JP2511939B2 - Reconstruction method of synthetic aperture radar image - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial field of application] The present invention relates to a synthetic aperture radar (Synthetic Aperture Radar, hereinafter referred to as "SAR") mounted on an artificial satellite or an aircraft.
The present invention relates to a digital processing method for reproducing a human-understandable image from image pickup data according to (1), and particularly to a processing method suitable for reproducing a high quality image at high speed.

[Conventional technology]

In the field of remote sensing using satellites or airplanes, a SAR that can obtain high-resolution images in the microwave band transmitting through clouds as a sensor for imaging the surface of the earth
Is attracting attention.

Figure 4 shows the overall system of SAR. The SAR including the radar sensor 401 and the antenna 402 is mounted on an artificial satellite or the like, and moves along the flight path 403 in the direction of arrow 404 to image the ground surface. The imaging data from the SAR is received by the ground station 405 and processed by the data processor 406 to create the image film 407, the data storage magnetic tape 408, and the like. In addition, 409 is a resolution cell, 410 is the range direction on the ground surface of the data sampled by SAR, 411 is the same azimuth direction, 412 is the antenna beam, and 413 is the cutting width.

The outline of the processing of the data collected by SAR is described below. For more details, refer to Proceedings of 13th International Symposium on R, Preliminary Report of 13th International Symposium on Remote Sensing of Environment, p.337-360, April 1977.
emote Sensing of Environmert, p.337-360, April, 197
Bennette and Cumm in 7)
ing) “Digital Synthetic Aperture Radar Image Creation, Airborne and Satellite Results” (“Digiat
l SAR Image Formation, Airborne and Satellite Resul
ts ").

In the SAR received image, one point on the original image is distributed with the spread of the point image pattern h (t, r) and cannot be used as it is. Here, r indicates the range direction, and t indicates the azimuth direction. The information spread in the received image is first compressed in the range direction,
Next, it is compressed in the azimuth direction. The range compression processing is performed by the correlation processing with the point image pattern data for each line of the image data. However, if the correlation process is executed as it is, it will take an enormous amount of processing time. Therefore, fast Fourier transform (hereinafter referred to as “FFT”), complex multiplication,
High-speed inverse Fourier transform (hereinafter referred to as "IFFT") is used to achieve speedup. In order to perform correlation processing using FFT, first, a point image pattern is generated by digital processing by a computer, and FFTs of both the point image pattern and one line of image data are calculated. Since the correlation of two data is a mere multiplication in the frequency domain after calculating the FFT, the product of the FFT calculation results of the above two data is taken and IFFT is applied to The relation result is obtained.

Similar to the range compression, the azimuth compression also uses FFT to obtain the correlation result at high speed.

However, even if the speedup using the FFT is performed as described above, the amount of calculation required for the SAR image reproduction processing is still enormous, and an ultra-high-speed arithmetic unit is required to perform the processing within a practical time. Is known to be.

On the other hand, in order to obtain a high quality SAR image, it is necessary to accurately estimate the point image pattern. In particular, the point image pattern in the azimuth direction is different for each scene because it depends on the moving speed of the SAR sensor and the distance to the ground surface.
The azimuth direction point image pattern can be calculated from the position / velocity data of the SAR sensor obtained by other means, but these data include errors, and usually the point image pattern is estimated automatically from the imaging data itself. High-quality SAR images are often obtained by the focus method.

The azimuth direction point image pattern h (t) is Is represented. Here, t is a coordinate in the azimuth direction, k and β are quantities called reproduction parameters, k is a Doppler frequency change rate, and β is a Doppler center frequency. A (t) is called an antenna pattern and is related to the intensity distribution of microwaves emitted from the SAR sensor in the azimuth direction. Although k and β are estimated by completely different methods, especially regarding the estimation of the Doppler center frequency β,
Transaction on Geo Science and
Remote Sensing GE 23, Number 1 (198
5 years) Pages 47 to 56 (IEEE Transactions on Geosci
F.Li et al. in ence and Rensing, GE-23, No.1 (1985) pp.47-56) “Doppler parameter estimation in satellite-borne synthetic aperture radar (“ Doppler Para
meter Estimation for Spaceborne Symthetic-Apertur
e Radars "). The method is to first reconstruct the image using the roughly estimated reconstruction parameters, then Fourier transform the image data again in the azimuth direction to obtain the power spectrum. , Estimate the Doppler center frequency from its shape.

According to this document, the center frequency can be estimated more accurately than the method in which the range-compressed data is Fourier-transformed in the azimuth direction and the Doppler center frequency is obtained from the power spectrum.

[Problems to be Solved by the Invention] However, according to the method of the above-mentioned document, after the point image pattern generation of the azimuth method, its FFT, multiplication with the received data, and the resulting IFFT, the FF is further processed.
It is necessary to calculate the power spectrum by performing T, and it is necessary to estimate the Doppler center frequency by approx.
It has a drawback that it requires a large amount of calculation and processing time that reaches 0% or more.

An object of the present invention is to provide a SAR image reproduction processing method that improves the above-mentioned drawbacks of the prior art, estimates the Doppler center frequency with high accuracy, and enables high-speed SAR image reproduction. .

[Means for solving problems]

The above-mentioned object is a power obtained as a square of the absolute value of the complex number data by operating a digital filter on the complex number data after the azimuth point image pattern multiplication in the azimuth direction frequency domain in the normal SAR image reproduction processing. It is achieved by estimating the Doppler center frequency using the spectrum.

[Work]

Filtering (convolution) in the frequency domain is equivalent to multiplication in the time domain. Therefore, by selecting an appropriate filter, the estimation accuracy similar to the removal of the detour effect by the conventional method proposed by Lee et al. Is realized. be able to. Furthermore, since the above effect is achieved by filtering in the frequency domain, it is possible to perform high-speed and accurate estimation while suppressing an increase in the amount of calculation required for center frequency estimation to a minimum.

〔Example〕

An embodiment of the present invention will be described below with reference to the flowchart of FIG.

In FIG. 1, the SAR reception data is first compressed in the range direction by the range compression processing 101. Next, the reproduction parameter calculation processing 102 calculates the initial value of the reproduction parameter and stores it in the reproduction parameter file 103. Next, the range-compressed received data is FFT processed in the azimuth direction 1
After being converted to the frequency domain by 04, the range curvature correction processing 105 corrects the range curvature distortion.

In the azimuth point image pattern generation processing 106, a point image pattern is generated by referring to the reproduction parameter file 103, and the FFT
After the conversion into the frequency domain in the process 107, the complex multiplication 108 with the range curvature corrected data is performed. When control is first passed to the branch processing 113 for a certain scene, filtering processing for removing the wraparound effect is performed.
Proceeding to 109, the contents of the reproduction parameter file 103 are updated according to the result of the Doppler center frequency estimation processing 110. When the processes 106 to 108 are repeated using the updated contents of the reproduction parameter file and the second branch process 113 is reached, the process proceeds to the IFFT process 111, and the final SAR image is processed by the multi-look addition / quantization process 112. Is played.

The reproduced result is stored on a magnetic tape or the like and applied to the user.

When the filter used in the filtering process 109 is, for example, a three-point filter G ₋₁ = 0.25, G ₀ = 0.5, G ₁ = 0.25, the data after complex multiplication is F _k , k = −Nd / 2, ..., Nd / 2-1 (Nd: number of processed data samples) Will be processed. Of course, this processing can be easily realized by software as it is, and can also be realized by dedicated hardware. FIG. 2 shows how the wraparound effect is removed in this case. It can be seen that the effect of the wraparound area 201 is controlled by filtering. Of course, if the length of the filter is increased, the amount of calculation increases, but it goes without saying that the wraparound effect can be reduced more effectively.

FIG. 3 shows the principle of the Doppler center frequency estimation processing 110. The horizontal axis 301 is the azimuth direction frequency, the vertical axis 302 is the power spectrum intensity, and the curve 303 is the power spectrum of the received data after filtering. Usually, in order to reduce noise and improve estimation accuracy, the power spectra of a plurality of lines are averaged. The initial value of Doppler center frequency 304 is the first
It is the value calculated in the process 102 in the figure, and this value 304 is set to the curve 30.
The value is corrected to a value 305 in which the left and right areas surrounded by 3 and the horizontal axis 302 are equal, and this is set as the estimated value of the Doppler center frequency. Here, the left and right areas are estimated as equal values, but of course the method of directly obtaining the peak P of the curve 303 does not change the effect of the present invention.

〔The invention's effect〕

According to the present invention, since the wraparound effect is removed by the filtering process in the azimuth direction frequency domain, there is an effect that the Doppler center frequency can be estimated at high speed and with high accuracy, and a high quality SAR reproduced image can be obtained.

[Brief description of drawings]

FIG. 1 is a flow chart showing a processing procedure of an embodiment of the present invention, FIG. 2 is a view showing an example of removal of a wraparound effect in the processing 109 of FIG. 1, and FIG. FIG. 4 is a block diagram showing the entire SAR system.

 ─────────────────────────────────────────────────── ─── Continued Front Page (56) References JP-A-62-73182 (JP, A) JP-A-61-187079 (JP, A) IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, GE-23 [1] (1985) ) P47-56 Dig Int Geosci Remote Sens Symp 1982 [1] (1982) P. wp3.7.1-wp3.7.5 ICASSP 86, TOKYO. 1986 [1] (1986) P36.1.1-36.1.4

Claims (1)

(57) [Claims] 1. A synthetic aperture radar image reproduction processing method for estimating a Doppler center frequency from an azimuth direction power spectrum of image data obtained by a synthetic aperture radar, wherein a digital filter is applied to data after azimuth point image pattern multiplication in a frequency domain. A synthetic aperture radar image reproduction processing method, characterized in that a power spectrum as a result of application is obtained, and a Doppler center frequency is estimated from the shape of the power spectrum.