JP5679904B2 - Image radar device - Google Patents

Description

  The present invention relates to an image radar device that acquires a radio wave image of a target that is an observation target by a radar, and particularly when the relative motion between the target and the image radar device is unknown or the estimation result of the relative motion includes an error. In order to compensate the image blur caused by the time variation of the distance and estimate the scaling factor when converting the image axis to the physical length axis, the generated image axis is physically The present invention relates to a technique for scaling to a long axis.

In the conventional image radar apparatus, the following processes (1) to (3) are repeatedly performed while changing the relative position between the radar and the target.
(1) A high-frequency signal transmission process in which a transmitter generates a high-frequency signal and outputs the high-frequency signal to a transmission / reception antenna via a transmission / reception switch, thereby radiating the high-frequency signal from the transmission / reception antenna toward a space (2) High-frequency signal reception processing for receiving a part of a high-frequency signal reflected by a target existing in space (3) The range-direction resolution of the received signal is range-compressed, and the amplitude characteristics of the received signal in the range direction Range profile acquisition processing to acquire a range profile

When the image radar apparatus obtains a range history that is a time history of the range profile by repeating the above-described series of processes while changing the relative position between the radar and the target, the image radar device generates a radar image from the range history.
Here, the range history is generally given as a two-dimensional distribution, and one axis of the two-dimensional distribution is the same as the range profile, and an elapsed time [s] (hereinafter, referred to as “high frequency signal irradiation”). (Referred to as “fast time”) or a range [m] obtained by multiplying the elapsed time by 1/2 of the radio wave speed.
The other axis of the two-dimensional distribution is given by an observation time (hereinafter referred to as “slow time” or “observation time”) or a hit representing an observation number while changing the relative positional relationship.

As typical radars for obtaining radar images, there are a synthetic aperture radar (SAR) and an inverse synthetic aperture radar (Inverse SAR: ISAR).
For example, the SAR has the effect of arranging a large-diameter antenna in the space by appropriately combining the range profile obtained by observing the observation target such as the ground surface while the radar moves in the space. The radar image with high resolution can be obtained also in the cross range direction orthogonal to the range.
By observing a target on which the radar moves, the ISAR obtains the same effect as the SAR, and obtains a radar image with a high resolution in the cross range direction.
For example, it can be easily understood that the same effect as the SAR can be obtained if the change of the radar position in the coordinate system fixed to the target is considered.

Relative motion is determined by its own motion, and relative motion is determined by the motion of the observation target, and the motion is relatively highly accurate compared to SAR, which allows the user to measure the motion with relatively high accuracy. In the ISAR, which is difficult to measure, a situation in which the relative motion is unknown or the estimation error becomes large is likely to occur.
However, even in the SAR, when the accuracy of the motion sensor mounted on the radar platform is low, the relative motion may increase.
From this point of view, in the following, SAR and ISAR are not particularly distinguished and will be referred to as image radar.
Note that imaging can also be performed between the SAR and the ISAR, that is, when both the radar and the target move (for example, when observing a ship navigating the sea from an aircraft). include.

In order to know the range position of the reflection point of the high-frequency signal at the target, it is only necessary to measure the range of the corresponding point on the range profile. However, in order to know the cross range position, distance change information is required, and thus, a range history is obtained by performing observations a plurality of times as described above.
For example, assuming a case in which a fixed target is observed from a radar that moves at a constant linear velocity, the distance decreases as the positions of the radar and the reflection point in the orbit direction become closer.
Therefore, by observing the change in distance, for example, the time when the position in the orbit direction becomes the same as that of the radar can be obtained as the observation time when the distance change becomes 0 (or the observation time when the distance becomes the smallest).
From this observation time and the radar speed, which is relative motion information, and the initial position of the radar, the position of the point in the orbit direction (corresponding to the cross range position) can be specified.

In actual radar imaging processing, it is difficult to directly measure the distance change of each reflection point from the received signal where the signals of multiple reflection points are superimposed, and specify the position in the direction orthogonal to the distance. This is implemented based on some signal processing.
Various imaging methods in the image radar have been proposed, and there are various models of relative motion for explaining the imaging method.
Here, for convenience of explanation, the change in the distance between the radar and the reflection point on the target due to relative motion is a combination of “change in the distance between the radar and the target center” and “rotational motion around the target center”. Adopt the model that is considered to be generated by.

First, considering a coordinate system fixed to a target, the radar moves while observing a stationary target in this coordinate system.
At this time, a compensation system is considered in which the distance between the target center and the radar is made constant, and a coordinate system that fixes the positions of the radar and the target center is considered.
In this coordinate system, the radar and the target center are stationary, and only the target rotates around the target center. Therefore, the distance change between the radar and each reflection point on the target is the distance between the radar and the target center. It can be considered that it was generated by the change in the angle and the rotational movement around the target center.
Note that the motion that causes a change in the distance between the radar and the target center may be referred to as a translational motion relative to a rotational motion, and this term will be used hereinafter.

If the distance between the target and the radar is sufficiently large compared to the size of the target, it can be considered that the influence of translational movement contributes equally to each reflection point. Therefore, this component is not only useful for specifying the position of each reflection point in the direction orthogonal to the distance direction, but also blurs the radar image in the range direction. In addition, depending on the component, the cross range direction may be blurred.
Therefore, it is necessary to correctly estimate and compensate for this component. This compensation process is referred to as translational motion compensation.

Translational compensation causes blurring in the range axis direction.
Estimate the range movement of each reflection point exceeding the range resolution, and estimate the range compensation to compensate for the range movement and the “secondary or more changes with respect to time of phase” that cause blur in the Doppler axis direction , It is roughly divided into phase compensation for compensating for the second-order and higher-order changes.

また、距離変化に相当する＋ｘ方向のラジアル速度Ｖ_P［ｍ／ｓ］は、下記の式（２）で与えられる。

Next, consider the case where the effect of translational motion is fully compensated.
Assume that the rotation center is at the origin of the Cartesian coordinate system and the target rotating at the angular velocity ω [rad / s] around the z-axis is observed from a sufficiently far position in the −x direction.
In this case, the distance L _P [m] based on the center of rotation of the point P given by [r _P cos θ _P , r _P sin θ _P , z] at a certain assumed instantaneous time is expressed by the following equation (1). Given in.

Further, the radial velocity V _P [m / s] in the + x direction corresponding to the distance change is given by the following equation (2).

If the x coordinate of the point P is x _P and the y coordinate is y _P , these are expressed as r _P cos θ _P and r _P sin θ _P.
Therefore, the two-dimensional distribution of the distance-radial speed of each reflection point represents the target reflection point position projected on the xy plane, excluding expansion and contraction of −ω times in the y-axis direction.
For this reason, if the angular velocity ω of the rotational motion that is the relative motion remaining after the translational motion compensation is known, the two-dimensional distribution of the distance-radial velocity is expanded / contracted by −1 / ω times in the y-axis direction, thereby The position projected on the xy plane can be specified.

In general radar, a Doppler frequency is used to obtain a radial velocity. This is obtained by Fourier transforming the range history after translational motion compensation in the hit direction for each range cell.
Assuming that the Doppler frequency of the point P on the range Doppler frequency image is F _P [Hz], F _P is expressed by the following equation (3), where the transmission wavelength is λ [m].

That is, the range Doppler image also represents the position of the target reflection point projected on the xy plane, excluding expansion and contraction in the y-axis direction.
In order to obtain an image in which the expansion and contraction in the y-axis direction is correct, the above-described angular velocity ω is required except for the known λ. And this value is generally unknown in situations where the relative motion is unknown.
In the general case where the rotation axis is not orthogonal to the range axis as in the above z-axis, the latter when the rotational motion is separated into rotation around the range axis and rotation around the axis orthogonal to the range. The angular velocity of the component is required. Since the rotation around the range axis does not contribute to the distance change, the radar image is not affected.

The unit vector _ic bold in the cross range axis direction is determined by the vector calculation of the following equation (4), where the unit vector in the range axis direction is _ir bold and the rotational angular velocity vector is ω bold. Since i _c , i _r , and ω represent vectors, i _c , i _r , and ω are represented in bold in Equation (4). Since bold characters cannot be used, they are represented as “ _ic bold”, “ _ir bold”, and “ω bold”.

The problem of scaling the Doppler frequency axis of a range Doppler image to the cross range axis, which is the physical length axis, is called the cross range scaling problem.
In the following, processing for improving the resolution in the cross range direction by signal processing, such as Fourier transform in the observation time direction, is referred to as cross range compression.

式（５）において、Δθ［ｒａｄ］は観測中の総回転角である。 Since the Doppler frequency resolution of the image is given by 1 / T [Hz] when the observation time is T [s], the cross range resolution Δc [m] converted to the cross range is expressed by the following equation (5). become that way.

In equation (5), Δθ [rad] is the total rotation angle under observation.

In principle, the greater the total rotation angle Δθ of the target under observation, the better the cross range resolution.
However, as easily understood from the equations (1) and (2) showing the distance L _P and the radial velocity V _P , each reflection point has its range and radial velocity (in other words, the Doppler frequency as the rotation angle changes). As the total rotation angle Δθ increases, each reflection point moves beyond the resolution cell of the range and the Doppler frequency on the image.
That is, in the simple Fourier transform process as described above, when the total rotation angle Δθ is large, the resolution may be deteriorated.

As one of image reproduction methods for solving this problem, there is a polar format method (hereinafter referred to as “conventional method 1”) described in Non-Patent Document 1 below.
In the conventional method 1, the range profile in each hit after translational motion compensation is further compensated so that the range of the center of rotation is completely zero, and then the frequency distribution (Fourth transform obtained by Fourier transforming the range profile) When the axis of the range profile is regarded as the range [m], the spatial frequency distribution is arranged on the frequency plane at the position corresponding to the transmission band and the target rotation angle in each observation.

Then, by re-sampling the data on a rectangular grid and performing inverse two-dimensional Fourier transform, the above-mentioned blur due to rotation is eliminated, and an image in which the cross range direction is also correctly scaled to a physical length is obtained. (If the original is a frequency distribution, the two axes of the resulting image will be time, but this can be easily converted to physical length by multiplying by a known factor (light speed / 2)) Is).
However, the conventional method 1 has the following problems.
(I) It is necessary to know the amount of change in the rotation angle in each observation (if the rotation angular velocity ω is constant) in order to realize the arrangement of the coordinate coordinates in the frequency plane. If not, or if the relative motion estimation error is large, there is a problem that it is difficult to realize.
(Ii) In addition to the necessary level of translational motion compensation, compensation that keeps the distance between the radar and the target center at 0 is necessary. However, when there is no information on relative motion, When the estimation error is large, there is a problem that it is difficult to realize.

Another image reproduction method includes a method described in the following Non-Patent Document 2 (hereinafter referred to as “conventional method 2”).
In this method, in addition to normal translational motion compensation, it is assumed that compensation has already been made in some way so that the change in distance between the radar and the target center is completely zero.
Then, under the assumption that the target equivalent rotational acceleration vector is constant during observation, the blur generated by the rotational motion is compensated in two stages in the order of the range axis direction and the Doppler axis direction.

In the first stage, it is used that the range change during observation of each reflection point on the equal rotation target has the following properties in the order of the range resolution.
(A) The coefficient of the primary change that can be approximately regarded as the primary change with respect to the hit is the coefficient of the primary change of the Doppler cell at each reflection point when the Doppler frequency at the center of rotation is 0, regardless of the magnitude of the rotational angular velocity. Approximately determined from only the position and the known coefficient In the conventional method 2, the first order change of the coefficient determined by the cell position is compensated for each blurred Doppler cell. Thereby, the blur in the range axis direction of the image is approximately eliminated.

In the second stage, the fact that “the phase change of the received signal caused by the change in the distance of each reflection point”, which causes the blur in the Doppler axis direction of the image, has the following property is used.
(A) The magnitude of the coefficient of the secondary change that can be regarded as a secondary change with respect to the hit is approximately proportional to each range position when the center of rotation is the 0 range.

In the conventional method 2, the PD (Phase Difference) method, which is one of the general methods for estimating the secondary phase change caused by the secondary distance change in translational motion compensation, is applied to the data of each range cell. Then, the secondary change coefficient of the phase is estimated for each range cell.
Next, from these estimation results, taking into account that the magnitude of the coefficient is proportional to the range position, the phase secondary coefficient in each range is estimated while smoothing out outliers.
Then, based on the quadratic coefficient of the phase in each range, compensation is performed to cancel the unnecessary phase change, and then imaging is performed again.

With the above processing, in the conventional method 2, it is possible to approximately compensate for the blurring of the image caused by the influence of the rotational motion without the need to estimate the rotational angular velocity.
However, the conventional method 2 has the following problems.
(I) Since cross range scaling cannot be realized, there is a problem that the physical position of each reflection point in the range direction cannot be obtained. (Ii) Approximate processing is performed for rotation compensation. There is a problem that an error occurs. (Iii) In normal translational motion compensation, it is assumed that translational motion compensation including compensation that completely eliminates the first-order change in the range not covered by compensation has been made. Since the implementation method is not clear, there is a problem that is difficult to implement

Jakowatz Jr., C.V., Wahl, D.E., Eichel, P.H., Ghiglia, D.C., and Thompson, P.A. “Spotlight-mode synthetic aperture radar: a signal processing approach,” Kluwer Achademic Publishers, 1999. Munoz-Ferreras JM, and Perez-Martfnez F., “Uniform rotational motion compensation for inverse synthetic aperture radar with non-cooperative targets,” IET Radar Sonar Navig., Vol. 2, No. 1, pp. 25-34, Feb. . 2008.

  Since the conventional image radar apparatus is configured as described above, if the relative motion between the target and the radar is unknown or the estimation error of the relative motion is large, the image generated by the rotational motion The blur cannot be compensated with high accuracy. In addition, since cross range scaling cannot be realized, there is a problem that the physical position of each reflection point in the range direction cannot be obtained.

  The present invention has been made to solve the above-described problems. Even when the relative motion between the target and the radar is unknown or the estimation error of the relative motion is large, an image generated by the rotational motion is provided. An object of the present invention is to obtain an image radar apparatus that can accurately compensate for blur and can perform cross-range scaling to obtain the physical position of each reflection point in the range direction.

  The image radar apparatus according to the present invention radiates a high-frequency signal toward a space, receives a part of the high-frequency signal reflected by a target existing in the space, and sets the resolution in the range direction of the received signal as a range. The range history, which is the time history of the range profile, is obtained by repeating a series of processing to obtain a range profile that is the amplitude characteristics of the received signal in the range direction while changing the relative position between the radar and the target. Range history acquisition circuit, range history acquired by range history acquisition circuit, translation motion compensation circuit that compensates for unnecessary distance change between radar and target caused by unnecessary translation motion, and distance by translation motion compensation circuit Range history compensated for changes, generated by rotational motion at each reflection point of the high-frequency signal at the target The phase term of the data string in the hit direction of each range is identified in the range history in which the range direction blur is compensated by the rotation range cell movement compensation circuit and the rotation range cell movement compensation circuit. The probability of a plurality of candidate values that are candidates for the secondary phase coefficient that is the coefficient of the secondary phase change included in the phase term is evaluated for each range cell, and the distribution in the range direction of the evaluation result of each candidate value is calculated. This is a change coefficient with respect to the range of the secondary phase coefficient based on the output secondary phase coefficient candidate evaluation circuit and the secondary phase coefficient evaluation value distribution which is the distribution in the range direction output from the secondary phase coefficient candidate evaluation circuit. Based on the secondary phase coefficient change amount estimation circuit for estimating the secondary phase coefficient range change coefficient and the secondary phase coefficient range change coefficient estimated by the secondary phase coefficient change amount estimation circuit, The second-order phase coefficient is calculated from the second-order phase coefficient range change coefficient estimated by the second-order phase coefficient change amount estimation circuit and the second-order phase coefficient change value estimation circuit, and the second-order phase coefficient change value estimation circuit. A zero-crossing range estimation circuit that estimates a range that is 0 and a range history that is compensated for a distance change by a translational motion compensation circuit so that the range estimated by the zero-crossing range estimation circuit is 0 And a radar image generation circuit that generates a radar image from the range history after compensation by the zero-order range compensation circuit based on the angular velocity estimated by the rotation angular velocity conversion circuit. .

  According to the present invention, a high-frequency signal is emitted toward a space, while a part of the high-frequency signal reflected by a target existing in the space is received, and the resolution in the range direction of the received signal is subjected to range compression. The range history that is the time history of the range profile is obtained by repeating a series of processing to acquire the range profile that is the amplitude characteristic of the received signal in the range direction while changing the relative position between the radar and the target. The range history acquired by the acquisition circuit, the range history acquisition circuit, the translational motion compensation circuit that compensates for the unnecessary distance change between the radar and the target caused by the unnecessary translational motion, and the distance change compensated by the translational motion compensation circuit In the range direction generated by the rotational motion at each reflection point of the high-frequency signal at the target. The phase term of the data string in the hit direction of each range is identified and included in the phase term in the range history in which the blur in the range direction is compensated by the rotational range cell movement compensation circuit A second order that evaluates the probability of a plurality of candidate values that are candidates for a second order phase coefficient that is a second order phase change coefficient for each range cell, and outputs a distribution in the range direction of the evaluation result of each candidate value Based on the phase coefficient candidate evaluation circuit and the secondary phase coefficient evaluation value distribution that is the distribution in the range direction output from the secondary phase coefficient candidate evaluation circuit, the secondary phase coefficient that is a change coefficient for the range of the secondary phase coefficient Based on the secondary phase coefficient variation estimation circuit for estimating the range variation coefficient and the secondary phase coefficient range variation coefficient estimated by the secondary phase coefficient variation estimation circuit, The secondary phase coefficient is zero based on the rotational angular velocity conversion circuit that estimates the angular velocity of motion, the secondary phase coefficient range variation coefficient estimated by the secondary phase coefficient variation estimation circuit, and the secondary phase coefficient evaluation value distribution. A zero-crossing range estimation circuit that estimates a range, and a zero-order range compensation circuit that compensates for a range history in which a distance change is compensated by a translational motion compensation circuit so that the range estimated by the zero-crossing range estimation circuit becomes zero And the radar image generation circuit is configured to generate a radar image from the range history after compensation by the zero-order range compensation circuit based on the angular velocity estimated by the rotational angular velocity conversion circuit. Even if the relative motion of the image is unknown or the estimation error of the relative motion is large, the image blur caused by the rotational motion is accurately compensated. In addition, it is possible to obtain the physical position of each reflection point in the range direction by performing cross range scaling.

It is a block diagram which shows the image radar apparatus by Embodiment 1 of this invention. It is a block diagram which shows the range history acquisition circuit 1 of the image radar apparatus by Embodiment 1 of this invention. It is a block diagram which shows the translational motion compensation circuit 2 of the image radar apparatus by Embodiment 1 of this invention. It is a block diagram which shows the secondary phase coefficient candidate evaluation circuit 4 for every range cell of the image radar apparatus by Embodiment 1 of this invention. It is a block diagram which shows the secondary phase coefficient variation | change_quantity estimation circuit 5 of the image radar apparatus by Embodiment 1 of this invention. FIG. 6 is a configuration diagram showing the inside of a spectral image utilization type trajectory inclination specifying circuit 61 in the secondary phase coefficient variation estimation circuit 5; It is a block diagram which shows the zero crossing range estimation circuit 7 in consideration of the secondary phase coefficient of the image radar apparatus according to Embodiment 1 of the present invention. It is a schematic diagram which shows an example of secondary phase coefficient evaluation value distribution. It is explanatory drawing which shows the processing content of the locus | trajectory inclination reliability distribution calculation circuit 71. FIG. It is explanatory drawing which shows the processing content of the locus | trajectory inclination reliability distribution calculation circuit 71 and the reliability distribution maximum position specifying device 72. FIG. It is a block diagram which shows the image radar apparatus by Embodiment 2 of this invention. It is a block diagram which shows the correction type | mold zero crossing range estimation circuit 10 before secondary phase coefficient consideration of the image radar apparatus by Embodiment 2 of this invention. It is a block diagram which shows the image radar apparatus by Embodiment 3 of this invention. It is a block diagram which shows the image radar apparatus by Embodiment 4 of this invention. It is a block diagram which shows the image radar apparatus by Embodiment 5 of this invention. It is a block diagram which shows the image radar apparatus by Embodiment 6 of this invention. It is a block diagram which shows the image radar apparatus by Embodiment 6 of this invention.

Embodiment 1 FIG.
1 is a block diagram showing an image radar apparatus according to Embodiment 1 of the present invention.
In FIG. 1, a range history acquisition circuit 1 radiates a high-frequency signal toward a space, receives a part of the high-frequency signal reflected by a target existing in the space, and determines the resolution in the range direction of the received signal. The range history, which is the time history of the range profile, is obtained by repeating a series of processing that compresses the range and obtains the range profile that is the amplitude characteristics of the received signal in the range direction while changing the relative position between the radar and the target. It is a circuit to acquire.

The translational motion compensation circuit 2 is a range history acquired by the range history acquisition circuit 1 and is a circuit that compensates for an unnecessary change in the distance between the radar and the target caused by unnecessary translational motion.
The rotational range cell movement compensation circuit 3 is a range history in which the change in distance is compensated by the translational motion compensation circuit 2, and is a circuit that compensates for blur in the range direction caused by rotational motion at each reflection point of the high-frequency signal at the target.

The secondary phase coefficient candidate evaluation circuit 4 for each range cell specifies the phase term of the data string in the hit direction of each range in the range history in which the blur in the range direction is compensated by the rotation range cell movement compensation circuit 3, and the phase term The probability of a plurality of candidate values that are candidates for a secondary phase coefficient that is a coefficient of a secondary phase change included in each range cell (or each sub-range cell into which the range resolution cell is further divided) It is a circuit that evaluates and outputs a secondary phase coefficient evaluation value distribution that is a distribution in the range direction of the evaluation result of each candidate value.
The secondary phase coefficient variation estimation circuit 5 is a secondary phase coefficient that is a variation coefficient for the range of the secondary phase coefficient based on the secondary phase coefficient evaluation value distribution output from the secondary phase coefficient candidate evaluation circuit 4 for each range cell. It is a circuit for estimating a range change coefficient.

The rotational angular velocity conversion circuit 6 is a circuit that estimates the target equivalent angular velocity of the rotational motion based on the secondary phase coefficient range variation coefficient estimated by the secondary phase coefficient variation estimation circuit 5.
The zero-crossing range estimation circuit 7 in consideration of the secondary phase coefficient 7 has the secondary phase coefficient range change coefficient estimated by the secondary phase coefficient change amount estimation circuit 5 and the secondary phase output from the secondary phase coefficient candidate evaluation circuit 4 for each range cell. This is a circuit that estimates the range where the secondary phase coefficient is 0 from the coefficient evaluation value distribution.

The zero-order range compensation circuit 8 is a circuit that compensates the range history in which the distance change is compensated by the translational motion compensation circuit 2 so that the range estimated by the zero-crossing range estimation circuit 7 in consideration of the secondary phase coefficient becomes zero. .
The rotation-considering radar image generation circuit 9 performs image reproduction in consideration of rotation based on the angular velocity estimated by the rotation angular velocity conversion circuit 6, thereby obtaining both images from the range history after compensation by the zero-order range compensation circuit 8. A circuit that generates a radar image whose axis is scaled to a physical length axis.

FIG. 2 is a block diagram showing the range history acquisition circuit 1 of the image radar apparatus according to Embodiment 1 of the present invention.
In FIG. 2, the transmitter 31 generates a high frequency signal and outputs the high frequency signal to the transmission / reception switch 32.
The transmission / reception switching device 32 is a switching device that switches the direction of signal flow between transmission and reception of a high-frequency signal.
The transmission / reception antenna 33 radiates the high-frequency signal output from the transmission / reception switch 32 toward the space, while receiving a part of the high-frequency signal reflected by the target existing in the space and switching the reception signal between transmission / reception. Output to the device 32.

The receiver 34 amplifies the received signal output from the transmission / reception switch 32 and detects it.
The range compressor 35 performs range compression on the resolution in the range direction of the received signal detected by the receiver 34 based on the transmission waveform of the high-frequency signal, and acquires a range profile that is an amplitude characteristic of the received signal in the range direction. Perform the process.
A series of processing in the transmitter 31, the transmission / reception switch 32, the transmission / reception antenna 33, the receiver 34, and the range compressor 35 is repeated while changing the relative position between the radar and the target, so that the time history of the range profile is obtained. A range history is acquired.

FIG. 3 is a block diagram showing the translational motion compensation circuit 2 of the image radar apparatus according to Embodiment 1 of the present invention.
In FIG. 3, a general translational motion compensation circuit 41 is a range history acquired by the range history acquisition circuit 1, and estimates the range movement of each reflection point exceeding the range resolution and the second or higher order change with respect to the phase time. Thus, the circuit compensates for the range shift and the second and higher order changes.
The primary phase change compensator 42 performs processing for compensating for the primary change in phase in the range history after compensation by the general translational motion compensation circuit 41.

FIG. 4 is a block diagram showing the secondary phase coefficient candidate evaluation circuit 4 for each range cell of the image radar apparatus according to Embodiment 1 of the present invention.
In FIG. 4, the range interpolator 51 interpolates the range history compensated for the blur in the range direction by the rotation range cell movement compensation circuit 3 in the range direction, and evaluates the range history after the interpolation in the range direction as a secondary phase coefficient candidate evaluation for each range cell. A process of outputting to the index calculation circuit 52 is performed.
The secondary phase coefficient candidate evaluation index calculation circuit 52 for each range cell is a candidate for the secondary phase coefficient indicating the secondary phase change added to the phase of the signal of the range for each range cell (or for each sub-range cell). This is a circuit that evaluates the probabilities of a plurality of candidate values and calculates a secondary phase coefficient evaluation value distribution in which the probabilities of the plurality of candidate values are digitized.

FIG. 5 is a block diagram showing the secondary phase coefficient variation estimation circuit 5 of the image radar apparatus according to Embodiment 1 of the present invention.
In FIG. 5, the spectral image utilization type trajectory inclination specifying circuit 61 performs two-dimensional Fourier transform on the amplitude distribution of the two-dimensional image indicated by the secondary phase coefficient evaluation value distribution output from the secondary phase coefficient candidate evaluation circuit 4 for each range cell. This is a circuit that obtains a spectrum image and estimates the inclination of a straight line on the two-dimensional image from the spectrum image.
The secondary phase coefficient change amount calculation circuit 62 is a circuit that calculates a secondary phase coefficient range change coefficient from the slope of the straight line estimated by the spectral image utilizing type trajectory slope specifying circuit 61.

FIG. 6 is a block diagram showing the inside of the spectrum image utilization type trajectory slope specifying circuit 61 in the secondary phase coefficient variation estimation circuit 5.
In FIG. 6, the trajectory slope reliability distribution calculation circuit 71 performs weighting of the amplitude distribution of the two-dimensional image indicated by the secondary phase coefficient evaluation value distribution output from the secondary phase coefficient candidate evaluation circuit 4 for each range cell. A spectral image is obtained by performing a two-dimensional Fourier transform on the subsequent amplitude distribution, and a plurality of linear integration paths having different slopes are set on the spectrum image through the origin corresponding to the DC component, and along the integration path In this circuit, the result of line integration of the amplitude distribution of the weighted spectrum image is output as the reliability with respect to the slope of each integration path.

The reliability distribution maximum position locator 72 performs a process of detecting the inclination of the integration path with the maximum reliability among the reliability with respect to the inclination of each integration path output from the locus inclination reliability distribution calculation circuit 71.
The trajectory slope converter 73 calculates the slope of the integration path having the maximum reliability detected by the reliability distribution maximum position locator 72, and the secondary phase coefficient evaluation value distribution output from the secondary phase coefficient candidate evaluation circuit 4 for each range cell. The process which converts into the inclination of the locus | trajectory on the two-dimensional image which is shown in FIG.

FIG. 7 is a block diagram showing the zero-crossing range estimation circuit 7 in consideration of the secondary phase coefficient of the image radar apparatus according to Embodiment 1 of the present invention.
In FIG. 7, the fixed slope integral type zero-crossing range estimation circuit 81 has a straight line inclination a estimated by the spectral image utilization type locus inclination specifying circuit 61, and one axis (for example, the vertical axis) set in advance. A zero crossing point that is a coordinate (x0 in the case of y = a (x + x0)) on another axis (for example, the horizontal axis x) when passing through a position where the coordinate along y) is 0 is estimated. Therefore, on the two-dimensional image indicated by the secondary phase coefficient evaluation value distribution output from the secondary phase coefficient candidate evaluation circuit 4 for each range cell, a plurality of integration paths having different slopes of a and x0 are set. This is a circuit that performs line integration along each integration path, and estimates that the candidate x0 having the maximum value of the line integration is in the range where the secondary phase coefficient is zero.

Next, the operation will be described.
First, the range history acquisition circuit 1 acquires the range history that is the time history of the range profile by repeating the following processes (1) to (3) while changing the relative position between the radar and the target.
(1) A transmission process in which the transmitter 31 generates a high-frequency signal and outputs the high-frequency signal to the transmission / reception antenna 33 via the transmission / reception switch 32, thereby radiating the high-frequency signal from the transmission / reception antenna 33 toward the space ( 2) Reception processing in which the transmission / reception antenna 33 receives a part of the high-frequency signal reflected by the target existing in space, and the receiver 34 detects the received signal. (3) The range compressor 35 is detected by the receiver 4. Range profile acquisition processing that compresses the resolution in the range direction of the received signal and obtains a range profile that is the amplitude characteristic of the received signal in the range direction

When the range history acquisition circuit 1 acquires the range history, the translational motion compensation circuit 2 compensates for an unnecessary change in the distance between the radar and the target generated by the unnecessary translational motion.
Specifically, an unnecessary distance change between the radar and the target caused by unnecessary translational motion is compensated as follows.
The general translation motion compensation circuit 41 of the translation motion compensation circuit 2 applies each of the general translation motion compensation processes so that each reflection point that exceeds the range resolution in the range history acquired by the range history acquisition circuit 1. The range shift and the second or higher order change with respect to the phase time are estimated to compensate for the range shift and the second or higher order change.
Various methods have been proposed as a general translational motion compensation process. However, in the general translational motion compensation process, the primary change of the phase in the range history is not targeted for compensation.

The primary phase change compensator 42 compensates for the primary change in phase in the range history after compensation by the general translational motion compensation circuit 41.
Hereinafter, the compensation process of the primary phase change by the primary phase change compensator 42 will be described in detail.
As shown below, the primary phase change compensator 42 approximates the range change estimated by the range compensation of the general translational motion compensation circuit 41 with a linear function with respect to the hit, and performs a general phase change corresponding to the general phase change. Subtract from the range history after translational compensation.

Here, the total number of hits is H, the hit number is h (h = −H / 2, −H / 2 + 1,..., H / 2-1), and the pulse repetition period is Δt [s]. The observation time t (h) [s] at the h-th hit is given by the following equation (6).
The total hit number H may be an even number or an odd number. In the case of an odd number, the hit number is set as h = − (H−1) / 2,..., (H−1) / 2.

The number of range cells is M, the range cell number is m (m = 0, 1,..., M−1), and the range resolution (range pixel spacing described later) is Δr [m].
In addition, a general range history after translational motion compensation is represented by s ₁ (h, m), and the slope of the linear function of the range change (range cell change amount per hit) is a ₁ [cell].
In this case, if the center frequency is f ₀ [Hz] and the speed of light is c [m / s], the phase change φ ₁ (h) [rad] corresponding to this range change is given by the following equation (7). .

The primary phase change compensator 42 uses the relationship of the equations (6) and (7) to obtain the phase change φ ₁ (h) that is not compensated for by the general translational motion compensation process using the following equation ( Compensate as in 8) to obtain the compensated range history s ₂ (h, m).

As described above, each reflection point on the target undergoes a range change according to its position under the influence of rotation.
The range change can be regarded as approximately the same as a primary change with respect to time in a situation where the angular velocity is constant and the angle change range is relatively small, and can be regarded as the same if the cross-range position is the same. (It is easy to understand that points with the same cross-range position have the same Doppler frequency (or the same radial speed)).
The processing of the primary phase change compensator 42 is intended to set the Doppler frequency at the cross range position where the range change becomes 0 to 0 [Hz] by general translational motion compensation.

The rotational range cell movement compensation circuit 3 compensates for the component blur in the range direction among the blurs generated by the rotational motion.
Hereinafter, the blur compensation process in the range direction by the rotation range cell movement compensation circuit 3 will be described in detail.

ここで、cosθ(t)，sinθ(t)は次式で表すことができる。

Here, if the rotation angle from time t = 0 generated by the rotational motion is θ (t), the rotation angle θ at the point where the range and the cross range at the time t = 0 are x _P and y _P. The range r _P (θ (t)) at (t) is given by the following equation (9). However, r ₀ is the range of the rotation center of the post-translational motion compensation.

Here, cos θ (t) and sin θ (t) can be expressed by the following equations.

よって、クロスレンジ位置ｙ_P毎に異なる１次のレンジ変化を補償してやればよいが、実際は、角速度ωとクロスレンジ位置ｙ_Pは不明である。 Here, ignoring the second and higher order terms and setting θ (t) = ωt, the range change δr (y _P , t) from t = 0 of the point of the cross range position y _P can be expressed by the following formula ( 12).

Therefore, it is only necessary to compensate for a different primary range change for each cross range position y _P , but in reality, the angular velocity ω and the cross range position y _P are unknown.

Therefore, consider a Doppler cell whose cell number is given by d (d = −H / 2, −H / 2-1,..., H / 2-1: d = 0 is Doppler frequency 0).
The cross range position (corresponding to y _P described above) of the center point of this cell is given by dΔc when the cross range resolution Δc is used.
Therefore, the range change at this point is dΔcωt.
Incidentally, as described above, in consideration of the fact that Δc = λ / (2ωT) is given, the following equation (13) is obtained.

That is, for the center point of the d-th Doppler cell, the range change is given by equation (13) regardless of the angular velocity ω.
Here, instead of the time t, the hit h is used, and δr (dΔc, t) in the equation (13) is expressed by the range change δr (d, h) of the hit h with respect to the Doppler cell number d. Equation (14) is obtained.

Based on the above, the rotational range cell movement compensation circuit 3 generates a range Doppler image S ₂ (d, m) by Fourier-transforming the range history s ₂ (h, m) after translational motion compensation in the hit direction. Compensation is performed so as to cancel the range change of each hit given by the equation (14) for each Doppler cell d.
Nature Specifically, in compensating for reflection points d _c doppler cell, the range history s ₂ (h, m) is passed through a signal near d _c doppler cells, preventing the passage of other signals The appropriate weighting factor W _dc (d, m) is introduced, and the weighted image S tilde _dc (d, m) of the following equation (15) (in the specification document, the upper part of S because of inability to represent "-", while "S tilde" and hereinafter) to get (a matter of course, when d _c doppler cells near the end of the image in consideration of the aliasing, pass Just set the bandwidth).

Rotation range cell movement compensation circuit 3, the weighted image S tilde _dc (d, m) obtains the, the weighted image S tilde _dc (d, m) and inverse Fourier transform on the Doppler cell direction, in the vicinity of d _c doppler cell The range history s tilde _dc (h, m) regarding the reflection point is acquired.
Then, the rotation range cell movement compensation circuit 3, the formula (14) of d [delta] r obtained by substituting the d _c to (d _c, h) Range history s to cancel the tilde _dc (h, m) compensation Then, the compensated range history s tilde _dc (h, m) is Fourier transformed in the hit direction to obtain a range Doppler image.

The rotation range cell movement compensation circuit 3 stores the entire range data d _c doppler cell in the range Doppler image, separately prepared to have sequences S ₃ (d, m) the cells of d = d _c in.
Note that the array S ₃ (d, m) obtained by performing the same processing for all _dc Doppler cells corresponds to the range Doppler image after the range movement compensation by the rotational motion.
By subjecting the array S ₃ (d, m) to inverse Fourier transform in the Doppler cell direction, the range history s ₃ (h, m) after the range movement compensation by the rotational motion is obtained.

In the processing of the rotation range cell movement compensation circuit 3, it is expected that the rotation center is located at least in the zero Doppler cell.
From this point of view, in the translational motion compensation circuit 2 in the previous stage, this cell is arranged in the zero Doppler (cross range) under the idea that the center of rotation exists at the Doppler cell (cross range) position where the range change is zero. Therefore, the first-order phase change is compensated.

先に説明した回転レンジセル移動補償回路３では、式（１６）に現れる三角関数を角度変化に対する１次以下の多項式で表しているが、ここでは、レンジセルの変化より感度の高い位相変化に着目することを考慮して、これより次数の高い２次以下の多項式で表すことにする。
この場合、式（１６）の位相変化φ_P（ｔ）は、下記の式（１７）で近似される。

Here, the phase change φ _P (t) of the point P is obtained by using the value of θ (t) = ωt in r _P (θ (t)) of the equation (9). Can be expressed as:

In the rotational range cell movement compensation circuit 3 described above, the trigonometric function appearing in the equation (16) is expressed by a first-order polynomial or less with respect to the angle change. Here, attention is paid to the phase change having higher sensitivity than the change of the range cell. In view of this, it is expressed by a second-order polynomial having a higher order than this.
In this case, the phase change φ _P (t) in the equation (16) is approximated by the following equation (17).

In Expression (17), the second-order phase change 2πx _P ω ² t ² / λ corresponds to a main component that causes the range Doppler image to be blurred in the Doppler axis direction by the rotational motion.
In consideration of the fact that the secondary phase coefficient is proportional to the range x _P based on the rotation center range, the following equation (18) is used here, where the secondary phase coefficient is a function α ₂ (x _P ) of the range x _P. ).

As apparent from the equation (18), if a set of x _P and α ₂ (x _P ) is obtained, | ω | can be determined.
Further, even when the secondary phase coefficient range change coefficient β _2, which is the ratio of the change to the range of the secondary phase coefficient expressed by the following equation (19), is obtained, | ω | it can.

When estimating | ω |, when using a set of x _P and α ₂ (x _P ), it is necessary to know the range x _P. In other words, while it is necessary to specify the range of the rotation center, when the secondary phase coefficient range change coefficient is used, x _P itself is not required, and therefore, the range of the rotation center is not required to be estimated. Therefore, it is considered that the estimation becomes easy.
Based on these viewpoints, the secondary phase coefficient change amount estimation circuit 5 estimates the secondary phase coefficient range change coefficient β ₂ , and the rotational angular velocity conversion circuit 6 sets the secondary phase coefficient range change coefficient β ₂ to | ω | Convert to.
In order to enable the estimation of the secondary phase coefficient range change coefficient β ₂ , the secondary phase coefficient candidate evaluation circuit 4 for each range cell evaluates the likelihood of each candidate value of the secondary phase coefficient in each range. A second-order phase coefficient evaluation value distribution that is a distribution of (1) is generated (see FIG. 8).

That is, the secondary phase coefficient candidate evaluation circuit 4 for each range cell specifies the phase term of the data string in the hit direction of each range in the range history in which the blur in the range direction has been compensated by the rotation range cell movement compensation circuit 3. The probability of a plurality of candidate values that are candidates for the secondary phase coefficient that is the coefficient of the secondary phase change included in the phase term is determined for each range cell (or the sub-range cell obtained by further dividing the range resolution cell) And a secondary phase coefficient evaluation value distribution that is a distribution in the range direction of the evaluation result of each candidate value is output.
The processing contents of the secondary phase coefficient candidate evaluation circuit 4 for each range cell will be specifically described below.

The range interpolator 51 of the secondary phase coefficient candidate evaluation circuit 4 for each range cell interpolates the range history compensated for the blur in the range direction by the rotation range cell movement compensation circuit 3 in the range direction, and the interpolated range history in the range direction. This is output to the secondary phase coefficient candidate evaluation index calculation circuit 52 for each range cell.
For interpolation in the range direction of the range history, general zero padding interpolation or the like can be used.
By this processing, the cell width in the range direction of the range history (hereinafter referred to as “range pixel spacing”) becomes smaller than the original range pixel spacing Δr, and the number of cells also increases. The interpolated range cell is also called a range cell as before.
The range pixel spacing after the range interpolation is Δr _s [m], and the number of range cells after the interpolation is M _s .

In subsequent processing, originally, it is necessary to handle the characteristics for range position x _P of the reflection points, when substitute characteristic for the center range of each range cell so, by incorporating a range interpolator 51, the reflection point The difference between the true range and the center range of the range cell to which the reflection point belongs can be reduced to reduce the influence of errors based on this.

When it is determined that the pixel spacing of the range is sufficiently small before the interpolation, the processing of the range interpolator 51 can be omitted.
In this case, the range interpolator 51 is actually unnecessary, but in this case as well, it can be considered that the interpolation has been performed by a factor of 1.
In consideration of this point, in the following, the contents of the processing will be explained in a unified manner including the case where range interpolation is not performed for the sake of efficiency of explanation.

The secondary phase coefficient candidate evaluation index calculation circuit 52 for each range cell is provided with a secondary phase coefficient candidate indicating a secondary phase change added to the phase of the signal of the range for each range cell (or for each sub-range cell). The probability of the plurality of candidate values is evaluated, and a secondary phase coefficient evaluation value distribution in which the probability of the plurality of candidate values is quantified is calculated.
The estimation problem handled here is closely related to the problem handled in general translational motion compensation.
For example, in the SAR, it is possible to realize the compensation of about the range resolution by using the information of the motion sensor mounted on the radar platform, but the phase compensation error that requires the accuracy of the wavelength order remains. Sometimes.
Various methods for estimating the phase error from the received signal itself have been proposed for such problems.

In particular, the second-order phase error has a large influence and is a basic error that frequently occurs, and various estimation methods specialized to this error have been proposed (for example, the map drift (MD) method and the phase difference). (PD) method).
It is useful to apply these methods to the estimation of the second order phase coefficient. However, in translational motion compensation, it is often assumed that the phase error to be estimated is almost equal in any range cell. Based on this assumption, the estimation result of phase change obtained by each range cell, the amount of change, intermediate index, etc. are integrated in the range cell direction (for example, average, integration, etc.) to improve the final estimation accuracy. Often.

However, in the problem that we deal with here, since the second-order phase change differs for each range, the processing in the conventional translational motion compensation cannot be used as it is.
That is, basically, the estimation result obtained for each range cell must be used as it is. Therefore, there is a possibility that accuracy deterioration occurs due to not integrating in the range cell direction, which causes a problem.

In the conventional method 2 (see Non-Patent Document 2) in which an image is formed by compensating for the second-order phase error in each range after compensating for the range cell movement due to the rotational motion described above, the PD described above is used. Applying the method, second-order phase change is estimated for each range cell.
Then, the estimation result group is concerned that an outlier that has largely failed in estimation is mixed, and a process of detecting and removing this is applied.
Thereafter, the estimation result is smoothed, and the secondary phase change of each range cell is finally determined.
However, in this method, as the number of outliers increases, the number of data used for estimation decreases and the estimation accuracy deteriorates. In addition, there is a high possibility that the outlier removal itself will fail, and the estimation accuracy deteriorates.

In the PD method, an evaluation value is calculated for each candidate secondary phase coefficient in the estimation process in each range cell, and the secondary phase coefficient having the highest evaluation value may be a true secondary phase coefficient. It is considered that the nature is high.
In the conventional method 2, based on this viewpoint, the secondary phase coefficient having the highest evaluation value in each range cell is used as a temporary estimation result in the range cell.

However, the true second-order phase coefficient evaluation value may not be the highest depending on the reflection point distribution on the image and the noise situation.
For example, in PD and MD, in principle, an amount corresponding to the Doppler shift between images generated only for the first half data and only the second half data of the synthetic aperture is estimated, and the secondary phase coefficient is calculated from the estimation result. It has gained. At this time, depending on the state of the reflection point distribution, the evaluation value of the candidate value that is significantly different from the true value may be increased due to the influence of interference between the reflection points.
There is a high possibility that the outlier in the conventional method 2 is generated under such influence. However, even in such a situation, even if the true evaluation value does not become the maximum, it retains a high value, so if this information is used well, the estimation accuracy may be improved. .
From this point of view, in the present invention, the secondary phase coefficient is estimated based on the entire distribution of evaluation values.

The processing contents of the secondary phase coefficient candidate evaluation index calculation circuit 52 for each range cell will be specifically described below.
The secondary phase coefficient candidate evaluation index calculation circuit 52 for each range cell uses a secondary phase estimation method that can calculate some evaluation value for a secondary phase coefficient candidate value such as PD or MD, A secondary phase coefficient evaluation value distribution for each candidate value is calculated.
Hereinafter, a case where PD is used will be described as an example.

It is assumed that the true value of the phase secondary phase coefficient in a certain range cell is given by a ₂ .
Assume that the received signal g (t) of the target range cell is given by the following equation (20) using a signal s (t) that does not include a secondary phase error.

The secondary phase coefficient candidate evaluation index calculation circuit 52 for each range cell calculates the signal g _s (t) of the first half of the observation time and the signal g _e (t) of the second half from the received signal g (t) of the range cell of interest Extract as shown in (21) and (22).

これを得るためには、ｇ_P（ｔ）をフーリエ変換して得られるスペクトルを確認すればよい。 Here, assuming a simple case where s (t) = γ _P ^ej2πfpt and only one point of the amplitude γ _P and the Doppler frequency f _P exists, g _P (t) obtained by the following equation (23) A secondary coefficient a ₂ appears in the primary change in phase.

In order to obtain this, a spectrum obtained by Fourier transforming g _P (t) may be confirmed.

If a term that does not vary with time is ignored, the spectrum G _P (f) of g _P (t) is given by the following equation (24).

When f that maximizes | G _P (f) | is f _max [Hz], a ₂ [1 / s ² ] is obtained by the following equation (25).

Based on the above, the secondary phase coefficient candidate evaluation index calculation circuit 52 for each range cell sets a value corresponding to | G _P (f) | as the secondary phase coefficient evaluation value distribution.
In the above description, the observation time t is treated as a continuous value, but the same applies even if the observation time is a discrete time represented by a hit h and a pulse repetition period Δt.

In addition, _s (h, m s) sequences _{s s (h 2, m s} ) hit the first half of the second half of the array _{s e (h 2, m s} ) (h 2 = -H / 4, ··· , H / 4-1), s _P (h ₂ , m _s ) = s _e (h ₂ , m _s ) s _s (h ₂ , m _s ) ^* is obtained by Fourier transform in the hit h ₂ direction. The array is represented by G ₀ (d ₂ , m _s ) (d ₂ = −H / 4,..., H / 4-1).
Here, the unit of the axis in the d ₂ direction of G ₀ (d ₂ , m _s ) is a cell, but the Doppler resolution has deteriorated twice due to the fact that the observation time is divided into two by equation (25). In consideration of the above, by multiplying cell number ₂ by A, the distribution can be converted to a distribution with respect to the secondary phase coefficient.

このＧ（ｄ_s，ｍ_s）を２次位相係数評価値分布とする。 Here, the above G ₀ (d ₂ , m _s ) is interpolated with zero padding in the d ₂ axis direction to increase the number of cells in this direction and reduce the cell width. The number of points is increased to N _intp times by zero padding interpolation.
The resulting sequence is represented by G (d _s , m _s ) (d _s = −H _s / 2,..., H _s / 2-1). However, H _s = (H / 2) N _intp .
In this case, conversion factor .DELTA.a ₂ from cell number d ₂ to the secondary phase factor is given by the following equation (27).

This G (d _s , m _s ) is defined as a secondary phase coefficient evaluation value distribution.

Although the method using the PD method has been described here, the method is not limited to the PD method, and other methods such as the MD method can be used instead.
In the MD method, images are generated in the first half and the second half of the hit, the shift between images is measured from the peak position of the cross-correlation between the images, and the shift between the images is converted into a secondary phase change coefficient. To do.
Therefore, the cross-correlation value for each shift can also be used as the evaluation value of the candidate secondary phase change coefficient, as in the case of the PD method.

When receiving the secondary phase coefficient evaluation value distribution from the secondary phase coefficient candidate evaluation circuit 4 for each range cell, the secondary phase coefficient variation estimation circuit 5 receives the secondary phase coefficient evaluation value distribution based on the secondary phase coefficient evaluation value distribution. A secondary phase coefficient range change coefficient that is a change coefficient with respect to the range is estimated.
Hereinafter, the estimation process of the secondary phase coefficient range variation coefficient by the secondary phase coefficient variation estimation circuit 5 will be specifically described.
As a two-dimensional image indicated by the secondary phase coefficient evaluation value distribution, a general image called s (y, x) is assumed.
The number of pixels in the x-axis direction of this image is Nx, and the number of pixels in the y-axis direction is Ny. When this is associated with G (d _s , m _s ), d _s corresponds to y, and m _s corresponds to x.

First, the spectral image utilization type trajectory inclination specifying circuit 61 of the second-order phase coefficient variation estimation circuit 5 is configured so that any straight line having the same inclination on the image is obtained by performing a two-dimensional Fourier transform on the image. The inclination of the straight line on the original image is estimated using the property of Fourier transform that passes through the origin (DC component) and is projected into a straight line having an inclination depending on the inclination of the original straight line.
In general, a straight line having an inclination a on the xy two-dimensional infinite plane passes through the origin on an fx-fy two-dimensional infinite frequency plane obtained by performing two-dimensional Fourier transformation on the straight line with an inclination A of It is converted into a straight line given by equation (28).

This relationship is represented by the inclination a of the straight line group on the two-dimensional image s (y, x) (see FIG. 9A) and the spectrum image S (fy, fx) obtained by performing two-dimensional Fourier transform on this. In the special case where the number of pixels Nx in the x direction and the number of pixels Ny in the y direction are equal to the straight line slope A (see FIG. 9B).
In the general case where the number of pixels Nx in the x direction is different from the number of pixels Ny in the y direction, the relationship between the two slopes is based on the effect that the spatial resolution in each axial direction becomes 1 / Nx and 1 / Ny, respectively. Is given by equation (29).

In any case, if the slope on the spectrum image is known, the slope a of the straight line group on the original image can be calculated based on the relationship of the above equation.
In particular, the merit of estimating the inclination on the spectrum image S (fy, fx) is that since the straight line passes through a fixed point (origin), it can be simplified to a straight line detection problem passing through the fixed point and the inclination can be estimated.

Based on the above, the trajectory slope reliability distribution calculation circuit 71 of the spectral image utilization type trajectory slope specifying circuit 61 performs two-dimensional Fourier transform on the amplitude distribution that is the absolute value of the two-dimensional image indicated by the secondary phase coefficient evaluation value distribution. A spectrum image is generated, and an amplitude distribution that is an absolute value of the spectrum image is acquired.
Detection of a straight line passing through a fixed point can be performed by detecting a peak of a line integration result of a spectrum image along a path that passes through the fixed point and has a different slope.
One or more expected values are prepared as the inclination a of the straight line group on the two-dimensional image, and the integration path on the spectrum image corresponding to each is set as the integration path group.

For example, the candidate value of a is a_cand (ia) (ia = 0, 1,..., Na-1: Na is the number of candidates for a).
When a candidate value is an equally-spaced value of Wa width centered on ac, a_cand (ia) is given by the following equation (30).
a_cand (ia) = ac + (Wa / Na) × (ia−1 / 2) (30)
The candidate values are not necessarily equal intervals, and it is possible to set unequal intervals as a_cand (ia). However, if there is a process that assumes that a_cand (ia) is equally spaced in the subsequent stage, a value that is equally spaced is set.
The slope A_cand (ia) of the integration path on the spectrum image with respect to a_cand (ia) is given by the following equation (31).
A_cand (ia) =-(Ny / Nx) 1 / a_cand (ia) (31)

Next, the locus inclination reliability distribution calculation circuit 71 sets the position fy0 (ii) (ii = 0, 1,..., 2Ny−1) of the 2Ny point in the fy direction as shown in the following equation (32). To do.
fy0 (iy) = iy−Ny (32)
Further, assuming that the pixel number in the fy direction of the spectrum image is given by 0 to Ny−1, fy (iy) is obtained by the following equation (33).
fy (iy) = fy0 (iy) mod Ny (iy = 0, 1,..., 2Ny−1) (33)

Next, the trajectory slope reliability distribution calculation circuit 71 gives the pixel number fx (iy) in the fx direction for each iy by the following equation (34).
fx (iiy) = (1 / A_cand (ia)) fy (iy) = − (Nx / Ny) a_cand (ia) (34)
Thereby, the integration path on the spectrum image is set with 2Ny sets of (fx (iy), fy (iy)).

The locus inclination reliability distribution calculation circuit 71 is based on the spectrum image amplitude distribution S (y, x) and the arrays fx (iy) and fy (iy) that set the integration path for each a_cand (ia). Perform line integral of the amplitude distribution of.
At this time, considering that the value of fx (iy) may be a decimal number, the value is determined by linear interpolation of the values of the pixels having two pixel numbers adjacent to fx (iy).
First, let floor (Z) be an operator that rounds off the decimal part of the real number Z. Pixel numbers fx_sml0 (iy) and fx_lrg0 (iy) smaller and larger than fx (iy) are expressed by the following equations (35), (36). Get in.
fx_sml0 (iy) = floor (fx (iy)) (35)
fx_lrg0 (iy) = fx_sml0 (iy) +1 (36)

Further, distances d_sml (iy) and d_lrg (ii) from fx (iy) to fx_sml0 (iy) and fx_lrg0 (iy) to be used for linear interpolation are obtained by the following equations (37) and (38).
d_sml (iy) = fx (iy) −fx_sml0 (iy) (37)
d_lrg (ii) = 1−d_sml (ii) (38)
In consideration of the fact that fx_sml0 (iy) and fx_lrg0 (iy) may indicate values outside 0 to Nx−1, the following pixel numbers fx_sml are determined by the following equations (39) and (40). (Ii), fx_lrg (iy) is acquired.
fx_sml (ii) = fx_sml0 (ii) mod Nx (39)
fx_lrg (iiy) = fx_lrg0 (iy) mod Nx (40)

Based on the above, the locus inclination reliability distribution calculation circuit 71 performs line integration of spectral images along paths having different inclinations.
At this time, as the line integration image, the following methods (a) to (c) are conceivable.
(A) Method of using S (fy, fx) as it is (b) When the distance from the origin of S (fy, fx) in the fy direction is smaller, the magnitude of the signal becomes smaller and the inclination detection sensitivity deteriorates. S (fy, fx) in each fy, S (fy, *) (where * is an expression representing all of the pixel numbers 0 to Nx−1 in the fx direction of the spectrum image. ) For the same purpose as the methods (c) and (b) using the spectrum image normalized by the maximum value or the square sum value (hereinafter referred to as a straight-line enhanced spectrum image), and the like. In this case, in consideration of subsequent processing, the pixel values of the above-described normalized all-spectrum image are further converted into square sum values for all the original pixels of S (fy, fx). Divided spectral image (total power-maintaining straight-line enhanced spectral image) In the following description the method using, expressed in (a) (b) For even spectral images obtained by any of the methods (c), the spectral image S1 (fy, fx).

The trajectory slope reliability distribution calculation circuit 71 obtains a value z (iy) for line integration for each iy by the following equation (41).
z (iy) = S1 (fy (iy)
, Fx_sml (iy)) × d_lrg (iy) + S1 (fy (iy)
, Fx_lrg (iy)) × d_sml (iy)
(41)

よって、以下では、軌跡傾き信頼度分布算出回路７１が線積分値Ｚ０を得るものとする。
なお、重みＷ（ｉｙ）については、その総和がある一定値Ｗｔであり（ただし、Ｗｔ＞０）、下記の式（４３）を満足するように設定する。

By integrating the value z (iy) for all iy, an integrated value for a_cand (ia) is obtained.
In this case as well, it is possible not only to simply add up the values as they are, but also to add up some weights W (iy), for example, to emphasize values near the origin.

Therefore, in the following, it is assumed that the locus inclination reliability distribution calculation circuit 71 obtains the line integral value Z0.
The weight W (iy) is set so as to satisfy the following equation (43) because the sum is a certain constant value Wt (Wt> 0).

  In particular, when Wt = 2Ny and W (iy) = 1 (iy = 0, 1,..., Ny), the same result is obtained as when z (iy) is summed as it is. Wt is not an essential value in the subsequent processing, but may be set to the same value when the same processing is repeatedly processed by changing a_cand (ia).

The integration value Z0 obtained based on the above processing along the integration path determined by each a_cand (ia) is represented by an array Z0 (ia) for ia.
The value of Z0 (ia) becomes higher as the integration path is along the locus on the spectral image, in other words, the closer a_cand (ia) is to the inclination of each locus on the original image.
That is, it can be considered that the inclination a_cand (ia) when the value of Z0 (ia) is high is more likely as a candidate value of the true inclination a. From this viewpoint, the Z (ia) for each a_cand (ia) is referred to as a locus inclination reliability distribution.

In the locus inclination reliability distribution shown in FIG. 10, the peak position is expected to be a true inclination a.
As described above, the trajectory slope reliability distribution calculation circuit 71 acquires the reliability distribution related to the trajectory slope.

When the trajectory slope reliability distribution calculation circuit 71 acquires the reliability distribution, the reliability distribution maximum position specifying unit 72 detects the inclination of the integration path having the maximum reliability in the reliability distribution.
That is, the reliability distribution maximum position locator 72 detects ia that maximizes Z (ia) by performing a general peak search.

The trajectory slope converter 73 obtains a trajectory slope candidate a_cand (ia) corresponding to ia detected by the reliability distribution maximum position locator 72, and uses the trajectory slope candidate a_cand (ia) as an estimation result a _{est of the} slope a. Is output to the secondary phase coefficient variation calculation circuit 62.
It should be noted that the estimation result a _est of the inclination a is the number of pixels (decimal number is also acceptable) when G (d _s , m _s ) is an input image, and the cell advances in the m _s direction while moving one cell in the d _s direction ).

Secondary phase coefficient change amount calculating circuit 62 receives the estimation result a _est slope a from the trajectory inclination converted 73, the estimation result a _est, and range pixel spacing [Delta] r _s after the aforementioned range interpolation, d _Using the conversion coefficient Δa ₂ from _s to the second-order phase coefficient, the second-order phase coefficient range change coefficient β ₂ , which is the rate of change with respect to the range of the second-order phase coefficient, is converted by the following equation (44).

この角速度の大きさ｜ω｜によって、前述のポーラーフォーマット法等における周波数平面上での信号の配置の角度を特定することができる。 When the rotational phase velocity conversion circuit 6 converts the estimation result a _est of the inclination a into the secondary phase coefficient range change coefficient β ₂ by the secondary phase coefficient change amount calculation circuit 62, the secondary phase is calculated by the following equation (45). The coefficient range change coefficient β ₂ is converted into a target equivalent rotational motion angular velocity magnitude | ω |.

The angle of signal arrangement on the frequency plane in the above-described polar format method or the like can be specified by the magnitude of this angular velocity | ω |.

The zero-crossing range estimation circuit 7 in consideration of the secondary phase coefficient estimates a range in which “the secondary phase coefficient is 0”, which is a property that the rotation center should have, and regards this range as the current rotation center range. . In addition, a range compensation amount for making this range 0 is estimated for the subsequent image reproduction.
That is, the fixed slope integral zero-crossing range estimation circuit 81 of the zero-crossing range estimation circuit 7 in consideration of the secondary phase coefficient 7 outputs the secondary phase coefficient evaluation value distribution G () output from the secondary phase coefficient candidate evaluation circuit 4 for each range cell. d _s , m _s ), the slope considering the scaling in both axial directions is the secondary phase coefficient range change coefficient β ₂ shown in the equation (44), and the range where the secondary phase coefficient is 0 is A plurality of different integration paths are set, and line integration along each integration path is performed.

  The slope fixed line integration type zero-crossing range estimation circuit 81 considers that the integration path where the value of the line integration becomes the maximum coincides with the locus as a result of performing the line integration along each integration path. 2, the range where the secondary coefficient is 0 is regarded as the range where the secondary phase coefficient is 0 (zero-crossing range), and the range is output to the 0th-order range compensation circuit 8.

When the zero-order range compensation circuit 8 receives a range in which the secondary phase coefficient is zero from the zero-crossing range estimation circuit 7 in consideration of the secondary phase coefficient, the translational compensation circuit 2 changes the distance so that the range becomes zero. Compensates for the compensated range history.
That is, the zero-order range compensation circuit 8 uses the range output from the zero-crossing range estimation circuit 7 in consideration of the second-order phase coefficient in the range history s ₂ (h, m) in which the distance change is compensated by the translational motion compensation circuit 2. Compensate over all hits to be zero. The compensation result is stored in the array s _rot (h, m).

In the range history s ₂ (h, m), the influence of the range change due to the translational motion is already compensated. In particular, by compensating for the first-order phase change, a point on the cross range where “range change is 0”, which is one of the conditions that the rotation center should have, is arranged in the 0 Doppler cell.
In addition, by arranging the range in which “the secondary phase coefficient is 0 (in other words, the change in the Doppler frequency being observed is 0)”, which is another condition that the rotation center should have, in the 0 range, Is 0 and the secondary phase coefficient is 0 ”, in other words, there is an effect that the position on the image being observed is 0, that is, the center of rotation can be arranged at zero Doppler.

The rotation-considering radar image generation circuit 9 performs image reproduction in consideration of rotation based on the angular velocity estimated by the rotation angular velocity conversion circuit 6, thereby obtaining an image from the range history after compensation by the zero-order range compensation circuit 8. A radar image is generated in which both axes are scaled to a physical length axis.
In other words, the rotation-considered radar image generation circuit 9 reproduces an image of the range history in which the range of the rotation center is compensated after the translational motion is compensated by a method that considers the rotation like a general polar format method. A radar image in which blur caused by the influence of rotation is correctly compensated and cross range scaling is correctly performed is generated.

The processing contents of the rotation-considering radar image generation circuit 9 will be specifically described below.
The rotation-considering radar image generation circuit 9 sets the range history (range history stored in the array s _rot (h, m)) that is expected to be completely compensated for the influence of translational motion and the range shift of the rotation center. U (h, n) (n = −M / 2,..., M / 2-1) which is a history (range spectrum history) regarding the spectrum of the range profile is obtained by Fourier transform in the direction.
This range spectrum is positioned in the frequency space at a position corresponding to the intermediate frequency and bandwidth of the transmission pulse and the angle η (h) that changes due to rotation (the angle in the direction of the rotation center viewed from the radar in the target fixed coordinate system). Deploy.
Here, η (h) is expressed by the following formula (46).

Here, assuming that the frequency in the n-th spectrum cell is f (n), for example, the center frequency is f _c and the transmission bandwidth is B, the frequency f (n) in the n-th spectrum cell is given by the following formula ( 47).

Rotation consideration radar image generating circuit 9, range spectrum history u (h, n) and, on the grounds of "Central Slice Theorem" known, f _r -f _c on frequency plane, n, the following formula for each h ( The music coordinates are arranged at the position determined by 48).

As described above, the frequency distribution regarding the target reflection intensity is obtained.
By re-sampling this frequency distribution with a rectangular grid at an appropriate position and subjecting it to a two-dimensional inverse Fourier transform, a target reflection intensity distribution in the time plane can be obtained.
By multiplying both axes of this image by a known (high speed / 2), it is possible to obtain a physical length axis. That is, a radar image in which cross range scaling is realized is obtained.
In this radar image, blur due to rotation is compensated by the effect of a series of processes on the frequency plane.

In the first embodiment, by adopting the above configuration, the following effects can be obtained.
[1]
In the compensation process for translational motion, in particular, since the first-order phase change that generally does not require compensation is compensated, the point on the cross range where the range change becomes 0 after translational motion is placed in the zero Doppler cell. The following effects can be obtained.

(A1)
When the range movement due to rotation is compensated based on the conventional method 2, it is necessary to change the compensation amount according to each Doppler position based on the Doppler cell at the center of rotation. Therefore, it is necessary to know which Doppler cell the rotation center is located in.
However, in the conventional method 2, it is merely stated that the translational motion is completely described, and a method of knowing the Doppler cell at the center of rotation and a method of arranging at the 0 Doppler frequency are specifically disclosed. Not.
However, in the first embodiment, this can be specifically realized by the above processing.
Instead of compensating for the first-order phase change, the center of rotation assumes that a Doppler shift corresponding to the phase change is made, and in the above compensation in each Doppler cell, each Doppler frequency is corrected by this shift. However, there is no essential difference from the method shown in the first embodiment.

(A2)
In the image reproduction considering the subsequent rotation, the radar and the rotation center need to be fixed. This process compensates for the movement of the rotation center corresponding to the Doppler frequency and contributes to the achievement of the condition that the rotation center is fixed.
The polar format method, which is the conventional method 1, is basically a method when the relative motion between the radar and the target can be grasped with considerably high accuracy, and the relative motion information can be given as known information with high accuracy. Such compensation was difficult in cases where it could not be expected.

[2]
In the first embodiment, when estimating the secondary phase coefficient for each range cell, the change of the secondary phase coefficient with respect to the range is estimated using the distribution of the entire evaluation value for each candidate value in each range. There is an effect that the accuracy can be increased.
On the other hand, the conventional method 2 corresponds to estimation based on only the maximum position of the evaluation value in each range. Therefore, there is a problem that the estimation accuracy is low due to being easily affected by an outlier due to interference or the like.

[3]
When estimating the secondary phase coefficient for each range, since the signal is interpolated in the range direction, the difference between the true range of the reflection point and the center of the range cell to which the reflection point belongs can be reduced.
Thereby, when calculating the evaluation index related to the estimation of the secondary phase coefficient for each range cell, it is possible to reduce the occurrence of an error based on the above range difference.
Similarly, in the conventional method 2 in which similar processing is performed for each range cell, since such a device is not applied, there is a problem that an error based on this is not reduced.

[4]
When estimating the coefficient of the range change of the secondary phase coefficient, this problem is reduced to the problem of estimating the slope of the straight line (group) on the image, and by using the property of Fourier transform, this is converted to a known fixed point. Since the problem is solved by simplifying the problem of detecting a straight line that passes through, the coefficient of the range change of the secondary phase coefficient can be stably estimated with high accuracy.
On the other hand, in the conventional method, estimation is performed based only on the information on the maximum position, and the maximum position depends on the reflection point distribution on the target and is affected by interference between the reflection points. May occur in very different positions. In such a case, even if outliers are detected and removed and estimation is performed again, the estimation accuracy deteriorates due to further lack of information. Further, if the frequency of occurrence of outliers increases, there is a problem that removal itself becomes difficult.
In the first embodiment, even if a maximum value occurs at a position different from the true value in a certain range, the information on a certain high peak existing in the true value (the information on a certain high peak even if not the maximum) ) Can be used to minimize degradation of accuracy.

[5]
Non-patent documents describing conventional methods suggest that the rotational angular velocity can be estimated from the relationship between the secondary phase coefficient and the range position.
However, this is actually a story that is determined by the combination of a range based on the center of rotation at a certain range position and its secondary phase coefficient. For this purpose, the range of the center of rotation is known. Need to be. Since the range of the rotation center is unknown, the method suggested in the non-patent literature is actually difficult to realize.
On the other hand, in the first embodiment, the rotational angular velocity is obtained from the range change of the secondary phase coefficient, and the rotational angular velocity can be estimated without requiring an unknown rotational center range. is there.

[6]
In order to realize the polar format method which is the conventional method 1, it is necessary to perform processing after fixing the rotation center to the zero range. In this specific implementation method, in particular, the relative motion between the radar and the target is unknown. However, in reality, it is difficult to perform high-precision imaging in consideration of rotation.
On the other hand, in the first embodiment, the compensation of the first-order phase change described above results in a cross range in which the range change, which is one of the conditions to be satisfied by the rotation center, that is, the rotation center is In order to arrange the included cross range at 0 Doppler, the radial speed corresponding to the Doppler frequency of the cross range is set to 0.
Furthermore, the second phase coefficient range in which the slope is already determined in the evaluation value distribution described above, which is another condition that the rotation center should satisfy, “the range where the second-order phase coefficient generated by rotational motion is 0”. The estimation can be realized by searching for the maximum value of the line integration results along a plurality of paths determined by coefficients and having different intercepts. That is, the range of the rotation center can be specified.
Therefore, since compensation that makes this range zero is realized, compensation that completely fixes the range of the center of rotation necessary for image reproduction considering rotational motion to zero is realized.

[7]
The above-described estimation of the rotational angular velocity and the realization of the compensation for completely fixing the rotation center range to zero have the effect of realizing image reproduction in consideration of the rotational motion using these.
That is, the conventional method 1 has an advantage that it can solve the problem that an image cannot be actually reproduced or a low-quality image can be generated due to an unknown estimation error being too large, particularly in a situation where the relative motion is unknown. .

[8]
In the conventional method 2, the compensation of the range change based on the rotation and the compensation of the Doppler frequency change are processed in multiple stages without knowing the information of the rotational motion, and the compensation of the image blur due to the rotation is realized. Is realized by replacing the actual motion with an approximate model, which may cause signal distortion based on this approximation and reduce image quality.
In the first embodiment, a method similar to the conventional method 2 is used only for estimating the rotation and fixing the rotation center, and actual imaging is realized by a precise method considering the rotation. There is an effect that the type of image quality degradation that occurs in the method 2 does not occur.

[9]
Conventional method 2 can compensate for blur caused by rotation, but cannot realize cross-range scaling.
On the other hand, in the first embodiment, it is possible to realize not only compensation for blur caused by rotation but also cross range scaling.

Embodiment 2. FIG.
11 is a block diagram showing an image radar apparatus according to Embodiment 2 of the present invention. In the figure, the same reference numerals as those in FIG.
The correction type zero-crossing range estimation circuit 10 before considering the secondary phase coefficient estimates the range where the secondary phase coefficient is 0 based on the distribution of the secondary phase coefficient evaluation value and the estimation result of the secondary phase coefficient range change coefficient. Further, this is a circuit for calculating a compensation amount for setting the range where the secondary phase coefficient is 0 to 0 range based on the estimation result.

FIG. 12 is a block diagram showing the zero-crossing range estimation circuit 10 before correction of the second-order phase coefficient in the image radar apparatus according to Embodiment 2 of the present invention.
In FIG. 12, the preprocessing correction summation type zero-crossing range estimation circuit 91 has a straight line inclination a estimated by the spectral image utilization type locus inclination specifying circuit 61, and one axis (for example, vertical axis) set in advance. A zero-crossing point that is a coordinate (x0 when the locus is represented by y = a (x + x0)) on another axis (for example, the horizontal axis x) when passing through a position where the coordinate along y) is 0 In order to estimate, the inclination of the locus on the two-dimensional image indicated by the secondary phase coefficient evaluation value distribution output from the secondary phase coefficient candidate evaluation circuit 4 for each range cell becomes 0 based on the slope a of the straight line. This is a circuit that corrects, sums up the corrected two-dimensional image on the x-axis, and estimates the zero-crossing point from the y value y0 that maximizes the total value of each y.

Next, the operation will be described.
The method of estimating the zero-crossing range is different between the second embodiment and the first embodiment.
In the problem of estimating x0 when the trajectory on the two-dimensional image indicated by the secondary phase coefficient evaluation value distribution can be expressed by y = a (x + x0) on the xy plane,
In the second embodiment, all pixels are translated by −ax by some method for each x so that y ′ = y−ax, with the position of x = 0 as the center.
At that time, the pixels coming out from the end of the image are moved to the opposite end as they are in the manner of cyclic shift. The above is a process that appears in general primary range compensation.

In the compensated image, the original trajectory is represented by y ′ = ax0.
Therefore, the zero-crossing range x0 to be obtained is expressed by the following equation (49) using y _max that is y that is the peak of the value in the distribution obtained by summing the compensated image in the x direction. .
x0 = y _max / a (49)

The preprocessing correction sum type zero-crossing range estimation circuit 91 estimates the zero-crossing range based on the above principle and outputs the zero-crossing range to the zero-order range compensation circuit 8.
Other processes are the same as those in the first embodiment.

By adopting the configuration of the second embodiment, the same effects as those of the first embodiment can be obtained.
Further, as an effect when compared with the first embodiment, the evaluation index distribution used for estimating the zero-crossing range is obtained by processing equivalent to general first-order range compensation and the summation in the uniaxial direction of the array. Therefore, it can be easily configured and the processing load may be reduced.

Embodiment 3 FIG.
13 is a block diagram showing an image radar apparatus according to Embodiment 3 of the present invention. In the figure, the same reference numerals as those in FIG.
The secondary phase coefficient reference zero-crossing range estimation circuit 11 estimates the secondary phase coefficient of each range based on the secondary phase coefficient evaluation value distribution output from the secondary phase coefficient candidate evaluation circuit 4 for each range cell. This is a circuit that determines that the range in which the secondary phase coefficient is closest to 0 among the secondary phase coefficients is the zero-crossing range, and outputs the range to the zero-order range compensation circuit 8.

Next, the operation will be described.
In the first and second embodiments, the zero crossing range is estimated using the inclination of the locus on the two-dimensional image indicated by the secondary phase coefficient evaluation value distribution. In the third embodiment, By looking only at the secondary phase coefficient of each range obtained from the secondary phase coefficient evaluation value distribution (candidate value of the secondary phase coefficient that maximizes the evaluation value in each range), the secondary phase coefficient is closest to 0. The range cell is estimated as the zero-crossing range.

That is, the secondary phase coefficient reference zero-crossing range estimation circuit 11 compares the secondary phase coefficients of the respective ranges obtained from the secondary phase coefficient evaluation value distribution, and specifies the range where the secondary phase coefficient is closest to 0. To do.
Then, it is determined that the range is a zero-crossing range, and the range is output to the 0th-order range compensation circuit 8.
The other processes are the same as those in the first and second embodiments.

By adopting the configuration of the third embodiment, the same effects as those of the first and second embodiments can be obtained.
In addition, as an effect when compared with the first and second embodiments, when estimating the zero-crossing range, the inclination of the trajectory on the two-dimensional image indicated by the secondary phase coefficient evaluation value distribution estimated in the previous stage is shown. Since it is not necessary to perform the line integration process (Embodiment 1) of the second-order phase coefficient evaluation value distribution and the compensation and summation process (Embodiment 2), the processing load is reduced and the circuit is simple. The effect which is made can be produced.

Embodiment 4 FIG.
14 is a block diagram showing an image radar apparatus according to Embodiment 4 of the present invention. In the figure, the same reference numerals as those in FIG.
The range history reference zero crossing range estimation circuit 12 after translation compensation refers to the range history in which the distance change is compensated by the translation motion compensation circuit 2, estimates the range where the range movement caused by the rotational motion is minimum, and calculates the range. This circuit outputs the zero-crossing range to the zero-order range compensation circuit 8.

Next, the operation will be described.
When the range history after translation compensation based zero-crossing range estimation circuit 12 receives the range history after translation motion compensation from translation motion compensation circuit 2, the range history is divided into range areas in the range direction (overlapping in the range direction). To divide.
The post-translation-compensated range history reference zero-crossing range estimation circuit 12 performs a range cell movement amount estimation process performed in general translational motion compensation in each segmented region.

It can be considered that the zero cross range exists in the divided region of the range where the movement amount becomes zero.
For the zero-crossing range in the segmented area, the value is the maximum in the distribution obtained by summing the amplitude values of the range history in the segmented area in the hit direction (or the sum of squares in terms of power comparison). It can be obtained as a peak position where the range or the distribution shape becomes sharpest.

Therefore, the post-translationally compensated range history reference zero-crossing range estimation circuit 12 determines that the range in which the value is maximum is the zero-crossing range in the distribution obtained by summing the amplitude values of the range history in the segmented region in the hit direction. presume. Alternatively, it is estimated that the range of the peak position with the sharpest distribution shape is the zero crossing range.
The range history reference zero-crossing range estimation circuit 12 after translation compensation outputs the estimated zero-crossing range to the zero-order range compensation circuit 8.
Processes other than the translation compensated range history reference zero-crossing range estimation circuit 12 are the same as those in the first embodiment.

By adopting the configuration of the fourth embodiment, the same effects as those of the first to third embodiments can be obtained.
In comparison with the first to third embodiments, since the range movement estimation method used in general translational motion compensation can be applied as it is, there is an effect that it can be used without much repair cost of the already held device. .

Embodiment 5 FIG.
15 is a block diagram showing an image radar apparatus according to Embodiment 5 of the present invention. In the figure, the same reference numerals as those in FIG.
The secondary phase coefficient estimation circuit 13 shows the distribution of secondary phase coefficient evaluation values output from the secondary phase coefficient candidate evaluation circuit 4 for each range cell based on a general method for detecting a straight line on an image such as Hough transform. The inclination and intercept of the trajectory on the two-dimensional image are specified, and the secondary phase coefficient range change coefficient that is a change coefficient with respect to the range of the secondary phase coefficient is estimated from the inclination and intercept of the trajectory. It is a circuit that estimates a range that is zero.

Next, the operation will be described.
When receiving the secondary phase coefficient evaluation value distribution G (d _s , m _s ) from the secondary phase coefficient candidate evaluation circuit 4 for each range cell, the secondary phase coefficient estimation circuit 13 detects a straight line on the image. By performing the Hough transform, the locus on the two-dimensional image indicated by the secondary phase coefficient evaluation value distribution G (d _s , m _s ) is detected.

When detecting the trajectory on the two-dimensional image, the secondary phase coefficient estimating circuit 13 determines the inclination of the trajectory (the number of pixels a _est that advances in the m _s direction while proceeding by one cell in the d _s direction (a decimal number is allowed)). , The zero-crossing range cell of the trajectory (the trajectory and the range cell m _X (interval of decimals) at the intersection of d _s = 0) is acquired.

When the secondary phase coefficient estimating circuit 13 obtains the inclination of the trajectory and the zero-crossing range cell, it uses the above-described range pixel spacing Δr _s , d _s after the range interpolation and the conversion coefficient Δa ₂ to the secondary phase coefficient. The secondary phase coefficient range change coefficient β ₂ is calculated from the following equation (50), and the zero-crossing range r _X is calculated from the following equation (51).

After calculating the secondary phase coefficient range change coefficient β ₂ and the zero-crossing range r _X , the secondary phase coefficient estimation circuit 13 outputs the secondary phase coefficient range change coefficient β ₂ to the rotation angular velocity conversion circuit 6 and outputs the zero. The intersection range r _X is output to the 0th-order range compensation circuit 8.
Processing other than the secondary phase coefficient estimation circuit 13 is the same as that in the first embodiment.

By adopting the configuration of the fifth embodiment, the same effects as those of the first to fourth embodiments can be obtained.
Moreover, in comparison with the said Embodiment 1-4, there exists an effect which can comprise an apparatus using general Hough conversion.
Furthermore, since the secondary phase coefficient range change coefficient β ₂ and the zero-crossing range r _X can be obtained simultaneously in the secondary phase coefficient estimation circuit 13, there is an advantage that the device configuration is simplified.

Embodiment 6 FIG.
FIG. 16 is a block diagram showing an image radar apparatus according to Embodiment 6 of the present invention. In the figure, the same reference numerals as those in FIG.
The zero-crossing range estimation circuit 21 specifies the inclination of the locus on the two-dimensional image indicated by the secondary phase coefficient evaluation value distribution output from the secondary phase coefficient candidate evaluation circuit 4 for each range cell, and the locus zero is determined from the inclination of the locus. This is a circuit for estimating the intersection range.

  The secondary phase compensation circuit 22 for each range cell calculates the secondary phase of each range from the secondary phase coefficient range variation coefficient estimated by the secondary phase coefficient variation estimation circuit 5 and the zero crossing range estimated by the zero crossing range estimation circuit 21. This is a circuit that calculates a compensation amount that cancels the phase change, and compensates the secondary phase change of each range in the range history after compensation by the rotation range cell movement compensation circuit 3 according to the compensation amount.

The radar image generation circuit 23 is a circuit that generates a range Doppler image by Fourier-transforming the range history after compensation by the secondary phase compensation circuit 22 for each range cell in the hit direction.
The cross range scaling circuit 24 is based on the angular velocity of the rotational motion estimated by the rotational angular velocity conversion circuit 6, and the cross range that is the physical length of the Doppler frequency axis on the range Doppler image generated by the radar image generation circuit 23. It is a circuit that performs cross-range scaling to convert to an axis.

Next, the operation will be described.
In the sixth embodiment, as in the conventional method 2, image blur due to rotation is compensated (range change due to rotation and Doppler change due to rotation are compensated in two stages), and the compensated image is Fourier transformed. By doing so, a range Doppler image is generated.
In the sixth embodiment, cross range scaling that cannot be performed by the conventional method is performed.

  The processing contents of the range history acquisition circuit 1, the translation motion compensation circuit 2, the rotation range cell movement compensation circuit 3, the secondary phase coefficient candidate evaluation circuit 4 for each range cell, and the secondary phase coefficient change amount estimation circuit 5 are the same as those in the first embodiment. It is the same.

Upon receiving the secondary phase coefficient evaluation value distribution from the secondary phase coefficient candidate evaluation circuit 4 for each range cell, the zero-crossing range estimation circuit 21 specifies the inclination of the locus on the two-dimensional image indicated by the secondary phase coefficient evaluation value distribution. Then, the zero crossing range of the locus is estimated from the inclination of the locus.
That is, the zero-crossing range estimation circuit 21 estimates the zero-crossing range of the trajectory by performing the same processing as any of the following circuits (a) to (e).
(A) Zero-crossing range estimation circuit 7 considering secondary phase coefficient
(B) Correction type zero-crossing range estimation circuit 10 before secondary phase coefficient consideration
(C) Secondary phase coefficient reference zero-crossing range estimation circuit 11
(D) Range history reference zero-crossing range estimation circuit 12 after translation compensation
(E) Secondary phase coefficient estimation circuit 13

In (a) above, as described in detail in the first embodiment, the input values used for estimation are different. Therefore, the wiring between the blocks for obtaining the input value and the related blocks are also different.
FIG. 16 shows a configuration diagram assuming that (a) and (b) are used. However, when (c) is used, the zero-crossing range estimation is performed from the secondary phase coefficient variation estimation circuit 5. Wiring to the circuit 21 is not necessary.
When (d) is used, instead of the need for wiring from the secondary phase coefficient candidate evaluation circuit 4 for each range cell and the secondary phase coefficient change amount estimation circuit 5 to the zero-crossing range estimation circuit 21, a translational motion compensation circuit is provided. It is necessary to add a wiring from 2 to the zero crossing range estimation circuit 21.
When (e) is used, the configuration is as shown in FIG.
These may be appropriately changed based on the input of the zero-crossing range estimation circuit 21 or a circuit corresponding thereto.

When the zero-crossing range estimation circuit 21 estimates the zero-crossing range, the secondary phase compensation circuit 22 for each range cell uses the zero-crossing range and the secondary phase coefficient range variation coefficient estimated by the secondary phase coefficient variation estimation circuit 5. Then, a compensation amount for canceling the secondary phase change of each range is calculated, and the secondary phase change of each range in the range history after compensation by the rotation range cell movement compensation circuit 3 is compensated according to the compensation amount.
That is, the secondary phase compensation circuit 22 for each range cell calculates the secondary phase coefficient of each range from the zero crossing range and the secondary phase coefficient range change coefficient.
Then, in the range history after compensation by the rotational range cell movement compensation circuit 3, a secondary phase change causing a Doppler change remaining in each range is calculated from the secondary phase coefficient, and the secondary phase change is canceled out. To compensate.

The radar image generation circuit 23 Fourier-transforms the range history (range history expected to be largely compensated for the blur due to the translational motion and the rotational motion) after the compensation by the secondary phase compensation circuit 22 for each range cell. Thus, a range Doppler image is generated.
In the range Doppler image generated by the radar image generation circuit 23, the main components of blur in the range and the Doppler frequency direction are compensated as in the conventional method 2.

When the radar image generation circuit 23 generates a range Doppler image, the cross range scaling circuit 24 performs Doppler on the range Doppler image based on the magnitude | ω | of the rotational motion estimated by the rotation angular velocity conversion circuit 6. Cross range scaling is performed to convert the frequency axis to the cross range axis corresponding to the physical length.
The conversion factor from the Doppler frequency [Hz] to the cross range [m] is represented by λ / (2 · | ω |) based on the above equation (3).
Therefore, the conversion coefficient may be multiplied by the Doppler frequency [Hz] of the range Doppler image to make the cross range axis.

The conversion factor from the Doppler cell number [cell] to the Doppler frequency [Hz] is expressed as 1 / T using the observation time T.
Therefore, when the axis in the Doppler direction of the range Doppler image is expressed by the Doppler cell number, the Doppler cell number is converted to the cross-range axis by multiplying by λ / (2 · | ω | T). Can do.
As described above, a range Doppler image generated by a flow similar to that of the conventional method 2 and compensated for the main component of blur due to rotation can be converted into a cross-range scaled image.

By adopting the configuration of the sixth embodiment, the same effects as those of the first to fifth embodiments can be obtained.
Further, in comparison with the first to fifth embodiments, the range history compensated for the range movement due to the rotation generated for the estimation of the rotational angular velocity and the rotation center range (zero crossing range) By reusing the secondary phase coefficient range change rate, it is possible to generate a range history in which the main component of blur due to rotation is compensated with little increase in processing load, and this processing load is rotated. It is possible to make a range Doppler by Fourier transform in the hit direction, which is lower than the image reconstruction method that takes into account the image, and to make this a range cross-range image by simple scaling using the estimated rotational angular velocity magnitude Therefore, it is effective from the viewpoint of reducing the processing load.

  In the present invention, within the scope of the invention, any combination of the embodiments, or any modification of any component in each embodiment, or omission of any component in each embodiment is possible. .

  1 range history acquisition circuit, 2 translational motion compensation circuit, 3 rotation range cell movement compensation circuit, 4 secondary phase coefficient candidate evaluation circuit for each range cell, 5 secondary phase coefficient change estimation circuit, 6 rotation angular velocity conversion circuit, 7 secondary phase Coefficient-considered zero-crossing range estimation circuit, 80th-order range compensation circuit, 9 Rotation-considered radar image generation circuit, 10 Second-order phase coefficient-considered correction-type zero-crossing range estimation circuit, 11 Second-order phase coefficient-based zero-crossing range estimation circuit, 12 translation-compensated range history reference zero-crossing range estimation circuit, 13 secondary phase coefficient estimation circuit, 21 zero-crossing range estimation circuit, 22 secondary phase compensation circuit for each range cell, 23 radar image generation circuit, 24 cross-range scaling circuit, 31 Transmitter, 32 Transmission / reception switch, 33 Transmit / receive antenna, 34 Receiver, 35 Range compressor 41 general translational motion compensation circuit, 42 primary phase change compensator, 51 range interpolator, 52 secondary phase coefficient candidate evaluation index calculation circuit for each range cell, 61 spectral image utilizing type trajectory slope specifying circuit, 62 secondary phase coefficient change amount Calculation circuit, 71 locus inclination reliability distribution calculation circuit, 72 reliability distribution maximum position locator, 73 locus inclination converter, 81 inclination fixed line integral type zero crossing range estimation circuit, 91 preprocessing correction sum total type zero crossing range estimation circuit .

Claims (22)

  While radiating a high-frequency signal toward the space, a part of the high-frequency signal reflected by the target existing in the space is received, and the resolution in the range direction of the received signal is subjected to range compression, and the received signal A range history acquisition circuit that acquires a range history that is a time history of the range profile by repeating a series of processes for acquiring a range profile that is an amplitude characteristic of the range direction while changing the relative position between the radar and the target, and The range history acquired by the range history acquisition circuit compensates for an unnecessary distance change between the radar and the target caused by unnecessary translation movement, and the distance change is compensated by the translation movement compensation circuit. In the range history, at each reflection point of the high-frequency signal at the above target, the range direction generated by the rotational motion The phase term of the data string in the hit direction of each range is specified in the range history in which the blur in the range direction is compensated by the rotational range cell movement compensation circuit and the rotational range cell movement compensation circuit. The likelihood of a plurality of candidate values that are candidates for the secondary phase coefficient that is the coefficient of the secondary phase change included in each is evaluated for each range cell, and the distribution in the range direction of the evaluation result of each candidate value is output. Based on the secondary phase coefficient candidate evaluation circuit and the secondary phase coefficient evaluation value distribution that is the distribution in the range direction output from the secondary phase coefficient candidate evaluation circuit, the change coefficient of the secondary phase coefficient with respect to the range. Based on the secondary phase coefficient range variation coefficient estimated by the secondary phase coefficient variation estimation circuit and the secondary phase coefficient variation estimation circuit that estimates the secondary phase coefficient range variation coefficient From the rotational angular velocity conversion circuit that estimates the angular velocity of the target equivalent rotational motion, the secondary phase coefficient range variation coefficient estimated by the secondary phase coefficient variation estimation circuit, and the secondary phase coefficient evaluation value distribution 2 The zero-crossing range estimation circuit for estimating the range where the next phase coefficient is 0, and the translational motion compensation circuit compensated for the distance change so that the range estimated by the zero-crossing range estimation circuit becomes 0. A zero-order range compensation circuit that compensates the range history, and a radar image generation circuit that generates a radar image from the range history after compensation by the zero-order range compensation circuit based on the angular velocity estimated by the rotational angular velocity conversion circuit. An image radar apparatus provided.   The translation motion compensation circuit estimates the range movement of each reflection point exceeding the range resolution and the second or higher order change with respect to the time in the range history acquired by the range history acquisition circuit. And a general translation motion compensation circuit that compensates for the second and higher order changes, and a primary phase change compensator that compensates for the primary change in phase in the range history after compensation by the general translation motion compensation circuit. The image radar apparatus according to claim 1. The general translational compensation circuit compensates for range shift and second-order or higher changes, and estimates the first-order change for the range shift hit of each reflection point.
The primary phase change compensator converts each primary change with respect to the range movement hit into a primary phase change, and cancels the phase change so that each range in the range history after compensation by the general translational motion compensation circuit is corrected. The image radar apparatus according to claim 2, wherein a first-order change in the phase of the image is compensated.   The secondary phase coefficient candidate evaluation circuit evaluates the probability of a plurality of candidate values that are candidates for secondary phase coefficients indicating the secondary phase change added to the phase of the signal of the range for each range cell, 2. The image radar apparatus according to claim 1, comprising a secondary phase coefficient candidate evaluation index calculating circuit for calculating a secondary phase coefficient evaluation value distribution in which the probabilities of a plurality of candidate values are digitized. .   The secondary phase coefficient candidate evaluation circuit interpolates the range history compensated for the blur in the range direction by the rotation range cell movement compensation circuit in the range direction, and outputs the interpolated range history to the secondary phase coefficient candidate evaluation index calculation circuit. The image radar apparatus according to claim 4, further comprising a range interpolator.   The second-order phase coefficient candidate evaluation index calculation circuit calculates an array obtained by multiplying an array that is a complex conjugate of the first half of the range history hit and the second half of the range history hit in the hit direction. 6. The image radar device according to claim 4, wherein an array obtained by Fourier transform is calculated as a secondary phase coefficient evaluation value distribution.   The second-order phase coefficient variation estimation circuit obtains a spectrum image by two-dimensional Fourier transforming the amplitude distribution of the two-dimensional image indicated by the second-order phase coefficient evaluation value distribution output from the second-order phase coefficient candidate evaluation circuit, and obtaining the spectrum image A trajectory slope specifying circuit that estimates the slope of a straight line on the two-dimensional image from the image, and a secondary phase coefficient change amount calculation that calculates a secondary phase coefficient range change coefficient from the straight line slope estimated by the trajectory slope specifying circuit The image radar apparatus according to claim 1, comprising a circuit.   The locus inclination specifying circuit performs weighting of the amplitude distribution of the two-dimensional image indicated by the secondary phase coefficient evaluation value distribution output from the secondary phase coefficient candidate evaluation circuit, and performs two-dimensional Fourier transform on the weighted amplitude distribution. A spectral image is obtained, and a plurality of linear integration paths having different inclinations are set through the origin corresponding to the DC component on the spectral image, and the amplitude distribution of the weighted spectral image is linearly integrated along the integration path. The reliability of the trajectory slope reliability distribution calculation circuit that outputs the result as the reliability of the slope of each integration path, and the reliability of the slope of each integration path output from the trajectory slope reliability distribution calculation circuit. Detects the slope of the integration path with the maximum reliability, and the slope of the integration path with the maximum reliability detected by the maximum confidence distribution position locator. Image radar apparatus according to claim 7, characterized in that it is composed of a trajectory slope conversion unit for converting the inclination of the trajectory on the two-dimensional image in which the secondary phase factor evaluation value distribution shown.   The zero-crossing range estimation circuit is a coordinate on the other axis when passing through a position where the inclination of the straight line estimated by the locus inclination specifying circuit is a and the coordinate along one axis of the preset image is 0 Is x0 when y is expressed as y = a (x + x0), the slope is a and x0 on the two-dimensional image indicated by the secondary phase coefficient evaluation value distribution output from the secondary phase coefficient candidate evaluation circuit. A plurality of integration paths with different candidates are set, line integration along each integration path is performed, and the x0 candidate having the maximum line integration value has a secondary phase coefficient of 0. 9. The image radar apparatus according to claim 7, wherein the image radar apparatus is estimated to be a range.   The rotational range cell movement compensation circuit generates a range Doppler image by Fourier transforming the range history compensated for the distance change by the translational motion compensation circuit in the hit direction, and for each Doppler cell of the range Doppler image, the Doppler cell is centered. The Doppler width is extracted, and the Doppler width is subjected to inverse Fourier transform in the Doppler cell direction and returned to the range history. On the range history, the primary range shift determined by the Doppler cell number of the Doppler cell is performed. The range history is Fourier-transformed in the hit direction again to return to the radar image, and the Doppler cell image in the radar image is stored in the compensated image storage array. The image radar device according to claim 1.   The rotational angular velocity conversion circuit converts the secondary phase coefficient range variation coefficient estimated by the secondary phase coefficient variation estimation circuit into the rotational angular velocity, and outputs the converted result as the target equivalent rotational motion angular velocity. The image radar apparatus according to claim 1, wherein:   The second-order phase coefficient candidate evaluation index calculation circuit generates an image generated by an array in the first half of a range history hit that has been compensated for blurring in the range direction by a rotation range cell movement compensation circuit and an array in the second half of the hit 6. The image radar apparatus according to claim 4, wherein a cross-correlation value between images to be calculated is calculated as a secondary phase coefficient evaluation value distribution.   The zero-crossing range estimation circuit is a coordinate on the other axis when passing through a position where the inclination of the straight line estimated by the locus inclination specifying circuit is a and the coordinate along one axis of the preset image is 0 2 is a two-dimensional image represented by the secondary phase coefficient evaluation value distribution output from the secondary phase coefficient candidate evaluation circuit based on the slope a of the straight line when x is 0 when represented by y = a (x + x0) Correction is made so that the slope of the upper trajectory becomes zero, and the two-dimensional image after correction is summed with the other one axis, and the range where the secondary phase coefficient is zero is estimated from the total value of each y The image radar device according to claim 7 or 8, wherein   While radiating a high-frequency signal toward the space, a part of the high-frequency signal reflected by the target existing in the space is received, and the resolution in the range direction of the received signal is subjected to range compression, and the received signal A range history acquisition circuit that acquires a range history that is a time history of the range profile by repeating a series of processes for acquiring a range profile that is an amplitude characteristic of the range direction while changing the relative position between the radar and the target, and The range history acquired by the range history acquisition circuit compensates for an unnecessary distance change between the radar and the target caused by unnecessary translation movement, and the distance change is compensated by the translation movement compensation circuit. In the range history, at each reflection point of the high-frequency signal at the above target, the range direction generated by the rotational motion The phase term of the data string in the hit direction of each range is specified in the range history in which the blur in the range direction is compensated by the rotational range cell movement compensation circuit and the rotational range cell movement compensation circuit. The likelihood of a plurality of candidate values that are candidates for the secondary phase coefficient that is the coefficient of the secondary phase change included in each is evaluated for each range cell, and the distribution in the range direction of the evaluation result of each candidate value is output. Based on the secondary phase coefficient candidate evaluation circuit and the secondary phase coefficient evaluation value distribution that is the distribution in the range direction output from the secondary phase coefficient candidate evaluation circuit, the change coefficient of the secondary phase coefficient with respect to the range. Based on the secondary phase coefficient range variation coefficient estimated by the secondary phase coefficient variation estimation circuit and the secondary phase coefficient variation estimation circuit that estimates the secondary phase coefficient range variation coefficient Rotational angular velocity conversion circuit that estimates the angular velocity of the target equivalent rotational motion, and secondary phase coefficient of each range is estimated based on the secondary phase coefficient evaluation value distribution output from the secondary phase coefficient candidate evaluation circuit. The zero-crossing range estimation circuit that outputs the range in which the secondary phase coefficient is closest to 0 among the secondary phase coefficients of each range, and the range output from the zero-crossing range estimation circuit is zero. The zero-order range compensation circuit that compensates the range history in which the distance change is compensated by the translation motion compensation circuit, and the compensation by the zero-order range compensation circuit based on the angular velocity estimated by the rotational angular velocity conversion circuit. An image radar apparatus including a radar image generation circuit that generates a radar image from a later range history.   While radiating a high-frequency signal toward the space, a part of the high-frequency signal reflected by the target existing in the space is received, and the resolution in the range direction of the received signal is subjected to range compression, and the received signal A range history acquisition circuit that acquires a range history that is a time history of the range profile by repeating a series of processes for acquiring a range profile that is an amplitude characteristic of the range direction while changing the relative position between the radar and the target, and The range history acquired by the range history acquisition circuit compensates for an unnecessary distance change between the radar and the target caused by unnecessary translation movement, and the distance change is compensated by the translation movement compensation circuit. In the range history, at each reflection point of the high-frequency signal at the above target, the range direction generated by the rotational motion The phase term of the data string in the hit direction of each range is specified in the range history in which the blur in the range direction is compensated by the rotational range cell movement compensation circuit and the rotational range cell movement compensation circuit. The likelihood of a plurality of candidate values that are candidates for the secondary phase coefficient that is the coefficient of the secondary phase change included in each is evaluated for each range cell, and the distribution in the range direction of the evaluation result of each candidate value is output. Based on the secondary phase coefficient candidate evaluation circuit and the secondary phase coefficient evaluation value distribution that is the distribution in the range direction output from the secondary phase coefficient candidate evaluation circuit, the change coefficient of the secondary phase coefficient with respect to the range. Based on the secondary phase coefficient range variation coefficient estimated by the secondary phase coefficient variation estimation circuit and the secondary phase coefficient variation estimation circuit that estimates the secondary phase coefficient range variation coefficient Referring to the rotational angular velocity conversion circuit that estimates the angular velocity of the target equivalent rotational motion and the range history in which the distance change is compensated by the translation motion compensation circuit, the range that has the smallest range movement caused by the rotational motion is determined. A zero-crossing range estimation circuit for estimation, and a zero-order range compensation circuit for compensating for the range history in which the distance change is compensated by the translational motion compensation circuit so that the range estimated by the zero-crossing range estimation circuit becomes zero An image radar apparatus comprising: a radar image generation circuit that generates a radar image from a range history after compensation by the zero-order range compensation circuit based on the angular velocity estimated by the rotation angular velocity conversion circuit.   While radiating a high-frequency signal toward the space, a part of the high-frequency signal reflected by the target existing in the space is received, and the resolution in the range direction of the received signal is subjected to range compression, and the received signal A range history acquisition circuit that acquires a range history that is a time history of the range profile by repeating a series of processes for acquiring a range profile that is an amplitude characteristic of the range direction while changing the relative position between the radar and the target, and The range history acquired by the range history acquisition circuit compensates for an unnecessary distance change between the radar and the target caused by unnecessary translation movement, and the distance change is compensated by the translation movement compensation circuit. In the range history, at each reflection point of the high-frequency signal at the above target, the range direction generated by the rotational motion The phase term of the data string in the hit direction of each range is specified in the range history in which the blur in the range direction is compensated by the rotational range cell movement compensation circuit and the rotational range cell movement compensation circuit. The likelihood of a plurality of candidate values that are candidates for the secondary phase coefficient that is the coefficient of the secondary phase change included in each is evaluated for each range cell, and the distribution in the range direction of the evaluation result of each candidate value is output. Identifying the slope and intercept of the trajectory on the two-dimensional image indicated by the secondary phase coefficient candidate evaluation circuit and the secondary phase coefficient evaluation value distribution that is the distribution in the range direction output from the secondary phase coefficient candidate evaluation circuit; A secondary phase coefficient range change coefficient that is a change coefficient with respect to the range of the secondary phase coefficient is estimated from the slope and intercept of the locus, and a range where the secondary phase coefficient is 0 is estimated. A second-order phase coefficient estimating circuit, and a rotational angular velocity conversion circuit for estimating an angular velocity of the target equivalent rotational motion based on the second-order phase coefficient range change coefficient estimated by the second-order phase coefficient estimating circuit; Estimated by a zero-order range compensation circuit that compensates for a range history whose distance change has been compensated by the translational motion compensation circuit and the rotational angular velocity conversion circuit so that the range estimated by the second-order phase coefficient estimation circuit becomes zero. An image radar apparatus comprising: a radar image generation circuit that generates a radar image from a range history after compensation by the zeroth-order range compensation circuit based on the angular velocity that has been obtained.   While radiating a high-frequency signal toward the space, a part of the high-frequency signal reflected by the target existing in the space is received, and the resolution in the range direction of the received signal is subjected to range compression, and the received signal A range history acquisition circuit that acquires a range history that is a time history of the range profile by repeating a series of processes for acquiring a range profile that is an amplitude characteristic of the range direction while changing the relative position between the radar and the target, and The range history acquired by the range history acquisition circuit compensates for an unnecessary distance change between the radar and the target caused by unnecessary translation movement, and the distance change is compensated by the translation movement compensation circuit. In the range history, at each reflection point of the high-frequency signal at the above target, the range direction generated by the rotational motion The phase term of the data string in the hit direction of each range is specified in the range history in which the blur in the range direction is compensated by the rotational range cell movement compensation circuit and the rotational range cell movement compensation circuit. The likelihood of a plurality of candidate values that are candidates for the secondary phase coefficient that is the coefficient of the secondary phase change included in each is evaluated for each range cell, and the distribution in the range direction of the evaluation result of each candidate value is output. Based on the secondary phase coefficient candidate evaluation circuit and the secondary phase coefficient evaluation value distribution that is the distribution in the range direction output from the secondary phase coefficient candidate evaluation circuit, the change coefficient of the secondary phase coefficient with respect to the range. Based on the secondary phase coefficient range variation coefficient estimated by the secondary phase coefficient variation estimation circuit and the secondary phase coefficient variation estimation circuit that estimates the secondary phase coefficient range variation coefficient Rotation angular velocity conversion circuit for estimating the angular velocity of the target equivalent rotational motion and the inclination of the locus on the two-dimensional image indicated by the secondary phase coefficient evaluation value distribution output from the secondary phase coefficient candidate evaluation circuit are specified. A zero-crossing range estimation circuit for estimating the zero-crossing range of the locus from the inclination of the locus, a secondary phase coefficient range variation coefficient estimated by the secondary phase coefficient variation estimation circuit, and the zero-crossing range estimation circuit Is calculated from the zero-crossing range estimated by the above equation, and the secondary phase change of each range in the range history after compensation by the rotational range cell movement compensation circuit is calculated according to the compensation amount. A secondary phase compensation circuit for compensation and a range doppler by Fourier-transforming the range history after compensation by the secondary phase compensation circuit in the hit direction Based on the angular velocity of the rotational motion estimated by the radar image generation circuit that generates the image and the rotational angular velocity conversion circuit, the physical length of the Doppler frequency axis on the range Doppler image generated by the radar image generation circuit An image radar apparatus comprising: a cross range scaling circuit that performs cross range scaling for converting to a cross range axis.   In the zero crossing range estimation circuit, the inclination of the locus on the two-dimensional image is a, and the coordinate on the other axis when passing through the position where the coordinate along the one axis of the preset image is 0 is y = When x0 is represented by a (x + x0), there are various candidates for x0 with a slope of a on the two-dimensional image indicated by the secondary phase coefficient evaluation value distribution output from the secondary phase coefficient candidate evaluation circuit. A plurality of different integration paths, and performing line integration along each integration path, and estimating that the candidate for x0 having the maximum value of the line integration is the zero-crossing range of the trajectory. The image radar device according to claim 17, characterized in that: The zero-crossing range estimation circuit is a coordinate on one axis when passing through a position where the inclination of a straight line representing the locus on the two-dimensional image is a and the coordinate along one axis of the preset image is 0 2 is a two-dimensional image represented by the secondary phase coefficient evaluation value distribution output from the secondary phase coefficient candidate evaluation circuit based on the slope a of the straight line when x is 0 when represented by y = a (x + x0) A correction is made so that the slope of the upper trajectory becomes 0, and the corrected two-dimensional image is summed with the other one axis, and a range in which the secondary phase coefficient is 0 from the total value of each y, The image radar device according to claim 17, wherein the image radar device is estimated as a zero-crossing range of the trajectory.   The zero-crossing range estimation circuit does not specify the slope of the locus on the two-dimensional image indicated by the secondary phase coefficient evaluation value distribution output from the secondary phase coefficient candidate evaluation circuit, and adds the secondary phase coefficient evaluation value distribution to the secondary phase coefficient evaluation value distribution. Based on this, the secondary phase coefficient of each range is estimated, and among the secondary phase coefficients of each range, the range where the secondary phase coefficient is closest to 0 is estimated as the zero-crossing range of the trajectory. The image radar device according to claim 17, characterized in that:   The zero-crossing range estimation circuit compensates for the change in distance by the translational compensation circuit without specifying the inclination of the locus on the two-dimensional image indicated by the distribution of evaluation values of the secondary phase coefficient output from the candidate evaluation circuit of the secondary phase coefficient. 18. The image radar apparatus according to claim 17, wherein a range having the smallest range movement caused by a rotational motion is estimated as a zero-crossing range of the locus with reference to the range history. While radiating a high-frequency signal toward the space, a part of the high-frequency signal reflected by the target existing in the space is received, and the resolution in the range direction of the received signal is subjected to range compression, and the received signal A range history acquisition circuit that acquires a range history that is a time history of the range profile by repeating a series of processes for acquiring a range profile that is an amplitude characteristic of the range direction while changing the relative position between the radar and the target, and The range history acquired by the range history acquisition circuit compensates for an unnecessary distance change between the radar and the target caused by unnecessary translation movement, and the distance change is compensated by the translation movement compensation circuit. In the range history, at each reflection point of the high-frequency signal at the above target, the range direction generated by the rotational motion The phase term of the data string in the hit direction of each range is specified in the range history in which the blur in the range direction is compensated by the rotational range cell movement compensation circuit and the rotational range cell movement compensation circuit. The likelihood of a plurality of candidate values that are candidates for the secondary phase coefficient that is the coefficient of the secondary phase change included in each is evaluated for each range cell, and the distribution in the range direction of the evaluation result of each candidate value is output. Identifying the slope and intercept of the trajectory on the two-dimensional image indicated by the secondary phase coefficient candidate evaluation circuit and the secondary phase coefficient evaluation value distribution that is the distribution in the range direction output from the secondary phase coefficient candidate evaluation circuit; A secondary phase coefficient range change coefficient that is a change coefficient with respect to the range of the secondary phase coefficient is estimated from the slope and intercept of the locus, and a range where the secondary phase coefficient is 0 is estimated. A second-order phase coefficient estimating circuit, and a rotational angular velocity conversion circuit for estimating an angular velocity of the target equivalent rotational motion based on the second-order phase coefficient range change coefficient estimated by the second-order phase coefficient estimating circuit; from range secondary phase factor estimated by the estimated by the secondary phase coefficient estimation circuit secondary phase coefficient range change coefficient and the secondary phase coefficient estimation circuit is 0, the secondary of each range A secondary phase compensation circuit that calculates a compensation amount that cancels the phase change and compensates a secondary phase change of each range in the range history after compensation by the rotation range cell movement compensation circuit according to the compensation amount, and the secondary phase compensation A radar image generation circuit that generates a range Doppler image by Fourier transforming the range history compensated by the circuit in the hit direction, and the rotation angular velocity conversion circuit. A cross range that performs cross range scaling that converts the Doppler frequency axis on the range Doppler image generated by the radar image generation circuit into a cross range axis that is a physical length based on the angular velocity of the determined rotational motion. An image radar apparatus comprising a scaling circuit.

Images