JPS58113876A - Auto-focus processing device for image of synthetic aperture radar - Google Patents

Abstract

PURPOSE:To sharply improve the precision of the titled device and the whole processing time required therein, by providing such a simple means as detects the shift of image output data between different two looks. CONSTITUTION:Data taken by a composite open-face radar are inputted in a range compressing unit 2. Then the data are re-arranged in an azimuth direction by a corner turning circuit 3 and are sent thereafter to an azimuth compressing unit 4. In an azimuth compressing circuit 44, a signal in a frequency region which is inputted after the compensation of range curvature is multiplied by an azimuth reference function and correlated therewith. The signal thus processed is filtered in a multilook filtering circuit 47 corresponding to a frequency of each look. Next, data of each look, which are a complex number, are converted into an absolute value for each look by an absolute value computing circuit 5 and delivered to an image shift detecting unit 7. The shift of image outputs between different two looks, i.e. an image shift, is detected by the image shift detecting unit 7, and said processing is continued until the image shift disappears.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention is a synthetic aperture radar autofocus image processing device that automatically focuses two-dimensional image data acquired by a synthetic aperture radar through digital processing.

By emitting radio waves from a side-looking radar mounted on a mobile platform such as an aircraft or an artificial satellite to the ground on the side of the mobile platform, and receiving and synthesizing the reflected waves while moving, a relatively small aperture can be created. Synthetic aperture radar, which can effectively combine a large-diameter antenna with a surface antenna, is well known.

FIG. 7 is an operational perspective view showing the principle of No. 11, which realizes a synthetic aperture radar using a side-looking radar mounted on a first scanning moving platform.

A mobile platform such as an aircraft (9) or a satellite that moves at a speed of V along a specific route or orbit set in advance according to a desired purpose is mounted on a small-diameter side-looking radar at a point A at an altitude of h above the ground. The antenna emits transmission pulses at regular time intervals. This transmitted pulse is radiated in the traveling direction and the angle direction with the beam width β expanding,
The reflected waves from the area BCDE on the ground are received by the side-looking radar.

These reflected waves are input one after another while the mobile platform is moving at a speed V, and are passed along the ground along a line 1. Amplitude information and phase information are recorded from 1 to 1 with the received signal at each time point while scanning between C and C. For example, a point target P on the ground located at a short distance in azimuth φ from the moving platform starts receiving the irradiation of the transmitted pulses while it is on the moving platform's line of travel, and increases as the platform moves. The reception of the transmission pulse ends at point G.

Reflected waves from point targets P are received while emitting transmit pulses, and the received signals contain distance information as well as phase information corresponding to constantly changing relative velocities.By processing this received signal, these point targets are The set of images is output as image information. The transmission pulse usually uses a linear 1M pulse that modulates the frequency of the LeF wave at a constant rate of change in order to improve distance resolution. This linear F'
This is part of the pulse compression technology commonly used in synthetic aperture radars to improve range resolution.This pulse compression technology reduces the pulse width instead of increasing the peak value output of the transmitted pulse. By adding linear FM to this pulse, the occupied bandwidth is widened and a resolution equivalent to that of a short pulse is obtained.In range compression for image processing, the signal is processed through dispersion slow abrasive powder, etc., which has the opposite frequency vs. time delay characteristics. The spectrum is concentrated at one point and output as a sharp pulse.

While the mobile platform is moving at a preset travel distance IL at a speed V, the side-looking radar acquires information on successive changes in relative orientation. Emit pulses and receive reflected waves from the target. After a certain period of time, another pulse is transmitted at the position of the fire, and in this way, the received signal containing the distance, relative velocity, or direction information obtained at each position is multiplied by 5 - times the amount of phase contained in the phase information. By synthesizing in response to the changes, it is possible to provide a function as a synthetic aperture radar that can effectively obtain the same effect as when using an antenna with an extremely long aperture diameter.

This synthetic aperture radar has excellent resolution, especially azimuth resolution, but in order to fully realize this function, it is necessary to perform phase correction for signals received at many different relative positions, and to As the distance decreases over time, range curpture correction must be performed to account for the fact that distance information about the same target is included in different range samples. This is the purpose of Image 7 Orcus. It is well known that the force resolution ΔD when focus processing is performed is expressed by the following equation (1).

ΔD = D / 2 m・
...(1) In equation (1), D is the aperture diameter of the side-looking radar antenna to configure the synthetic aperture radar, and m varies depending on the conditions when providing the synthetic aperture radar function. The following coefficient indicates the utilization rate of the antenna of the side-looking radar and is usually between 0.5 and 1.

If focusing is not performed like in a non-focusing side-looking radar, the azimuth resolution is Δ
It is also well known that it is denoted by D'.

ΔD′ = 1/2・fth “ ・・・・・・(
2) In equation (2), λ is the wavelength of the transmitted signal, and R is the distance to the target. Obviously, in the case of formula (1) in which focusing is performed, high resolution can be obtained that is not affected by the transmission frequency used or the distance at the target beam.

Therefore, if an image is to be reproduced faithfully, the above-described correction must be incorporated and processed during azimuth compression.

FIG. 2 is a received signal g special diagram showing the characteristics of a received signal from a point target on the ground. The received h signal due to the transmission pulses transmitted at a constant repetition interval Δt is shown in the figure as the oblique distance R to the scanning width d corresponding to the transmission pulse width changes over time in response to changes in the azimuth angle φ. The beam of the antenna of the side-looking radar that constitutes the synthetic aperture radar is received for a time t from the time it starts irradiating the point target to the time it finishes irradiating it, and therefore the beam from the point target is The received signal spreads in both distance and direction as the moving platform advances, and requires processing to compress the signals and reconstruct them as the original point target.

In addition, the signal received from the target is usually compressed by dividing the area covered by the radar beam from the time the radar beam starts radiating the target until the end of radiating the target, that is, the area BCDE in the case of Figure 1, into multiple equal sections. Correlation processing is performed between the data in each section. Each such section is called a look. FIG. 3 is a look perspective view for explaining the look in synthetic aperture radar image processing, and al, a2 ---21, etc. represent this look.

In azimuth compression processing, the azimuth direction spectrum of the received signal is divided into a plurality of such looks in order to reduce coherent noise after image processing, each of which is independently azimuth compressed, and then multi-look processing, which will be described later, is performed. ing. Of the range and azimuth compression processes mentioned above, range compression first compresses a two-dimensionally spread target in the Lemme direction, and is generally processed by correlating the conjugate function of the received signal and the transmitted signal. are doing.

Azimuth compression is performed by correlating range-compressed data with a conjugate function of a received signal that includes phase information that changes in response to azimuthal thermal changes. This range,
and the conjugate functions required for correlation in azimuth compression processing are range reference functions.

It is an azimuth function.

Therefore, how accurately these two reference functions are set to perform compression processing is a basic condition for obtaining a high-quality image with good contrast and little image shift.

Now, if we represent the reflected signal from a point target with distance r, orientation xK, and reflectance l as a two-dimensional impulse response h(x, r), then h(x, r) can be expressed by the following equation (3). I can do it.

9- h(x, r)=hl(x, r)Hh2(x, r)
・”(3) In equation (3), hs(Xlr) is a function that includes phase information that changes in response to changes in the relative distance between the moving platform and the target, and 112(x, r) is the waveform of the transmitted pulse. It is a function corresponding to information, that is, distance information, and the symbol ≠ indicates \convolution integral.

The received signal consists of these hl(x, r) and h2(x,
It can be reproduced by performing convolutional integration using the reference function derived from r).

h+(Xlr) contains azimuth information, that is, phase information that changes as the moving platform advances, and is a function of the Doppler frequency as described later, and h2(x, r) is expressed as a function of the modulation rate of the linear F'M. Ru.

Range compression performed using a range reference function derived from h2(Xlr) only needs to process distance information for each transmission pulse repetition, and this can be accomplished relatively easily using the linear FM modulation rate of the transmission pulse.

In this case, it is sufficient to correlate the data input sequentially.

The azimuth reference function derived from 10-h+cx+r) needs to be constantly updated because the input signal contains phase information that varies with distance.

This azimuth reference function can be expressed as a function of the time rate of change of the Doppler frequency received by the input received signal, as described below.

As the moving platform approaches and moves away from the point target, the reflected signal undergoes a Doppler shift. If this Doppler shift is a time function FD(t), then FD(t)
has a change amount that varies depending on the parameters of the speed and measure of the moving flatform, the speed of the target due to the earth's rotation, and the direction of the transmitted beam, and the signal spectrum in the azimuth direction is broadened.

Further, the distance change ΔR between the moving platform and a certain point target can be expressed by the following equation (4).

This FD(t) is a value that changes almost linearly with time if the above parameters are constant. Therefore △Jt)
As shown in FIG. 2, changes quadratically. This means that the information from a particular point target becomes time-varying range sample data. Therefore, in order to allow the target information to be compressed in the azimuth direction to be taken into the azimuth compression means, the range curvature correction that changes the sample position in the range direction over time is performed in accordance with △R(t) in equation (4). It is necessary to do so.

The azimuth reference function derived from ht(X+r) is
It is a function corresponding to the phase information that changes with R(t) to the distance change shown by equation (4), and can be expressed as a function that is conjugate to the phase change due to the Doppler frequency shift, including the above-mentioned Doppler shift function FI) (t). If the Doppler shift function FD(t) changes almost linearly with time, then if the slope of the straight line is FD', then this slope FD', that is, the form of a complex function that includes the temporal rate of change of Doppler frequency. It can be expressed as Therefore, the most important condition for azimuth compression, which is most favorable for image reproduction, is how accurately the azimuth reference function can be obtained in which the Doppler frequency temporal change rate FD' is set.

This azimuth reference function constantly changes as the azimuth changes, and the Doppler shift time rate of change that determines the azimuth reference function is the amount of Doppler shift obtained from parameters such as the velocity of the moving platform, altitude and transmit beam direction, and the velocity of the target due to the earth's rotation. These parameters are typically calculated based on relatively low accuracy information obtained from attitude sensors mounted on mobile platforms and large-scale ground-based measurement radars, and are prone to errors and are difficult to measure. In addition to requiring a significant amount of time, extensive ground tracking is required.

By including range curvature correction, optimally performing compression to this agino-space compression, and focusing the blurred image, the synthetic aperture radar will be of high quality without deviation.

Conventionally, there are generally the following two methods for image reproduction processing including this type of focus processing.

The first method is optical image processing in which processing such as range compression and azimuth compression is performed using an optical lens system.

According to this method, although it is possible to correct the phase change for each received pulse and therefore cover the entire amount of Doppler change that the moving platform, that is, the synthetic aperture radar, receives as a parameter of the phase change, it is possible to correct the phase change for each received pulse. It is essentially difficult to achieve resolution close to that of the Buddha.

Furthermore, in this optical image processing, the acquired data is
It is displayed on a photographic film that moves in proportion to the speed of the moving platform, and the phase information of the received pulse is recorded in correspondence with the range information, and this two-dimensional record has various optical characteristics. Since range compression and azimuth compression are performed through a lens system consisting of a lens group, there are drawbacks in that errors are likely to occur in the processing, and flexibility is essentially low.

The second method is digital image processing using digital processing. According to this method, it is possible to perform image processing that eliminates the drawbacks of the first method. However, this second
The azimuth compression method is as follows: In order to make the function rigid, the parameters of the altitude, speed, and antenna beam direction j of the moving platform are determined and calculated with high accuracy17.
Must have.

As mentioned above, these parameters are usually measured using a sensor mounted on a moving brush foam and a measuring device on the ground. Disadvantages 75: It is difficult to determine the parameters, it takes a lot of time to determine the parameters, and it requires a large-scale tracking unit.

The purpose of the present invention is to eliminate the above-mentioned conventional drawbacks and to significantly improve accuracy and overall processing time by replacing the simple means of detecting the shift in m1 image output data between two looks. An object of the present invention is to provide a synthetic radar autofocus image processing system that is highly flexible and does not require a large-scale ground tracking unit.

In the synthetic aperture radar image reproduction that reproduces ground conditions as an image using radio waves using a side-looking radar mounted on a mobile platform such as an artificial satellite or the like, the Lj-device of the present invention
a range compression means for compressing received data having a dimensional spread in the distance direction of the mobile platform by multiplying it by a range reference function that is a conjugate function of a transmission pulse; a corner turning means for rearranging in the azimuth direction; a Doppler frequency time change rate signal generating means for outputting a Doppler frequency time change rate signal whose value is set to a predetermined initial value and whose value changes according to an external control signal; By multiplying the output data obtained by the corner turning means by the azimuth reference function including the Doppler frequency temporal change rate, the azimuth is compressed in the azimuth direction of the moving platform, and the azimuth direction spectrum of the data is divided into a plurality of pieces ( means for performing multi-look filter processing (dividing into a plurality of looks); absolute value calculation means for calculating the absolute value of each look from the element data obtained by the multi-look filter processing; The delay time at which the correlation value between two different signals among the signals is maximum is sent to the Doppler frequency time change rate signal generating means as the control No. 41, and the Doppler frequency time change rate is changed in accordance with the delay time. and an image shift detecting means that sequentially calculates the rate signal by -1-shift and then repeats [7] until the correlation value reaches a maximum when the delay time is zero.

Next, the present invention will be explained in detail with reference to the drawings.

FIG. 4 is a schematic diagram showing an embodiment of the synthetic aperture radar autofocus image processing apparatus according to the present invention.

The orbit attitude parameter calculation circuit 1, which calculates the parameters necessary for setting the range reference function and the initial 1'th fathom of the azimuth reference function, is under the control of a built-in program in the system control unit 9, which is received via the control line S+f. To,
From the previously known information 17- regarding the trajectory, speed, attitude, etc. of the mobile platform, parameters such as the velocity vector and position vector in the earth coordinate system regarding the mobile platform and the target necessary for the following processing are calculated, and further The Doppler frequency is calculated using the parameters and stored in the built-in memory of the system control unit IO.

Data acquired by the synthetic aperture radar is input from a data file 101 using magnetic deso to a range compression unit 2 for range compression. The range compression section 2 includes a range compression circuit 22 that performs calculations necessary for the compression process, and a range reference function generation circuit 21 that generates a range reference function necessary for this calculation.

The range compression circuit 22 includes a corner velocity Fourier transform circuit that converts a received signal in the time domain into a frequency domain, a multiplication circuit that multiplies the received signal converted into the frequency domain by a range reference function, and a multiplication circuit that performs an operation. Fast Fourier performance 3 that returns the output of the chest circuit to the time domain! (9) These circuits transform the input signal into the frequency domain, and then perform range compression by multiplying the input signal by the frequency reference signal and performing correlation processing.

The range characteristic function generation circuit 21 receives transmission pulse information stored in the built-in memory of the system control unit 9 under the control of the program of the system control unit 9 received via the control line S2, and receives this information. A range reference function having a value corresponding to is generated by a function generation circuit and converted into the frequency domain.

The range reference function is represented by 112 (μ) in the following equation (5), and is expressed as a complex conjugate function of the transmission signal.

−τ/2≦μ≦τ/2 In equation (5), k is the linear FM modulation rate of the transmission pulse, and τ is the transmission pulse width, so h2(μ
) can be set from known transmission conditions.

The non-range compressed data output from the range compression unit 2 is
After being stored in the cheater file 102, it is read out and input to the corner turning circuit 3.

The corner turning circuit 3 controls the program of the system control unit fd', which receives the range compressed outputs stored in the range direction via the control line S3 for the next azimuth compression process. Corner turning processing is performed to rearrange the parts in the azimuth direction. In this case, the data is rearranged by a data replacement operation using the magnetic disk 301.

The output of corner turning circuit 3 is cheetah faille 103
It is sent to the azimuth compression section 4 via.

The azimuth compression unit 4 includes a Doppler frequency shift circuit 41, a fast Fourier transform circuit 42 . Range curvature correction circuit 4
3. Range curvature correction circuit 43° Azimuth compression circuit 44. Doppler shift/range carbateya function generation circuit 4
5. It has an azimuth diagnostic function generation circuit 46 and a multi-look filtering circuit 47, and these circuits perform azimuth compression.

The Doppler frequency shift circuit 41 performs processing in the frequency domain for azimuth compression as well as for range compression, so the Doppler frequency shift circuit 41 shifts the entire input signal spectrum so that it falls within the range of processing by fast Fourier transform.

In this case, the 0 Doppler frequency is the Doppler frequency of the reflected wave from the target irradiated by the beam center of the synthetic aperture radar.

The Doppler frequency required for the Doppler frequency shift with respect to the orbit, the Doppler frequency required for the range curvature sleeve correction described later, and the parameters for generating the range curvature function are calculated by the orbit attitude parameter calculation circuit 1 and stored in the memory of the system control unit 9. The current Doppler frequency is input to the Doppler frequency/temporal rate of change control circuit 8 under the control of the program of the system control unit 9 received via the control line S4, and then determined. The function generation signal is input to the gJ path 45, thereby converting it into a desired function and outputting it to the Doppler frequency shift circuit and the range curve correction circuit 43.

The output of the Doppler frequency shift circuit 41 is converted into a frequency domain by a fast Fourier transform circuit 42 and then input to a range carburetor sode main path 43.

21- The range curvature correction circuit 43 is configured so that the distance between the moving platform and the target changes in a quadratic manner as is clear from equation (4) and FIG. Since data samples may be included in different range samples, the data sample position of the input received signal is changed in accordance with time so that all the target information to be compressed can be input to the azimuth compression circuit 44. In other words, range curvature correction is performed. The range carbatey function can basically be expressed as ΔR(t) obtained from equation (4), and if it can be considered that the Doppler frequency changes almost like blood particles over time, then it can be expressed as the Doppler frequency change over time. It can be expressed as a quadratic function of the rate.

This Doppler frequency time change rate is also controlled by the Doppler frequency/time change rate control circuit 8 under the control of the program of the system control unit 9 which receives the Doppler frequency stored in the system control unit 9 via the control line S4. is inputted to the Doppler shift/rensicurve function generation circuit 45 via the signal line 1 as the initial 22-term value of the Doppler frequency temporal change rate by the time change rate signal generation circuit based on time differentiation, and this function generation circuit 45 The desired range curvature function is activated, the output is sent to the range curvature correction circuit 43, and this function is applied to the eleventh section to obtain the range curvature correction.

The azimuth compression circuit 44 receives the input frequency domain signal after the range curve correction and the azimuth gauze function generation circuit 4.
A multiplication circuit performs a multiplication operation with the azimuth reference function in the frequency domain, which is also input from 5, to perform a correlation. For the azimuth reference function, we assume that tt F D (t) changes approximately linearly with time [to obtain ~
It can be expressed as h+(t) in the following equation (6).

In equation (6), FD' is the time rate of change of the Doppler frequency,
T again! and ']'2i, I is a constant determined by the beam width of the antenna.

In this embodiment, the Doppler frequency temporal change rate d1, which is received from the Doppler frequency/time change rate control circuit 8 via the signal line along with information regarding the Doppler frequency range and shift l-1, is used as the initial value (-). and decide,
This Doppler frequency temporal change rate has a certain value that is roughly estimated from previously known incident information such as the trajectory, speed, and attitude of the moving platform. This Doppler frequency time change rate is used as the 11th, and the initial approximate value (#
The azimuth compression circuit 44 generates and inputs this (7).
A built-in multiplication circuit performs a multiplication operation with the output of the range curvature correction circuit to obtain a correlation. The multi-look filtering circuit 47 reads out the output data of the azimuth compression circuit 440 and performs multi-look filtering under the control of the program of the system control 141 unit 9 received via the control line S6.

The multi-look filtering circuit 47 includes one multi-look filter for filtering the output of the azimuth compression circuit 44 corresponding to the frequency of each look, and converts the rising wave frequency for each look back into the time domain. Azimuth compression circuit 4
The output of the azimuth compression unit that performs azimuth compression in this way is stored in the data file 104 and then sent to the amplitude absolute value calculation circuit 5. is input.

The amplitude absolute value calculation circuit 5 sends the output for each look of the multi-look filtering circuit 47 to the cheater file 104.
This circuit converts the data of each look, which is a complex number, into an absolute value by an absolute value W1 calculation circuit for each look, and this output is output to the image 77 detection unit 7 and also to the multi-look processing circuit. Absolute value addition is performed using the adder circuit 601, and images of so-called non-coherent addition and look are superimposed, and the output is displayed on a CRT screen via an output line 601, with the speckle noise peculiar to coherent addition processing being reduced. output to the device.

The image shift detection unit 7 detects a deviation in the 25- image output between two different looks, that is, an image shift, and detects the image output of the two different looks converted into an absolute value by the amplitude absolute value calculation circuit 5. Multi-look correlation calculation circuit 7
1, and depending on the built-in correlation calculation circuit, different 2
Performs cross-correlation calculation of time domain data between looks. This output has a value at which the correlation value between two different looks is the highest, that is, a value corresponding to the image shift. A correction signal having a magnitude proportional to the signal is output as a correction amount, and sent to the Doppler frequency/time rate of change control circuit 8.

The Doppler frequency temporal change rate control circuit 8 that receives this output calculates the Doppler frequency temporal change rate in accordance with the output of the correction amount calculation circuit 82 while being controlled by the program of the system control unit H via the control line S4. This Doppler frequency time change rate signal is input to the azimuth reference function generation circuit 46 via the signal line 2 to change the azimuth reference function, and while repeating this operation, image shift between looks 26- is performed. Processing continues until the image is exhausted, and the image output is automatically adjusted to complete the focus.

At this time, the relationship between the elimination of image shift between two looks and the correlation function for calculating the correlation value between two looks is as follows. In other words, the two look (image) signals are
(t) + y(t) (where t indicates time and distance on the image), the correlation function of both signals R(
t) is expressed by the following equation (8).

From equation (8), the delay time t i' when R(t) is maximum
rs indicates the shift (shift) between images, and therefore means that when there is no image shift, B (t) becomes maximum at a delay time of zero. In this way, autofocusing of the image is performed while setting the optimum azimuth reference function by image shift detection, and the obtained output is outputted to a display device or the like via the image data output line 601.

In the above explanation, range compression and azimuth compression are performed after converting the input No. 48 into the wave number domain, and it goes without saying that compression in the time domain is also theoretically possible. That is, the present invention detects an image shift by calculating the correlation between two different looks obtained by azimuth compression and multi-look filtering from the amplitude component of the signal after compression processing, and The key point is to quickly set an appropriate azimuth reference function, and various other configurations can be considered.

Furthermore, in this example, the movement of the moving platform is linear, and therefore the Doppler frequency shift amount is treated as a function that changes almost linearly over time; however, the movement of the moving platform is curved, and therefore the Doppler Even if the frequency lift cannot be considered to be uniformly approximately linear across the entire processing range, it is possible to divide the entire processing range into multiple ranges that can be considered approximately linear. It is clear that the present invention can be applied to approximate processing of these individual ranges by dividing the range.

Although the embodiment of the present invention has been described above, the embodiment is based on a hard-wired electronic circuit, but it is obvious that the present invention can be applied to other modified examples as well. For example, in FIG. 4, the processing contents of the image shift detection section 7 and the Doppler frequency temporal change rate control circuit 8 can be incorporated into the built-in program of the entire system control section 9. )・−
It can also be built into the azimuth compression unit 4 or the like depending on the hardware configuration of the dwired type.

Which of these methods to select can be arbitrarily determined depending on the amount of image data to be processed and the required processing time. Furthermore, it is clear that the multi-look correlation calculation circuit 71 may perform any number of correlation calculations for two or more looks.Furthermore, in this embodiment, in order to determine the initial position of the azimuth reference function used for azimuth compression, the motion of the moving platform is The initial value of the Doppler frequency flow rate signal, which is determined from known information such as conditions, can be set using information obtained from a 7-ohm sensor mounted on a mobile platform and a speech sensor on the ground. The measurement sensor described above also sets the Doppler frequency temporal change rate signal for azimuth reference function correction, which is necessary to correct the azimuth reference function in response to changes in the phase information of the receiving angle 29- input. It is also clear that it can be determined by the information obtained from .

Any of these modifications can be easily implemented without departing from the spirit of the invention.

As explained above, according to the present invention, by providing a simple means of detecting a shift in image output data between two different looks, accuracy and overall processing time can be significantly improved. It also has the advantage of being highly flexible and capable of performing synthetic aperture radar autofocus image processing that does not require a large-scale ground tracking ring.

[Brief explanation of the drawing]

Fig. 1 is an operational perspective view showing the principle of operation of the synthetic aperture radar, Fig. 2 is a received signal characteristic diagram showing the characteristics of the received signal of the synthetic aperture radar, Fig. 3 is a look perspective view, and Fig. 4 is a FIG. 3 is a three-dimensional diagram showing one embodiment of the present invention. In Fig. 4, 1... orbit attitude parameter calculation circuit, 2...
・Range compression section, 3... Corner turning circuit, 4... Azimuth compression section, 5... Absolute amplitude calculation circuit, 6... Multi Look processing circuit, 7... Image shift extraction section, 8...
・Dob 2 frequency/time rate of change side leg circuit, 9...
System control unit. 31- ・Kiadan Tsui- too~

Claims (1)

[Claims] Aircraft 1 In synthetic aperture radar image reproduction that reproduces ground conditions as a radio wave image using a side-looking radar mounted on a mobile platform such as an artificial satellite, the received data having a two-dimensional spread acquired by the synthetic aperture radar is transmitted as a pulse. a range compression means for compressing the moving platform in the distance direction by multiplying it by a range reference function that is a conjugate function of the moving platform; and a corner turning means for rearranging the output obtained by the range compression means in the azimuth direction of the moving platform. a Doppler frequency temporal change rate signal generating means for outputting a Doppler frequency temporal change rate signal whose value is set to a predetermined value as an initial value and whose value is changed by an external control signal; and an output obtained by the corner turning means. By multiplying the data by the azimuth reference function including the Doppler frequency temporal change rate, azimuth is compressed in the azimuth direction of the moving platform, and the azimuth direction spectrum of the data is divided into several pieces (divided into multiple looks).
means for performing multi-look filter processing to perform multi-look filter processing; absolute value calculation means for calculating the absolute value of each look from the complex data obtained by the multi-look filter processing; Sending the delay time at which the correlation value between the signals is maximum to the Doppler frequency time change rate signal generating means as the control signal;
and an image shift detection means for changing the Doppler frequency temporal change rate signal in accordance with the delay time and repeatedly calculating the correlation value when the delay time is zero until the correlation value reaches a maximum. Synthetic aperture radar autofocus image processing device.