JPS6368980A - Image reproduction processing system for synthetic aperture radar - Google Patents

Abstract

PURPOSE:To attain reproduction high in picture quality by making one- dimensional sampling of reproduced complex image in azimuthal direction, correcting phase deviation in range direction multiplying a phase rotation factor and making one-dimensional resampling in range direction. CONSTITUTION:A synthetic aperture radar SAR provided with a radar sensor 1 and an antenna 2 is mounted on an artificial satellite etc., and photographs ground surface while moving on the flying path 3 in the direction of arrow mark 4. Photographing data from the radar SAR are received by a ground station 6. In such a case, phase deviation in azimuthal direction is corrected by multiplying complex image after azimuth compression by phase rotation factor exp(-2pijBt). Then, one-dimensional resampling is made in azimuthal direction. Image data resampled in azimuthal direction is multiplied by phase rotation factor exp(2pijBt) again, and phase deviation in range direction is corrected. Finally, one-dimensional resampling is made in range direction to complete two-dimensional resampling process, and a reproduced image is obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention is directed to a synthetic aperture radar (Synthetic Aperture R) mounted on an artificial satellite or an aircraft.
adar.

This relates to a digital processing system for reproducing human-understandable images from imaging data obtained by rSARJ (hereinafter referred to as rSARJ), and is particularly suitable for reproducing images with high image quality.
This invention relates to an R image reproduction processing method.

[Conventional technology]

Regarding the method of reproducing SAR images through digital processing by a computer, see, for example, ``6 by Bennett and Cumming in Proceedings of the 13th International Symposium on Environmental Remote Sensing, pp. 337-360 (1977).
``Digital SAR Image Generation: Airborne and Satellite Results'' (Bennette).

CCumm1n: ”Digital SARIma
ge Formation: Ajrbnne and
5atellite Re5ults”, 13t
hProceedings of International
onal Symposium onRemote S
Ensuring of Environment, pp.
337-360 (1977)) J. Also, in the above literature, SA
A method of enlarging the image by performing resampling on the complex image after azimuth compression as a method of improving the interval resolution of the R reproduced image and reproducing a high-quality image. Problem] Since the above-mentioned conventional technology does not take into account the phase shift of the complex image after azimuth compression, the above-mentioned sampling causes edge deterioration such as an increase in high-frequency noise components and the generation of striped artifacts. This will be explained below.

In the SAR image after range compression, information on each point on the earth's surface is distributed one-dimensionally in the azimuth direction. The function representing this spread is called the azimuth direction point spread pattern h(t) = exp[2πj (αt''+βt)]
- It can be written as (1). Here, t represents the coordinate in the azimuth direction of the SAR sensor, j is an imaginary unit, α and β are quantities called reproduction parameters, and α is the Doppler change rate β is called the Doppler center frequency. Since α and β depend on the distance and relative speed between the satellite and the cold target point, they change slowly in the range direction. Therefore, the point spread pattern in the azimuth direction is 1
Not constant throughout the scene.

In an actual SAR image, the azimuthal point image pattern in the image after range compression is subject to geometric distortion called range curvature distortion, which is caused by changes in the relative distance between the SAR sensor and each point on the ground surface. , is distributed not on a straight line in the azimuth direction but on a curve. In the following, it is assumed that this distortion has already been corrected by range coverage correction.

The azimuth direction point spread pattern (1) is distributed with a certain width around the Doppler center frequency β in the frequency domain.

Therefore, the data resulting from azimuth compression by correlating the SAR image after range compression with the azimuth direction point image pattern is also distributed around the Doppler center frequency β in the frequency domain. In other words, in a complex image after azimuth compression, there is a phase shift between adjacent pixels in the azimuth direction, and conventional methods that perform interpolation such as cubic convolution on this complex image prevent deterioration of spatial resolution. Contrary to the purpose, this will cause deterioration of image quality.

On the other hand, a method of correcting the above-mentioned phase shift by parallel movement in the frequency domain or multiplication by a phase rotation factor in the real time domain is also conceivable. This method is equivalent to multiplying each line by a phase rotation factor of exp (-2πjβt) in azimuth compression, but since β is the amount that changes in the range direction, a different phase rotation factor is applied to each line. As a result of the multiplication, a phase shift occurs between pixels adjacent in the range direction.

As mentioned above, it is inevitable that a complex image after azimuth compression will have a phase shift in either the azimuth or range direction, and if sampling is performed without taking this phase shift into consideration, the image quality will deteriorate as described above. There was a problem that deterioration occurred.

The present invention solves the above problems and provides SAR of both image quality.
The objective is to provide a processing method for reproducing images.

[Means for solving problems]

The above purpose is to perform two-dimensional resampling on the SAR complex image after azimuth compression by dividing it into one-dimensional resampling in the range direction and one-dimensional resampling in the azimuth direction, and perform one-dimensional resampling in either the range direction or the azimuth direction. This is achieved by multiplying the complex image by a phase rotation factor to correct the phase shift.

[Effect]

Generally, two-dimensional resampling can be divided into vertical and horizontal one-dimensional resampling. Therefore, in the case of SAR,
First, it can be divided into one-dimensional resampling only in the azimuth direction, and then one-dimensional resampling only in the range direction. Therefore, when performing one-dimensional resampling in the azimuth direction, the phase shift in the azimuth direction of the reproduced complex image is corrected, and then the phase shift in the range direction is corrected by multiplying the image after the azimuth direction resampling by a phase rotation factor. After that, one-dimensional resampling in the range direction is performed. As a result, one-dimensional resampling in each direction of range and azimuth is always performed on data whose phase shift has been corrected, so that deterioration in image quality due to resampling can be prevented. The same applies to the case where one-dimensional resampling in the range direction is performed first.

〔Example〕

An embodiment of the present invention will be described below with reference to FIGS. 1 and 2.

Figure 2 shows the entire SAR system. Radar sensor 1. A SAR having an antenna 2 is mounted on an artificial satellite or the like and images the surface while moving on a flight path 3 in the direction of an arrow 4. Imaging data from the SAR is received by a ground station 6 and processed by a data processor 6 to create a video film 7, a magnetic tape 8 for data storage, and the like.

Note that 9 indicates a resolution cell, 10 indicates a range direction on the ground surface of data collected by SAR, 11 indicates an azimuth direction, 12 indicates an antenna beam, and 13 indicates a cutting width.

FIG. 1 is a flowchart of complex image resampling processing during SAR image reproduction processing according to the invention method.

It is assumed that the SAR complex image immediately after the azimuth compression process has the same phase in the range direction. At this time, as described above, there is a phase shift of exp[2πjβt] in the azimuth direction.

First, in step 21, the complex image 25 after azimuth compression is multiplied by a phase rotation factor exp (-2πjβt) to correct the phase shift in the azimuth direction.Next, in step 22, one-dimensional resampling in the azimuth direction is performed. Next, in the step 23, exp[2πjβt] is again applied to the image data resampled in the azimuth direction.
The phase shift in the range direction is corrected by multiplying by a phase rotation factor.

Finally, in step 24, one-dimensional resampling in the range direction is performed to complete the two-dimensional resampling process and obtain a reproduced image 26.

As described above, in this embodiment, the azimuth direction resampling is performed first, but when performing the range direction resampling first, the image after the range direction resampling is -multiplied by the phase rotation factor, It goes without saying that the same effect can be obtained by correcting the positional deviation in the azimuth direction.

〔Effect of the invention〕

As described above, according to the present invention, resampling processing is always performed on data whose phase shift has been corrected.
This has the effect of preventing deterioration of image quality due to phase shift and enabling high-quality reproduction of SAR images.

[Brief explanation of the drawing]

FIG. 1 is a flow diagram of SAR complex image sampling processing according to an embodiment of the present invention, and FIG.

Claims (1)

[Claims] 1. In an image processing system that reproduces an image from imaging data obtained by synthetic aperture radar, a feature is that the reproduced complex image is subjected to one-dimensional resampling in the range or azimuth direction and the image is multiplied by a phase rotation factor. Image reproduction processing method for synthetic aperture radar.