JP5419632B2 - Synthetic aperture radar equipment - Google Patents

Description

  The present invention relates to a synthetic aperture radar apparatus that is mounted on a platform such as an aircraft or a satellite and captures a high-resolution image of the ground surface or the sea surface, and particularly relates to an improvement in image reproduction processing technology.

  Conventionally, a synthetic aperture radar device mounted on an aircraft or satellite platform has been provided with an SAR (Synthetic Aperture Radar) sensor having an antenna, and the platform on which the radar device is mounted transmits and receives radio waves for observation. A two-dimensional high-resolution image is obtained by performing signal processing on the obtained radio wave signal (see, for example, Patent Document 1).

  In the conventional apparatus described in Patent Document 1, the target in the observation area is two-dimensionally spread on the received signal received by the SAR sensor. By performing the signal processing called, it is possible to compress to one point. The image reproduction processing used at this time includes various methods such as a polar format method, a range Doppler method, and a chirp scaling method.

  In the image reproduction process, FFT (Fast Fourier Transform) calculation processing, IFFT, in the range direction (radiation direction of radio waves) or azimuth direction (direction orthogonal to the radio wave irradiation direction) is performed on two-dimensional data. (Inverse Fast Fourier Transform) arithmetic processing, coefficient multiplication processing (see, for example, Non-Patent Document 1) or the like is performed.

The coefficient multiplication process of the chirp scaling method described in Non-Patent Document 1 mainly corresponds to a process of rotating and correcting the phase of each data. In this case, one of the coefficients multiplied by each data after the range IFFT The component S (τ, f _n ) is expressed by the following equation (1) using the distance R ₀ from the ground surface to the platform, the frequency f _{0 of} the radio wave used, and the relative velocity V _{r of the} platform with respect to the ground surface. .

係数乗算処理用の係数として必要な情報は、送信波に対する受信波の相対位相であり、式（１）の位相（指数部）をθとすると、位相θは、以下の式（３）のように表される。 However, in Formula (1), j is an imaginary unit.
Further, D (fn, Vr) of the exponent part is expressed by the following formula (2).

Information necessary as a coefficient for coefficient multiplication processing is the relative phase of the received wave with respect to the transmitted wave. When the phase (exponential part) of Equation (1) is θ, the phase θ is as shown in Equation (3) below. It is expressed in

式（３）において、２ｎπ（ｎは整数）を除去した最終的な位相θ’は、係数を算出するために必要な情報であり、０≦θ’＜２πの範囲内の値となる。
言い換えれば、最終的な必要情報となる位相θ’を求めるためには、式（３）で求められる位相θを２πで除算した商の小数点以下の数値（商と、商の整数部分との差）が必要である。

In Equation (3), the final phase θ ′ from which 2nπ (n is an integer) is information necessary for calculating the coefficient, and is a value in the range of 0 ≦ θ ′ <2π.
In other words, in order to obtain the phase θ ′ which is the final necessary information, the numerical value below the decimal point of the quotient obtained by dividing the phase θ obtained by the expression (3) by 2π (the difference between the quotient and the integer part of the quotient) )is necessary.

However, depending on the numerical value given to Equation (1), the phase θ can be an extremely larger value than 2π. In particular, when an SAR image is to be obtained from an artificial satellite, the frequency f ₀ to be used is on the order of GHz and the distance to the ground is 100 km or more, so the phase component of equation (1) has a very large value, There is also a possibility that the calculation error of the phase θ ′ that is required in an increased manner.
Specifically, although the values of the distance R ₀ and the frequency f ₀ are much larger than “1”, D (f _n , V _r ) represented by the equation (2) is a real number (0 A value of ˜1), which is substantially a value in the vicinity of “1”.

Therefore, the calculation error of the phase θ is likely to increase, and the increase of the phase error causes a disturbance of the SAR image and affects the resolution.
As a result, when the arithmetic circuit is configured with conventional single-precision floating point numbers or fixed-point numbers, sufficient arithmetic accuracy cannot be obtained, and the resolution of the final SAR image is deteriorated.

  However, it is known that the phase calculation error as described above hardly poses a problem if the calculation is performed with a double precision floating point. In the conventional apparatus, the SAR image reproduction processing is normally performed on the ground, and there are no physical restrictions (such as placement space and power consumption) in the arithmetic apparatus, which hinders arithmetic operations using double-precision floating point. Did not occur.

  However, when the SAR image reproduction process is performed on a platform such as an artificial satellite, the size and power consumption of the arithmetic device are limited. Therefore, when the image reproduction process is performed by a dedicated arithmetic device (arithmetic circuit). It is practically difficult to construct an arithmetic circuit with a double precision floating point.

JP 2004-198275 A

I. G. Cumming and F.M. H. Wong, “Digital processing of synthetic algorithm, radical data algorithms and implementation,” Arttech House, 2004.

The conventional synthetic aperture radar device has a problem that when the arithmetic circuit is configured with a single precision floating point number or a fixed point number, sufficient arithmetic accuracy cannot be obtained and the resolution of the final SAR image is deteriorated. .
In addition, when performing SAR image playback processing on a platform such as an artificial satellite, the size and power consumption of the computing device are limited, so double-precision floating point calculations are performed to improve accuracy in image playback processing. When the circuit is configured, there is a problem that the circuit scale is increased and the cost is increased.

  The present invention has been made in order to solve the above-described problems, and provides an image reproduction processing means that is smaller in apparatus size than that configured with double-precision floating-point numbers and that can perform calculations with sufficient calculation accuracy. It is an object of the present invention to provide a synthetic aperture radar apparatus.

A synthetic aperture radar device according to the present invention is a synthetic aperture radar device which is mounted on a movable platform and obtains a SAR image of the ground surface or the sea surface, and radiates a high frequency pulse signal and receives a reflected signal of the high frequency pulse signal. A SAR sensor that outputs a received signal for obtaining a SAR image, a motion sensor that measures the motion of the platform and outputs the instantaneous position of the platform when transmitting and receiving a high-frequency pulse signal as motion data, and a received signal an image reproduction processing means for performing image reproduction processing on the basis of, when obtaining the SAR image by using the chirp scaling method based on the motion data, coefficient calculation means for calculating the coefficient values to be used in the coefficient multiplication processing of the chirp scaling method with the door, coefficient calculation means, engagement used in the coefficient multiplication processing from the image reproduction processing unit Based on the request, when calculating the coefficient value corresponding to the request, the range direction and y-coordinate at the time of transmission of the high frequency pulse signal, the azimuth direction as the x-coordinate, the coefficients of the phase, and the variable y coordinate and x-coordinate By expressing with the above formula, it is divided into a minute fluctuation component that fluctuates slightly and an invariant component that does not change at all, and the phase component of the coefficient is calculated for each, and then each phase component of the coefficient is digitized and added. Thus, the coefficient value that is finally output is calculated .

  According to the present invention, it is possible to realize a synthetic aperture radar apparatus including an image reproduction processing unit that can be reduced in apparatus scale and can be operated with sufficient calculation accuracy as compared with the case of being configured with double precision floating point numbers. Can do.

It is a block diagram which shows the structure of the synthetic aperture radar apparatus which concerns on Embodiment 1 of this invention. It is a flowchart which shows the program content of the synthetic aperture radar apparatus which concerns on Embodiment 1 of this invention. It is a flowchart which shows the outline | summary of the chirp scaling method in Embodiment 1 of this invention. It is a block diagram which shows the specific structure of the coefficient calculation means by Embodiment 1 of this invention. It is a block diagram which shows the specific structure of the coefficient calculation means by Embodiment 2 of this invention. It is a block diagram which shows the specific structure of the coefficient calculation means by Embodiment 3 of this invention.

Embodiment 1 FIG.
1 is a block diagram showing a configuration of a synthetic aperture radar apparatus according to Embodiment 1 of the present invention.
In FIG. 1, a synthetic aperture radar device mounted on a movable platform (not shown) such as an aircraft or a satellite includes a SAR sensor that outputs a received signal 2, a motion sensor 3 that outputs motion data 4, and a reception. Data holding means 5 for holding the signal 2, coefficient calculating means 6 for calculating coefficients necessary for image reproduction (described later), and image reproduction processing means 7 for generating the SAR image 8 based on various data and coefficients. I have.

  The SAR sensor 1 is a generic term for sensor devices, and includes an antenna, a transmitter, and a receiver (all not shown), generates a high-frequency pulse signal, transmits and receives the high-frequency pulse signal, and reproduces the SAR image 8. The reception signal 2 necessary for the acquisition is acquired.

  Specifically, the SAR sensor 1 first generates a high-frequency pulse signal, radiates the high-frequency pulse signal from the antenna to the space, and receives the echo signal reflected by the target via the antenna. Subsequently, the reception signal from the antenna is amplified and converted to an intermediate frequency, and then converted to a digital signal to output a final reception signal 2.

  The motion sensor 3 measures the motion of the platform on which the synthetic aperture radar device is mounted, and outputs the instantaneous position of the platform at the time of transmission / reception of a high-frequency pulse signal as motion data 4.

The data holding unit 5 holds the reception signal 2 input from the SAR sensor 1 as data, and holds intermediate data 11 and 12 input from the image reproduction processing unit 7.
The coefficient calculation means 6 calculates various parameters based on the motion data 4 obtained from the motion sensor 3, and receives coefficient request signals from the image reproduction processing means 7, and responds to each coefficient multiplication process based on the various parameters. A coefficient value G (z) is calculated and input to the image reproduction processing means 7.

  The image reproduction processing means 7 calculates various parameters from the motion data 4, requests each coefficient to the coefficient calculation means 6, and uses the coefficients obtained from the coefficient calculation means 6 to use the data of the received signal 2. An image reproduction process is performed on the SAR image 8 to generate a SAR image 8.

Although not shown in FIG. 1, the synthetic aperture radar apparatus actually includes a control unit that performs overall control. The control means is connected to the data holding means 5, the coefficient calculation means 6 and the image reproduction processing means 7 via a system bus or a control signal line, and controls each means 5-7.
However, in the following description, for the sake of simplification, the data holding unit 5, the coefficient calculating unit 6, and the image reproduction processing unit 7 will be described as executing data processing by themselves.

Next, the overall operation according to the first embodiment of the present invention shown in FIG. 1 will be described with reference to the flowcharts of FIGS.
FIG. 2 is a flowchart schematically showing an image reproduction processing program executed by the synthetic aperture radar apparatus (control means (not shown)).
As the image reproduction process, the polar format method, the range Doppler method, the chirp scaling method, and the like described above are known. Here, the case of the chirp scaling method is shown as an example.

FIG. 3 is a flowchart showing the chirp scaling process, and clearly shows coefficient multiplication processes M1 to M5 included in steps ST4, ST6, and ST8 in FIG.
That is, steps ST101 to ST103 in FIG. 3 correspond to step ST4 in FIG. 3, steps ST104 to ST106 in FIG. 3 correspond to step ST6 in FIG. 3, and steps ST107 to ST109 in FIG. Corresponds to step ST8 in FIG.

  As shown in FIG. 3, the chirp scaling method uses coefficient multiplication processing M1 to M5 (steps ST101, ST103, ST105, ST107, ST109) with the received signal 2 or intermediate data in the data holding means 5, and coefficient multiplication processing M1. FFT in data units in the azimuth direction (step ST102) executed between the coefficient multiplication process M2 and the coefficient multiplication process M2, and FFT (step ST104) and IFFT in data units in the range direction executed before and after the coefficient multiplication process M3. (Step ST106) and IFFT in units of data in the azimuth direction (step ST108) executed between the coefficient multiplication process M4 and the coefficient multiplication process M5.

In FIG. 2, the synthetic aperture radar apparatus first performs signal transmission / reception processing (step ST1) by the SAR sensor 1.
That is, a high-frequency pulse signal is generated, the high-frequency pulse signal is radiated from the antenna of the SAR sensor 1 to the space, an echo signal reflected by the target is received by the antenna of the SAR sensor 1, and the received signal is amplified. Then, it is converted into an intermediate frequency and converted into a digital signal to obtain a received signal 2.
Subsequently, the data holding means 5 reads the received signal 2 input from the SAR sensor 1 and stores it as data (step ST2).

On the other hand, the motion sensor 3 outputs the platform motion data 4 measured together with the transmission and reception of pulses, and the coefficient calculation means 6 reads the motion data 4 measured by the motion sensor 3 and sets various parameters based on the motion data. calculate.
Further, the coefficient calculation means 6 reads the platform motion data 4 measured by the motion sensor 3 in association with the image reproduction processing means 7, and calculates coefficients used for the image reproduction processing based on the motion data 4. (Step ST3).

Next, in FIG. 2 and FIG. 3, the image reproduction processing means 7 performs the coefficient multiplication and azimuth FFT (step ST 4), out of the data of the received signal 2 stored in the data holding means 5, in the azimuth direction. Data is read out in units of rows, and the coefficient calculation means 6 is requested for coefficients used in coefficient multiplication processing M1 and M2 corresponding to each pixel to be processed.
In response to this coefficient request, the coefficient calculation means 6 calculates a necessary coefficient value G (z) from the various parameters calculated in step ST3 and inputs it to the image reproduction processing means 7.

  Thereafter, the image reproduction processing means 7 performs a coefficient multiplication process M1 in units of one row of data in the azimuth direction (step ST101), and subsequently executes an azimuth FFT process (step ST102), and again using a different coefficient. Then, coefficient multiplication processing M2 is performed (step ST103).

Subsequently, the image reproduction processing means 7 determines whether or not the coefficient multiplication / azimuth FFT (step ST4) is completed (step ST5), and it is determined that the coefficient multiplication / azimuth FFT is not completed (that is, NO). Then, the process returns to step ST4 (ST101).
Thereby, the image reproduction processing means 7 repeatedly executes the coefficient multiplication / azimuth FFT (steps ST4, ST101 to ST103) until the processing is completed for all the rows.

  On the other hand, if it is determined in step ST5 that the coefficient multiplication / azimuth FFT has been completed (that is, YES), the image reproduction processing means 7 inputs the data for which the coefficient multiplication / azimuth FFT processing has been completed to the data holding means 5. . This data is held in the data holding means 5 as intermediate data 11.

  Next, the image reproduction processing means 7 reads out the data in the range direction from the intermediate data stored in the data holding means 5 in units of one column in order to perform coefficient multiplication / range FFT / IFET (step ST6).

  Similarly to the case of step ST4 (ST102), the image reproduction processing means 7 requests the coefficient calculation means 6 for a coefficient used for the coefficient multiplication process M3 corresponding to each pixel to be processed. In response to this coefficient request, the coefficient calculation means 6 calculates a necessary coefficient value G (z) from the various parameters calculated in step ST3 and inputs it to the image reproduction processing means 7.

  Thereafter, the image reproduction processing means 7 performs FFT processing (step ST104) on the data in the range direction in units of columns, then performs coefficient multiplication processing M3 (step ST105), and further performs range IFFT processing (step ST106). ).

Subsequently, the image reproduction processing means 7 determines whether or not the coefficient multiplication / range FFT / IFET (step ST6) is finished (step ST7), and the coefficient multiplication / range FFT / IFET is not finished (that is, NO). If determined, the process returns to step ST6 (ST104).
Thereby, the image reproduction processing means 7 repeatedly executes the coefficient multiplication / range FFT / IFET (steps ST6, ST104 to ST106) until the processing is completed for all the columns.

  On the other hand, if it is determined in step ST5 that the coefficient multiplication / range FFT / IFET has been completed (that is, YES), the image reproduction processing means 7 uses the data holding means to store the data for which the coefficient multiplication / range FFT / IFET processing has been completed. Enter 5. This data is held in the data holding means 5 as intermediate data 12.

Next, the image reproduction processing means 7 reads out the data in the azimuth direction among the intermediate data stored in the data holding means 5 in units of one line in order to perform coefficient multiplication and azimuth IFFT (step ST8).
The image reproduction processing means 7 requests the coefficient calculation means 6 for coefficients for coefficient multiplication processes M4 and M5 corresponding to each pixel to be processed.
In response to this coefficient request, the coefficient calculation means 6 calculates the necessary coefficient value G (z) from the various parameters calculated in step ST3, as in steps ST4 and ST6, and sends it to the image reproduction processing means 7. input.

Thereafter, the image reproduction processing means 7 performs the coefficient multiplication process M4 in units of one row of data in the azimuth direction (step ST107), and subsequently executes the azimuth IFFT process (step ST108), and again using different coefficients. Then, coefficient multiplication processing M5 is performed (step ST109).
Subsequently, the image reproduction processing means 7 sequentially outputs the final data subjected to the azimuth IFFT processing as the SAR image 8 (step ST9).

Finally, the image reproduction processing means 7 determines whether or not the coefficient multiplication / azimuth IFET processing (steps ST8 and ST107 to ST109) and the SAR image output processing (step ST9) are finished (step ST10), and is not finished ( In other words, if NO is determined, the process returns to step ST8 (ST107).
Thereby, the image reproduction processing means 7 repeatedly executes steps ST8, ST107 to ST109 and step ST9 until the processing is completed for all the rows.

  On the other hand, if it is determined in step ST10 that the coefficient multiplication / azimuth IFET process and the SAR image output process have been completed (that is, YES), the processing routines in FIGS. 2 and 3 are terminated.

Next, the specific configuration and operation of the coefficient calculation means 6 in FIG. 1 will be described with reference to the block diagram of FIG.
Here, as an example, a case where the calculation of the above-described formula (1) is performed by the coefficient calculation means 6 will be described.

Further, 1024 azimuth direction data width _{A z} of the received signal 2 ^{(= 2 10),} and 1024 data width _{R g} in the range direction ^{(= 2 10),} coordinates (azimuth coordinates) x in the azimuthal pixels, An integer from 0 to 1023 and a pixel coordinate (range coordinate) y in the range direction are integers from 0 to 1023.
In order to simplify the description, in the equation (1), the frequency f _n is a variable of the azimuth coordinate x, and is expressed as the following equation (4).

In the equation (1), the distance R ₀ is a variable of the range coordinate y and can be expressed as the following equation (5).

Further, in the expression (1), values such as the relative velocity V _r are determined by the motion data 4, but here, it can be regarded as a fixed value in one SAR image 8.

  In FIG. 4, the coefficient calculation means 6 includes a parameter calculation means 21, phase tables 22a and 22b, multipliers 23a and 23b, dividers 24a and 24b, a shifter 25, an adder 26, and a coefficient table 27. It has.

The parameter calculation means 21 receives the motion data 4 and calculates parameters for phase calculation and table data to be stored in the phase tables 22a and 22b from each numerical value in the motion data 4.
The phase tables 22a and 22b store the table data sent from the parameter calculation means 21, and output the table data corresponding to the azimuth coordinates x.

The multiplier 23a multiplies the phase calculation parameter (calculated value a ′ described later) from the parameter calculation means 21 by the range coordinate y.
The multiplier 23b multiplies the table data from the phase table 22a and the range coordinate y.

The divider 24a calculates a remainder obtained by dividing the multiplication result of the multiplier 23a by 2 ¹⁶ (= 65536).
The divider 24b calculates the remainder obtained by dividing the addition result of the adder 26 by 2 ¹⁶ as the division result z.

Shifter 25 based on an instruction of the parameter calculation means 21, m times a multiplication result of the multiplier 23b (m is an integer) after, input to the adder 26 to calculate the remainder of division by 2 ^16.
The adder 26 adds the division result of the divider 24a, the calculated value of the shifter 25, and the table data from the phase table 22b.
The coefficient table 27 outputs a coefficient value G (z) corresponding to the division result z of the divider 24 b and inputs it to the image reproduction processing means 7.

Next, the operation of the coefficient calculation means 6 will be described with reference to FIGS.
First, the operation in the exercise data reading process (step ST3) in FIG. 2 will be described.
In step ST3, when the parameter calculation means 21 in the coefficient calculation means 6 receives the motion data 4, it creates a parameter for phase calculation and table data.

Specifically, the above formula (1) is expanded to calculate various parameters.
In the formulas (1) to (3), as described above, D (f _n , V _r ) is a real number and is in the range of 0 ≦ D (f _n , V _r ) ≦ 1, Specifically, it takes a value near “1”.

The relative velocity V _{r of the} platform is a fixed value, and the frequency f _n is a variable of the azimuth coordinate x.
Furthermore, when the distance R ₀ is expressed as the above-described equation (5), the variable D ′ (x) related to the azimuth coordinate x is expressed as the following equation (6).

In Expression (6), the variable D ′ (x) is 0 ≦ D ′ (x) ≦ 1, and takes a value in the vicinity of “0”.
Further, the phase θ represented by the above equation (3) can be transformed as the following equation (7) as variables of the azimuth coordinate x and the range coordinate y using the equations (4) and (5). can do.

The parameter calculation means 21 multiplies the coefficient a by 2 ¹⁶ to take an integer value (truncates the fractional part), calculates the remainder divided by 2 ¹⁶ , and obtains the calculated value a ′.
The parameter calculating unit 21 takes the integer value by multiplying the case of changing the azimuth coordinate x from 0 to 1023, the variable b {1-D '(x )} 2 16 to each value ^{of 2 16} The remainder divided by is calculated, and this calculated value is stored in the phase table 22b.
Further, the parameter calculation means 21 calculates the maximum value D _max of the variable a · D ′ (x) when the azimuth coordinate x is changed from 0 to 1023, and calculates an integer m satisfying the following equation (8). To do.

2 ^−m−1 ≦ D _max <2 ^−m (8)

However, if D _max ≧ 1, m = 0.
The variable D (f _n , V _r ) represented by the expression (2) has a monotonically increasing relationship with respect to the absolute value of the frequency f _n , and the variable D ′ (x) has x = 0 or x = 1023, which is the maximum, the maximum value D _max can be easily calculated.

After calculating the integer m, the parameter calculating unit 21 takes the integer value by multiplying the case of changing the azimuth coordinate x from 0 to 1023, the variable a · D '2 ¹⁶ to each value of (x), The remainder divided by 2 ¹⁶ is calculated, and this calculated value is stored in the phase table 22a.

Next, the operation when the coefficient value G (z) is output from the coefficient calculation means 6 will be described.
Since the coefficient of the equation (1) is related to the coefficient multiplication process M4 (step ST107) in FIG. The coefficient value G (z) is output from (1) to (4).

  (1) When information on the azimuth coordinate x (any value from 0 to 1023) is sent from the image reproduction processing means 7, the phase tables 22a and 22b output stored data corresponding to the azimuth coordinate x. .

(2) Subsequently, the multiplier 23a multiplies the calculated value a ′ from the parameter calculation means 21 and the value of the range coordinate y and outputs a multiplication result, and the divider 24a outputs the multiplication result of the multiplier 23a. Is divided by 2 ¹⁶ to calculate the remainder.
The multiplier 23b multiplies the table data from the phase table 22a and the value of the range coordinate y and outputs a multiplication result. The shifter 25 divides the multiplication result of the multiplier 23b by 2 ^m , -1 "is multiplied to take an integer value, and further multiplied by 2 ¹⁶ to calculate the remainder.

(3) Next, the adder 26 takes the sum of the division result of the divider 24a, the calculated value of the shifter 25, and the table data from the phase table 22b, and the divider 24b is the addition result of the adder 26. Is divided by 2 ¹⁶ , and the division result z is input to the coefficient table 27.

(4) Finally, the coefficient table 27 outputs a coefficient value G (z) corresponding to the division result z of the divider 24b.
Specifically, when the division result z (0 ≦ z <2 ¹⁶ ) of the divider 24b is used, the coefficient value G (z) from the coefficient table 27 is represented by a complex number as in the following Expression (9). The

The coefficient value G (z) corresponds sin function in the range of 0～2Pai (imaginary part) and cos function (real part) to ^{2 16} equal parts table.
The above processes (1) to (4) performed by the coefficient calculation means 6 divide the numerical value of the coefficient a into an integer part and a part below the decimal point, for example, and that the azimuth coordinate x and the range coordinate y are integers. This corresponds to eliminating the component corresponding to 2nπ in the formula (3).
Thus, by normalizing with 2 ¹⁶ the following fractional part, it is possible to easily eliminate an integer part after multiplying.

In the above operation by the coefficient calculating means 6, the numerical value (motion data 4) given to the parameter calculating means 21 needs to be a double-precision floating point number, but all the numerical values stored in the phase tables 22a and 22b are adjusted. It is a numerical value.
When these values are expressed as binary fixed-point numbers, the process of calculating the remainder by dividing by 2 ¹⁶ corresponds to the process of extracting the lower 16 bits and is easy to calculate. Similarly, the process of dividing by 2 ^m can also be processed with an mbit shift.

  Further, the phase tables 22a and 22b can be realized by a memory (such as a dual port memory) in which input and output are individually separated, and the reading side may use the azimuth coordinates x as a read address.

The coefficient table 27 can be realized by a ROM (Read Only Memory) or the like, and the division result z of the divider 24b may be used as an address.
The coefficient table 27 is has a table where the sin function and cos function ²¹⁶ equal parts, when the phase accuracy insufficient, by changing the table size, it is possible to obtain a desired accuracy . For example, the process of dividing by 2 ¹⁶ may be changed to the process of dividing by 2 ^k , and the integer k (> 16) corresponds to the precision bit length. On the other hand, if the accuracy is too high, the device scale can be reduced by shortening the bit length.

  In this way, the calculation of the phase calculation is divided into a fixed part with a large value and a part that fluctuates slightly, and after adding the final digit, the final phase is calculated, In a normal floating point calculation, the phase of a minute fluctuation portion that tends to be buried in a fixed portion having a large value can be calculated with high accuracy, and the scale of the apparatus can be reduced.

In the above description, only the processing corresponding to the expression (1) is shown, but in the processing of other coefficients, the phase table is decomposed into the term of the azimuth coordinate x and the term of the range coordinate y as described above. Prepare and calculate the coefficients.
In the first embodiment of the present invention, it is necessary to prepare the phase tables 22a and 22b for each phase component, but it is a phase table of azimuth (x) words or range (y) words at most. Compared to, the size is sufficiently small so that there is no problem.

4 shows the case where the numerical values sent to the phase tables 22a and 22b are the azimuth coordinates x, but depending on the coefficient calculation formula, the numerical values sent to the phase tables 22a and 22b are different from the case of FIG. The numerical value may be the range coordinate y.
Further, although the case where the value processed by the multipliers 23a and 23b is the range coordinate y is shown, the value processed by the multipliers 23a and 23b may be the azimuth coordinate x, and the multiplication between the phase tables 22a and 22b. It can be. However, in any of the above cases, the configuration of the coefficient calculation means 6 is the same as in FIG.

Further, in calculating each coefficient corresponding to the coefficient multiplication processing M1 to M5, multiplication of an exponential function is also necessary. This can be processed by exponent addition (phase addition), and finally an adder. After the addition at 26, the phase component may be input to the coefficient table 27.
Therefore, only one coefficient table 27 is prepared as shown in FIG.

  As described above, the synthetic aperture radar apparatus according to Embodiment 1 (FIGS. 1 to 4) of the present invention is a synthetic aperture radar apparatus that is mounted on a movable platform and obtains the SAR image 8 of the ground surface or the sea surface. The SAR sensor 1 that radiates a high-frequency pulse signal, receives a reflection signal of the high-frequency pulse signal, and outputs a reception signal 2 for obtaining the SAR image 8, and measures the movement of the platform. A motion sensor 3 that outputs the instantaneous position of the platform at the time of transmission / reception as motion data 4, an image playback processing means 7 that performs image playback processing based on the received signal 2, and a coefficient value G for image playback based on the motion data 4 Coefficient calculation means 6 for calculating (z).

The coefficient calculation unit 6 calculates a coefficient value G (z) corresponding to the request based on the request from the image reproduction processing unit 7.
Also, the coefficient calculation means 6 divides the change component in the range direction and the change component in the azimuth direction at the time of transmitting the high-frequency pulse signal, calculates the phase component of the coefficient for each, and then multiplies each phase component of the coefficient And the coefficient value G (z) finally output is calculated by adding.

  Further, the coefficient calculation means 6 divides the change in the range direction or the azimuth direction at the time of transmission of the high-frequency pulse signal into a minute fluctuation component that fluctuates slightly and an invariant component that does not change at all, and the accuracy of the minute fluctuation component increases. The phase component of the coefficient is calculated as described above, and then the coefficient value G (z) that is finally output is calculated by multiplying and adding each phase component of the coefficient.

  As a result, the apparatus scale can be reduced as compared with the case where the apparatus is composed of double precision floating point numbers, and a synthetic aperture radar apparatus provided with the image reproduction processing means 7 capable of calculating with sufficient calculation accuracy can be obtained.

Embodiment 2. FIG.
In the first embodiment (FIG. 4), two multipliers 23a and 23b and two dividers 24a and 24b are provided in the coefficient calculation means 6, but as shown in FIG. A single multiplier 23 and divider 24 may be provided in 6A.

FIG. 5 is a block diagram showing the coefficient calculation means 6A of the synthetic aperture radar apparatus according to Embodiment 2 of the present invention. The same components as those described above (see FIG. 4) are denoted by the same reference numerals as those described above. The description is omitted.
The overall configuration of the synthetic aperture radar apparatus according to the second embodiment of the present invention is as shown in FIG. 1, and the basic processing operation is also as shown in FIGS.

  In this case, when calculating the component of each coefficient, only the phase tables 22a to 22d corresponding to each component are prepared, and the multipliers 23a and 23b and the dividers 24a and 24b described above (FIG. 4) The multiplier 23 and the divider 24 are shared.

  In FIG. 5, the coefficient calculating means 6A includes a parameter calculating means 21, phase tables 22a to 22d, a multiplier 23, a divider 24, a shifter 25, an adder 26, a coefficient table 27, a selector 28a, 28b and a register 29.

The phase tables 22a and 22b hold the table data from the parameter calculation means 21 and input the table data corresponding to the azimuth coordinates x from the image reproduction processing means 7 to the selector 28a.
The phase tables 22c and 22d hold the table data from the parameter calculation unit 21 and input the table data corresponding to the range coordinate y from the image reproduction processing unit 7 to the selector 28a.

The selector 28a selects any one of the table data from the phase tables 22a to 22d according to the coefficient required for the processing, and inputs the selected data to the multiplier 23.
The selector 28 b selects either the azimuth coordinate x or the range coordinate y and inputs it to the multiplier 23.

The multiplier 23 multiplies the input information from the selectors 28 a and 28 b and inputs the multiplication result to the shifter 25.
Shifter 25 based on an instruction of the parameter calculation means 21, and inputs the multiplication result of the multiplier 23 after multiplied by m, the adder 26 calculates a remainder of division by 2 ^16.

The register 29 holds the addition result of the adder 26 and feeds back the addition result to the adder 26 to add the numerical values arranged in time series.
Thereafter, as described above, the coefficient table 27 inputs the coefficient value G (z) corresponding to the division result z of the divider 24 to the image reproduction processing means 7.

In FIG. 5, when calculating each phase component, the processing can be performed with a single multiplier 23 and adder 26 by shifting the processing time.
In this case, the time required to calculate one coefficient becomes longer, but the apparatus scale can be reduced.

  In FIG. 5, only the phase tables 22a to 22d are shown, but each coefficient is calculated by preparing a table (not shown) corresponding to each phase component of the coefficient multiplication processes M1 to M5. be able to.

In the first and second embodiments, the processing of one SAR image 8 has been described. However, even when a plurality of SAR images 8 having different motion data 4 are processed due to a change in the position of the platform or the like. Can be processed.
That is, the coefficient calculation means 6 and 6A may perform processing by recalculating the numerical values stored in the phase tables 22a to 22d for each SAR image 8 having different motion data 4.

  In this case, it is necessary to recalculate the phase tables 22a to 22d, but the size of each table is at most an azimuth length (x) or a range length (y) size, and an image size (azimuth length × range length). Since it is sufficiently smaller than the above, the overall processing time is not greatly affected.

  In the first and second embodiments, the chirp scaling method has been described as an example. However, even with other methods, the phase of each coefficient is a variable of the coordinate x in the azimuth direction or the coordinate y in the range direction. If the phase component can be decomposed or sufficiently accurate approximation is possible, the processing can be performed in the same manner.

As described above, the coefficient calculating means 6A according to the second embodiment (FIG. 5) of the present invention calculates the phase component of the coefficient, and the phase table for the change in the range direction or the azimuth direction when transmitting the high-frequency pulse signal. 22a to 22d are created in advance, and then the final coefficient value G (z) is calculated by multiplying and adding the table data from the phase tables 22a to 22d of each component of the coefficient.
In addition, when the motion data 4 changes, the coefficient calculation means 6A recalculates the phase component of the coefficient and updates the phase tables 22a to 22d.

Thereby, even when the coefficient for image reproduction processing is calculated with a fixed-point number, a numerical value with high accuracy can be obtained with a small bit length.
Therefore, higher accuracy can be obtained with a smaller apparatus scale than when using a floating point number such as double precision, and the resolution of the output SAR image 8 can be improved.

In addition, since the configuration of the entire arithmetic device can be reduced, the weight of the device and the cost of the product can be reduced.
Further, since the phase calculation is a fixed-point number, the phase accuracy is almost constant and the phase accuracy in the lowest case is guaranteed, so that the lower limit of the resolution of the SAR image 8 to be output is guaranteed. An effect is obtained.

Embodiment 3 FIG.
In the second embodiment (FIG. 5), in order to suppress the device scale, the coefficient calculator 6A prepares only the phase tables 22a to 22d corresponding to the respective components and uses the common multiplier 23 and Although the divider 24 is provided, a plurality of multipliers 23a to 23d, shifters 25a to 25d, and adders 26a and 26b may be provided as shown in FIG.

FIG. 6 is a block diagram showing the coefficient calculation means 6B of the synthetic aperture radar apparatus according to Embodiment 3 of the present invention. Components similar to those described above (see FIGS. 4 and 5) are denoted by the same reference numerals. Detailed description is omitted.
The overall configuration of the synthetic aperture radar apparatus according to the third embodiment of the present invention is as shown in FIG. 1, and the basic processing operation is also as shown in FIGS.

  In FIG. 6, the coefficient calculation means 6B includes a parameter calculation means 21, phase tables 22a to 22d, multipliers 23a to 23d, a divider 24, shifters 25a to 25d, adders 26a and 26b, and a coefficient table. 27 and a selector 28.

The phase tables 22a and 22b hold the table data from the parameter calculation means 21 and input the table data corresponding to the azimuth coordinates x from the image reproduction processing means 7 to the individual multipliers 23a and 23b.
The phase tables 22c and 22d hold table data from the parameter calculation means 21 and input table data corresponding to the range coordinate y from the image reproduction processing means 7 to the individual multipliers 23c and 23d.

The multipliers 23a and 23b multiply the table data from the phase tables 22a and 22b and the range coordinate y, and input the multiplication results to the individual shifters 25a and 25b.
The multipliers 23c and 23d multiply the table data from the phase tables 22c and 22d and the azimuth coordinates x, and input the multiplication results to the individual shifters 25c and 25d.

Shifter 25a, 25b, and inputs the calculated value based on the instruction of the parameter calculator 21 (m multiplied by the remainder of division by ^{2 16)} to the adder 26a.
Shifter 25c, 25d inputs the calculated value based on the instruction of the parameter calculator 21 (m multiplied by the remainder of division by ^{2 16)} to the adder 26b.

The selector 28 selects and outputs one of the addition results of the adders 26a and 26b according to the coefficient required for the processing.
Thereafter, as described above, the coefficient table 27 inputs the coefficient value G (z) corresponding to the division result z of the divider 24 to the image reproduction processing means 7.

In this way, by preparing a plurality of arithmetic devices corresponding to each coefficient component, it is possible to perform coefficient arithmetic at high speed.
When calculating the coefficient multiplication processes M1 to M5, a multiplier, a shifter, and an adder may be configured in parallel according to each process. In this case, the coefficient can be calculated at high speed, although the apparatus scale becomes large.

  As described above, according to the third embodiment of the present invention, a high-precision numerical value can be obtained at high speed even when the coefficient for image reproduction processing is calculated with a fixed-point number. Compared to the second embodiment, an effect that the processing time of the image reproduction process can be shortened is obtained.

  1 SAR sensor, 2 received signal, 3 motion sensor, 4 motion data, 5 data holding means, 6, 6A coefficient calculating means, 7 image reproduction processing means, 8 SAR image, 11, 12 intermediate data, 21 parameter calculating means, 22a -22d phase table, 23, 23a, 23b multiplier, 24, 24a, 24b divider, 25, 25a-25d shifter, 26, 26a, 26b adder, 27 coefficient table, 28, 28a, 28b selector, 29 register, G (z) Coefficient value.

Claims (4)

A synthetic aperture radar device which is mounted on a movable platform and obtains a SAR image of the ground surface or the sea surface,
A SAR sensor that emits a high-frequency pulse signal, receives a reflection signal of the high-frequency pulse signal, and outputs a reception signal for obtaining the SAR image;
A movement sensor that measures the movement of the platform and outputs the instantaneous position of the platform at the time of transmission and reception of the high-frequency pulse signal as movement data;
Image reproduction processing means for performing image reproduction processing based on the received signal;
In obtaining the SAR image using the chirp scaling method based on the said motion data, and a coefficient calculating means for calculating a coefficient value to be used in the coefficient multiplication processing of the chirp scaling method,
The coefficient calculating means includes
Wherein based on a request of coefficients to be used in the coefficient multiplication processing from the image reproduction processing unit, when calculating the coefficient value corresponding to the request, y coordinate range direction at the time of transmission of the high frequency pulse signal, the azimuth direction Is the x coordinate, and the phase of the coefficient is expressed by an equation using the y coordinate and the x coordinate variable to divide the component into a minute variation component that varies slightly and an invariant component that does not change at all. Calculate the phase component of the coefficient,
After that, each phase component of the coefficient is digitized and then added, thereby calculating a coefficient value to be finally output . The coefficient calculating means includes
As a formula using the phase θ of the coefficient as a variable of the y coordinate and the x coordinate,

Express by
The synthetic aperture radar apparatus according to claim 1. The coefficient calculating means includes
When calculating the phase component of the coefficient, create in advance a phase table for changes in the range direction or azimuth direction at the time of transmission of the high-frequency pulse signal,
Then, by adding after combined digits table data from the phase table for each component of the coefficient of claim 1 or claim 2, characterized in that to calculate the coefficient values to be finally output Synthetic aperture radar device. The coefficient calculating means includes
The synthetic aperture radar device according to claim 3 , wherein when the motion data changes, the phase component of the coefficient is recalculated and the phase table is updated.

Images