right\}}}{A\left[I^2(n,\mathrm{Ka}3)\right]};$

Step 3: curve

Utilize above-mentioned three corresponding point (Ka1, C1), (Ka2, the C2) and (Ka3, C3) calculated to carry out second-degree parabola matching, obtain matched curve.

The curve of this step can adopt any approximating method of the prior art to carry out, and formula as follows also can be adopted to carry out:

$C=\frac{(\mathrm{Ka}-{\mathrm{Ka}}_2)(\mathrm{Ka}-{\mathrm{Ka}}_3)}{({\mathrm{Ka}}_1-{\mathrm{Ka}}_2)({\mathrm{Ka}}_1-{\mathrm{Ka}}_3)}{C}_1\text{+}\frac{(\mathrm{Ka}-{\mathrm{Ka}}_1)(\mathrm{Ka}-{\mathrm{Ka}}_3)}{({\mathrm{Ka}}_2-{\mathrm{Ka}}_1)({\mathrm{Ka}}_2-{\mathrm{Ka}}_3)}{C}_2+\frac{(\mathrm{Ka}-{\mathrm{Ka}}_1)(\mathrm{Ka}-{\mathrm{Ka}}_2)}{({\mathrm{Ka}}_3-{\mathrm{Ka}}_1)({\mathrm{Ka}}_3-{\mathrm{Ka}}_2)}{C}_3$

Wherein Ka is independent variable, and C is dependent variable.

Step 4: slope calculations

The maximal value of the quafric curve that digital simulation obtains and corresponding chirp rate value Ka0, this chirp rate value is exactly the estimated value of chirp rate.This step adopts existing techniques in realizing.

When the formula that step 4 uses the application to provide carries out curve fitting time, ${\mathrm{Ka}}_0=\frac{1}{2}\frac{C_1({\mathrm{Ka}}_2^2-{\mathrm{Ka}}_3^2)+C_2({\mathrm{Ka}}_3^2-{\mathrm{Ka}}_1^2)+C_3({\mathrm{Ka}}_1^2-{\mathrm{Ka}}_2^2)}{C_1({\mathrm{Ka}}_2-{\mathrm{Ka}}_3)+C_2({\mathrm{Ka}}_3-{\mathrm{Ka}}_1)+C_3({\mathrm{Ka}}_1-{\mathrm{Ka}}_2)}.$

Step 5: imaging

Utilize Ka0 to SAR raw data carry out distance to compression (namely distance is to Fourier transform and inverse transformation) and orientation to compress (namely orientation is to Fourier transform and inverse transformation), obtain diameter radar image.

Because in method, the value of three exploration chirp rates affects the accuracy of estimation of parabolic method.In practical application, the scope of the kinematic error data guestimate chirp rate that can record according to carried SAR geometric relationship and various motion sensor (as inertial navigation, Inertial Measurement Unit etc.).Near the chirp rate scope that this is rough, get Ka1, Ka2, Ka3 get Ka1, the golden section value of Ka2: Ka3=0.618 (Ka2-Ka1)+Ka1.

Claims (4)

1., based on a chirp rate method of estimation for picture contrast, it is characterized in that, comprise the steps:

Step one: pre-service

Initial chirp rate is calculated according to radar inertial guidance data wherein, V represents the speed of radar carrier, and R represents radar range-to-go, λ represents radar wavelength, and utilize these two chirp rates to carry out distance respectively to compression and orientation to compression to original SAR data, obtain two dimensional image, calculate the contrast of gained two dimensional image simultaneously respectively $C1=\frac{\sqrt{A\left\{{[I^2(n,\mathrm{Ka}1)-A[I^2(n,\mathrm{Ka}1)\left]\right]}^2\right\}}}{A\left[I^2(n,\mathrm{Ka}1)\right]},$ $C2=\frac{\sqrt{A\left\{{[I^2(n,\mathrm{Ka}2)-A[I^2(n,\mathrm{Ka}2)\left]\right]}^2\right\}}}{A\left[I^2(n,\mathrm{Ka}2)\right]},$ Be wherein I (n, Ka), and (n=1 ..., N) orientation to the amplitude of one dimensional image, A() representation space is averaging computing;

Step 2: calculate chirp rate

Calculate chirp rate Ka3=0.618 (Ka2-Ka1)+Ka1 by Fibonacci method, and utilize this chirp rate to carry out distance to compression and orientation to compression to original SAR data, obtain two dimensional image, calculate the contrast of gained two dimensional image simultaneously $C3=\frac{\sqrt{A\left\{{[I^2(n,\mathrm{Ka}3)-A[I^2(n,\mathrm{Ka}3)\left]\right]}^2\right\}}}{A\left[I^2(n,\mathrm{Ka}3)\right]};$

Step 3: curve

Utilize above-mentioned three corresponding point (Ka1, C1), (Ka2, the C2) and (Ka3, C3) calculated to carry out second-degree parabola matching, obtain matched curve,

Step 4: slope calculations

The maximal value of the quafric curve that digital simulation obtains and corresponding chirp rate value Ka0, this chirp rate value is exactly the estimated value of chirp rate,

Step 5: imaging

Utilize Ka0 to SAR raw data carry out distance to compression and orientation to compression, obtain diameter radar image.

2. a kind of chirp rate method of estimation based on picture contrast as claimed in claim 1, is characterized in that: to compression, described distance refers to that namely distance is to Fourier transform and inverse transformation, to compression, orientation refers to that orientation is to Fourier transform and inverse transformation.

3. a kind of chirp rate method of estimation based on picture contrast as claimed in claim 2, is characterized in that: the curve of step 3 adopts formula as follows to carry out:

$C=\frac{(\mathrm{Ka}-{\mathrm{Ka}}_2)(\mathrm{Ka}-{\mathrm{Ka}}_3)}{({\mathrm{Ka}}_1-{\mathrm{Ka}}_2)({\mathrm{Ka}}_1-{\mathrm{Ka}}_3)}{C}_1\text{+}\frac{(\mathrm{Ka}-{\mathrm{Ka}}_1)(\mathrm{Ka}-{\mathrm{Ka}}_3)}{({\mathrm{Ka}}_2-{\mathrm{Ka}}_1)({\mathrm{Ka}}_2-{\mathrm{Ka}}_3)}{C}_2+\frac{(\mathrm{Ka}-{\mathrm{Ka}}_1)(\mathrm{Ka}-{\mathrm{Ka}}_2)}{({\mathrm{Ka}}_3-{\mathrm{Ka}}_1)({\mathrm{Ka}}_3-{\mathrm{Ka}}_2)}{C}_3$

Wherein Ka is independent variable, and C is dependent variable.

4. a kind of chirp rate method of estimation based on picture contrast as claimed in claim 3, is characterized in that: Ka0 adopts following formula to calculate:

${\mathrm{Ka}}_0=\frac{1}{2}\frac{C_1({\mathrm{Ka}}_2^2-{\mathrm{Ka}}_3^2)+C_2({\mathrm{Ka}}_3^2-{\mathrm{Ka}}_1^2)+C_3({\mathrm{Ka}}_1^2-{\mathrm{Ka}}_2^2)}{C_1({\mathrm{Ka}}_2-{\mathrm{Ka}}_3)+C_2({\mathrm{Ka}}_3-{\mathrm{Ka}}_1)+C_3({\mathrm{Ka}}_1-{\mathrm{Ka}}_2)}.$