JP2014194398A - Radar data processing device, radar data processing method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accurately measure a Doppler speed of an observation object.SOLUTION: A hypothesis creation section 108 creates a new hypothesis on the basis of an existing hypothesis. The hypothesis is formed from a plurality of peak trajectories, and a new peak trajectory can be obtained by connecting an observation value of the present time or a prediction value to an existing peak trajectory in a time-sequential manner. The prediction value is calculated by applying a predetermined change model to the existing peak trajectory. A hypothesis evaluation section 109 calculates likelihood of the change model on the basis of the prediction value and the observation value, calculates hypothesis reliability that is reliability of the hypothesis on the basis of the likelihood of the change model applied to the peak trajectories forming the hypothesis, and selects a best hypothesis that is a most satisfactory hypothesis. A primary scattering estimation section 110 determines a peak trajectory corresponding to primary scattering of Bragg scattering on the basis of sources of the peak trajectory from among the peak trajectories included in the best hypothesis and calculates a Doppler speed of an observation object.

Description

  The present invention relates to a radar data processing device, a radar data processing method, and a program.

  The marine radar system is a system for observing the velocity of the ocean surface current. The ocean radar system receives a radio wave scattered backward from the radio wave irradiated to the sea surface from the ocean radar, and detects the primary scattering of the Bragg scattering from the received wave, so that the Doppler velocity corresponding to the surface flow velocity, that is, Measure the gaze direction flow velocity. The first-order scattering of the Bragg scattering is detected as a peak having the maximum received intensity in the Doppler spectrum.

  For example, the peak detected from the received wave may not correspond to the primary scattering of the Bragg scattering, such as when the received intensity due to disturbances such as unnecessary waves and interference waves caused by ships, aircraft, etc. becomes maximum in the Doppler spectrum. is there. As a result, due to the influence of disturbance on the received waves, the surface velocity of the ocean surface may be measured incorrectly, degrading the observation quality of the ocean radar system.

  For example, Patent Document 1 discloses a marine radar that can remove the influence of unnecessary waves when measuring the Doppler velocity. This ocean radar supplies the median filter with velocity estimation data for multiple ranges aligned in the direction of radio wave extraction, extracts the representative value of the ocean current velocity component, and the velocity that exceeds a predetermined threshold from this representative value. Remove estimation data.

JP 2002-323558 A

  According to the marine radar described in Patent Document 1, if the Doppler velocity at the observation point affected by the unnecessary signal is significantly different from the Doppler velocity at the surrounding observation points, the influence of the unnecessary signal can be removed. it can.

  However, when the Doppler spectrum has a peak Doppler frequency that is close to that due to an unnecessary signal and that corresponding to primary scattering due to Bragg scattering, the marine radar described in Patent Document 1 makes it difficult to distinguish between the two. There is.

  Further, when an abrupt change that differs from the surrounding observation points occurs due to an earthquake that occurs inside or outside the observation range, the change is due to an unnecessary signal in the marine radar described in Patent Document 1. As a result, there is a risk of being excluded from the speed estimation data.

  The present invention has been made in view of the above-described circumstances, and an object thereof is to provide a radar data processing apparatus and the like that can accurately measure the Doppler velocity of an observation target.

  To achieve the above object, a radar data processing apparatus according to the present invention acquires data indicating a Doppler spectrum for each observation point, and extracts peak information including an observation value that is a frequency corresponding to a peak in the Doppler spectrum. Whether there is a correlation between the observed value and the predicted value, which is the peak position predicted by applying a predetermined change model to the peak information extraction unit and the peak trajectory that connects the peaks of past observations in time series A hypothesis memory that stores hypothesis information indicating a hypothesis created by a combination of peak trajectories, and the current observed value or predicted value in the peak trajectory included in the hypothesis indicated by the hypothesis information. It includes a consistent combination of peak trajectories obtained by connecting to a series, and the correlation determination unit determines that there is a correlation. A hypothesis generator that creates a hypothesis that includes a peak trajectory that is connected to the peak trajectory corresponding to the predicted value in time series with the observed value determined to be correlated with the predicted value, and based on the predicted value and the observed value The likelihood of the change model is calculated, and the hypothesis reliability, which is the reliability of the hypothesis, is calculated based on the likelihood of the change model applied to the peak locus included in the hypothesis created by the hypothesis creation unit. A hypothesis evaluation unit that selects the best hypothesis having the best reliability and a peak locus corresponding to the primary scattering of Bragg scattering based on the specifications of the peak locus from the peak locus included in the best hypothesis. And a primary scattering estimator for calculating the Doppler velocity of the observation target.

  According to the present invention, it is possible to accurately measure the Doppler velocity of an observation target.

It is a block diagram which shows the structure of the radar data processing apparatus which concerns on Embodiment 1 of this invention. It is a figure which shows an example of the hypothesis information which the hypothesis memory has memorize | stored. It is a figure which shows an example of the determination result by a correlation determination part. FIG. 3 is a diagram illustrating an example of a child hypothesis created using a hypothesis included in the hypothesis information illustrated in FIG. 2 as a parent hypothesis. 3 is a flowchart showing a flow of radar data processing according to the first embodiment. It is a flowchart which shows the flow of hypothesis reliability calculation processing. It is a flowchart which shows the flow of hypothesis reliability calculation processing. The Doppler spectrum of ocean radar is classified into high / medium / low reliability using the fact that the spectrum of the observation point with good S / N appears symmetrically in the positive frequency region and the negative frequency region. It is a figure which shows the example of an observation in an observation point. It is a block diagram which shows the structure of the radar data processing apparatus which concerns on Embodiment 2 of this invention. It is a block diagram which shows the structure of the radar data processing apparatus which concerns on Embodiment 3 of this invention.

  Embodiments of the present invention will be described below with reference to the drawings. The same elements are denoted by the same reference symbols throughout the drawings.

Embodiment 1 FIG.
A radar data processing apparatus 100 according to Embodiment 1 of the present invention is an apparatus provided in an ocean radar system that observes the flow velocity of a surface current in the ocean, and is based on a Doppler spectrum that is spectral data by Bragg scattering. The Doppler velocity of the surface flow at the observation point is calculated continuously.

  As shown in FIG. 1, the radar data processing apparatus 100 according to the present embodiment is connected to a signal processor 102 connected to a receiver 101 and a display 103. The radar data processing apparatus 100 functionally includes a change model memory 104, a hypothesis memory 105, a peak information extraction unit 106, a correlation determination unit 107, a hypothesis creation unit 108, a hypothesis, as shown in FIG. An evaluation unit 109 and a primary scattering estimation unit 110 are provided.

  The receiver 101 is composed of, for example, an antenna. When receiving a radio wave such as a reflected wave of a radio wave transmitted from a transmitter (not shown) to the sea surface, the receiver 101 detects an IQ (In-phase Quadrature) signal (digital signal) after phase detection. Convert and output.

  The signal processor 102 performs processing such as FFT (Fast Fourier Transform) and DBF (Digital Beam Forming) on the signal output from the receiver 101, thereby outputting Doppler spectrum data indicating the Doppler spectrum for each observation point. .

  In a normal case where the electric field received by the receiver 101 is not affected by disturbance, in the Doppler spectrum, the reflected wave satisfying the Bragg scattering resonance condition, that is, the power corresponding to the frequency of the primary scattered reflected wave (reception intensity) ) Form a peak. The disturbance is an unnecessary wave, an interference wave, or the like caused by a navigating ship or aircraft.

  The display device 103 is a device that acquires output data from the radar data processing device 100 and displays an image corresponding to the output data, and is, for example, a liquid crystal display.

  The change model memory 104 stores change model information in advance. The change model information includes one or more change models. In the change model information according to the present embodiment, one or more change models are associated with each observation point.

  The change model is a model related to the temporal change of the Doppler velocity, and corresponds to a motion model of a frequency (peak position) corresponding to a peak in the Doppler spectrum.

  The hypothesis memory 105 stores hypothesis information indicating a hypothesis regarding the surface flow. Hypothesis information is updated each time an observation is made. The hypothesis memory 105 further stores reliability information indicating the posterior reliability (details will be described later) for each change model.

  The hypothesis is composed of a plurality of peak trajectories. The peak trajectory is a trajectory obtained by linking time-series peak positions observed or predicted at each past observation. For the peak locus, for example, either the observed peak position (observed value) or the predicted peak position (predicted value) is selected for each past observation point, and the selected peak position is smoothed by the tracking filter. Created by concatenating in the time direction. For example, a Kalman filter may be employed as the tracking filter. When the nonlinearity of the change model is strong, a nonlinear Kalman filter such as an extended Kalman filter may be employed as the tracking filter.

Here, an example of a hypothesis indicated by the hypothesis information updated at the time of observation at time t2 will be described with reference to FIG. This figure, the previous observation time (time t1), observed or expected peak position Z _{point arrives} and Z _{T1_2} is, during the most recent observation (time t2), the peak position Z _{point arrives} and Z _{T2_2} were observed or predicted An example is shown. Peak position Z _{point arrives} and Z _{T2_2} correspond to the peak due to surface current is the observation target, the peak position Z _{T1_2} and Z _{point arrives} are vessels, and corresponds to the peak due to unwanted signals caused by such navigation of aircraft To do.

For example, at time t2, the peak trajectories are PT1 (t2) = {Z _{t1_2} , Z _{t2_1} }, PT2 (t2) = {Z _{t1_1} , Z _{t2_1} }, PT3 (t2) = {Z _{t1_2} , Z _{t2_2} }, PT4 (t2) = {Z _{t1_1} , Z _{t2_2} } is created, and hypotheses are H1 (t2) = {PT1 (t2), PT4 (t2)}, H2 (t2) = {PT2 (t2), PT3 (t2) } Is created. Here, {} in the peak trajectory indicates a set of peak positions adopted in each observation, and {} in the hypothesis indicates a set of peak trajectories constituting the hypothesis.

  The peak information extraction unit 106 calculates a noise level for the Doppler spectrum at each observation point received from the signal processor 102, and extracts the Doppler frequency corresponding to the power greater than the noise by a certain value or more as the peak position. Further, the peak information extraction unit 106 calculates the power at the extracted peak position by converting it into S / N (signal / noise). The peak information extraction unit 106 outputs peak information including the extracted peak position and the calculated S / N.

  Correlation determining section 107 determines whether or not there is a correlation between the peak locus included in the hypothesis information stored in hypothesis memory 105 and the current observed value (for example, time t3 after time t2). The correlation determination unit 107 outputs correlation determination information including the determination result.

  Specifically, the correlation determination unit 107 reads the change model information stored in the change model memory 104 and the hypothesis information stored in the hypothesis memory 105. The correlation determination unit 107 calculates a predicted value at the time of the current observation by applying each of the change models included in the change model information to each peak locus included in the hypothesis information. The correlation determination unit 107 determines whether or not there is a correlation between the peak locus and the current observation value based on the calculated predicted value and the current observation value.

  Here, whether or not there is a correlation between the peak locus and the observed value is determined by calculating a value (degree of correlation) indicating the degree of correlation between the peak locus and the observed value, and comparing the degree of correlation with a threshold value. The

  For the correlation degree, for example, the Mahalanobis distance obtained by the calculation represented by the equation (1) may be employed. For example, when the Mahalanobis distance is smaller than a predetermined threshold value d, the correlation determination unit 107 may determine that “there is correlation”. When the Mahalanobis distance is equal to or greater than the threshold value d, the correlation determination unit 107 may determine “no correlation”.

・・・（１）

... (1)

Here, z is the observed value this time. x _{k, a (−)} is a predicted value (individual predicted value) for each change model. S _a is a residual covariance matrix (individual residual covariance matrix) for each change model. H _a is an observation matrix for each change model.

  By executing such correlation determination processing, the correlation determination unit 107 outputs correlation determination information including, for example, the determination result shown in FIG.

Here, FIG. 3 shows that the current observed value Z _{t3_1} has no correlation with the peak trajectories PT1 (t2) and PT4 (t2) and has a correlation with the peak trajectories PT2 (t2) and PT3 (t2). Indicates that In addition, the figure shows that the observed value Z _{t3_2} has no correlation with the peak trajectory PT1 (t2), PT2 (t2) and PT3 (t2), and has a correlation with the peak trajectory PT4 (t2). It shows that.

  The hypothesis creation unit 108 creates a hypothesis (child hypothesis) based on all combinations of the current observation value and the hypothesis (parent hypothesis) indicated by the hypothesis information stored in the hypothesis memory 105. The hypothesis creation unit 108 stores hypothesis information indicating the created child hypothesis in the hypothesis memory 105, and thereby updates the hypothesis information in the hypothesis memory 105.

Specifically, the hypothesis creating unit 108 sets a combination of a parent hypothesis and a current observation value that are not inconsistent with the current observation value among all combinations of the current observation value and the parent hypothesis as a child hypothesis. Here, “no contradiction” means that a peak trajectory including the same peak position is not included in one hypothesis, in other words, a peak trajectory included in one hypothesis is the same in the past and present. It means not sharing the peak. For example, the combinations of peak trajectories at time t2 include {PT1 (t2), PT2 (t2)}, {PT1 (t2), PT3 (t2)}, {PT1 (t2), PT4 (t2)}, {PT2 There are six types (t2), PT3 (t2)}, {PT2 (t2), PT4 (t2)}, and {PT3 (t2), PT4 (t2)}. {PT1 (t2), PT2 (t2)} share Z _{t2_1} and {PT1 (t2), PT3 (t2)} share Z _{t1_2} . {PT2 (t2), PT4 (t2) } Is a point that shares Z _{t1_1} , and {PT3 (t2), PT4 (t2)} are points that share Z _{t2_2} , which are contradictory combinations. Therefore, at time t2, as described above, H1 (t2) = {PT1 (t2), PT4 (t2)} and H2 (t2) = {PT2 (t2), PT3 (t2)} are created as hypotheses. Is done.

  Further, the hypothesis creating unit 108 creates a child hypothesis including at least one peak locus created by connecting the peak locus determined to be correlated by the correlation determining unit 107 and the current observation value in the time direction. In other words, the combination of the peak trajectory determined by the correlation determination unit 107 as having no correlation for all the change models and the current observation value is excluded from the hypothesis. As a result, an increase in the number of hypotheses can be suppressed.

Here, when the parent hypothesis and the peak trajectory described with reference to FIG. 2 are stored in the hypothesis memory 105, the hypothesis creation unit 108 _compares the present observation values Z _{t3_1} and Z _{t3_2} with the correlation shown in FIG. An example of a child hypothesis created based on the determination result will be described with reference to FIG.

Based on the parent hypothesis H1 (t2) = {PT1 (t2), PT4 (t2)}, the peak trajectory PT4 (t2) and the judgment result that there is a correlation between the observed value Z _{t3_2} , the child hypothesis H1 (t3) = {PT1 (t3), PT4 (t3)} is created. Based on the parent hypothesis H2 (t2) = {PT2 (t2), PT3 (t2)} and the determination result that there is a correlation between the peak trajectory PT2 (t2) and the observed value Z _{t3_1} , the child hypothesis H2 (t3) = {PT2 (t3), PT3 (t3)} is created. Based on the parent hypothesis H2 (t2) = {PT2 (t2), PT3 (t2)}, the peak trajectory PT3 (t2) and the determination result that the observation value Z _{t3_1} is correlated, the child hypothesis H2_1 (t3) = {PT2_1 (t3), PT3_1 (t3)} is created.

Here, the peak trajectories at time t3 are PT1 (t3) = {Z _{t1_2} , Z _{t2_1} , φ}, PT2 (t3) = {Z _{t1_1} , Z _{t2_1} , Z _{t3_1} }, PT3 (t3) = {Z _{t1_2} , Z _{t2_2} , φ}, PT4 (t2) = {Z _{t1_1} , Z _{t2_2} , Z _{t3_2} }, PT2_1 (t3) = {Z _{t1_1} , Z _{t2_1} , φ}, PT3_1 (t3) = {Z _{t1_2} , Z _{t2_2} , Z _{t3_1} }.

  The hypothesis evaluation unit 109 calculates, for each hypothesis created by the hypothesis creation unit 108, a hypothesis reliability that is a value indicating the reliability of the hypothesis based on the reliability information stored in the hypothesis memory 105. The hypothesis evaluation unit 109 determines the best hypothesis that is the most reliable hypothesis based on the calculated hypothesis reliability. The hypothesis evaluation unit 109 outputs best hypothesis information including the determined best hypothesis and its hypothesis reliability.

  When the primary scattering estimation unit 110 acquires the best hypothesis information, the primary scattering estimation unit 110 determines the peak locus corresponding to the primary scattering from the peak locus constituting the best hypothesis at each observation point based on the specifications of the peak locus. In the determination of the peak locus corresponding to the primary scattering, for example, the peak locus having the longest existence period of the peak locus and the existence period being equal to or longer than the predetermined period is determined as corresponding to the primary scattering. Good.

  The first-order scattering estimation unit 110 converts the Doppler frequency corresponding to the peak position of the first-order scattering at each observation point into a Doppler velocity that is a visual line direction velocity. The first-order scattering estimation unit 110 outputs output data including the Doppler velocity obtained by the conversion and the hypothesis reliability of the best hypothesis to the display 103.

  In general, when an object that causes an unnecessary signal, such as a ship or an aircraft near an observation point, is moving, it often moves in a substantially constant direction. For this reason, it rarely remains within the range corresponding to one observation point. Therefore, the peak locus having the longest existence period and not less than a predetermined period is likely not to be an unnecessary signal but to correspond to primary scattering. It is possible to accurately measure the Doppler velocity of the surface layer flow by determining that such a peak locus corresponds to the first-order scattering and calculating the Doppler velocity based on the result.

  So far, the configuration of the radar data processing apparatus 100 according to Embodiment 1 of the present invention has been described. From here, operation | movement of the radar data processing apparatus 100 is demonstrated with reference to figures.

  When Doppler spectrum data is acquired from the signal processor 102, the radar data processing device 100 executes radar data processing shown in FIG.

  In general, since the marine radar system continuously irradiates the surface of the radar with the radar during operation, Doppler spectrum data is continuously output from the signal processor 102. Meanwhile, radar data processing is repeatedly executed.

  As shown in the figure, when the peak information extraction unit 106 acquires a signal from the signal processor 102, the peak information extraction unit 106 includes a peak position and an S / N corresponding thereto based on the Doppler spectrum of each observation point included in the signal. Peak information is generated (step S101). The peak information extraction unit 106 outputs the generated peak information. Assume that this observation time is time t3.

  The correlation determination unit 107 acquires hypothesis information stored in the hypothesis memory 105. The correlation determination unit 107 determines whether or not there is a correlation between the peak locus included in the acquired hypothesis information and the current observation value (time t3) (step S102). The correlation determination unit 107 outputs correlation determination information including the determination result.

  The hypothesis creation unit 108 acquires hypothesis information stored in the hypothesis memory 105. The hypothesis creating unit 108 creates a hypothesis (child hypothesis) based on all combinations of the hypothesis (parent hypothesis) indicated by the hypothesis information and the current observation value (step S103). The hypothesis creating unit 108 updates the hypothesis information in the hypothesis memory 105 with the hypothesis information indicating the created child hypothesis (step S104).

  The hypothesis evaluation unit 109 calculates a hypothesis reliability for each hypothesis created by the hypothesis creation unit 108 based on the reliability information stored in the hypothesis memory 105 (step S105).

  Here, the flow of hypothesis reliability calculation processing (step S105) will be described with reference to FIGS. 6A and 6B.

  As shown in the figure, the hypothesis evaluation unit 109 repeats the following steps S202 to S209 for all observation points (step S201; loop A). In loop A (step S201), the hypothesis evaluation unit 109 repeats the following steps 203 to S209 for all hypotheses (step S202; loop B). In the loop B (step S202), the hypothesis evaluation unit 109 repeats the following steps S204 to S208 for all peak trajectories (step S203; loop C).

In the loop C (step S203), the hypothesis evaluation unit 109 uses the posterior reliability μ _{k-1, b (+)} indicated by the reliability information in the hypothesis memory 105, and the prior reliability of each change model in the current observation. μ _{k, a (−)} is calculated (step S204). Prior reliability μ _{k, a (-)} is the sum of the product of posterior reliability and transition probability in the previous observation of each change model for all change models, as shown in the following formula (2), for example. Is calculated by

Here, m is the number of change models used. π _ba is a transition probability from the change model b to a set in advance.

The hypothesis evaluation unit 109 calculates the posterior reliability μ _{k, a (+)} of each change model based on the prior reliability μ _{k, a (−)} for all peak trajectories (step S205), and the hypothesis memory 22. Save to The posterior reliability μ _{k, a (+)} is calculated by the following equation (3), for example.

Here, the likelihood ν _{k, a} for each change model follows a multivariate normal distribution with a residual covariance matrix S _a centered on the predicted value x _{k, a (−)} for each change model, It is calculated by the following formula (4).

・・・（４）

... (4)

  The hypothesis evaluation unit 109 calculates a smooth value between the predicted value and peak position of each change model for all peak trajectories (step S206), and stores it in the hypothesis memory 22 for use in prediction at the next observation.

The hypothesis evaluation unit 109 calculates an integrated prediction value x _{k (−)} that is a prediction value integrated for all peak loci based on the prior reliability μ _{k, a (−)} (step S207). Further, the hypothesis evaluation unit 109 calculates an integrated prediction error covariance matrix P _{k (−)} that is a prediction error covariance matrix integrated for all peak trajectories based on the prior reliability μ _{k, a (−).} (Step S208).

Here, the integrated prediction value x _{k (−)} is calculated by the following equation (5), for example, and the integrated prediction error covariance matrix P _{k (−)} is calculated by the following equation (6), for example.

・・・（６）

... (6)

When Steps S204 to S208 are executed for all peak trajectories (Step S203; Loop C), the hypothesis evaluation unit 109 calculates the hypothesis reliability β _i normalized for each observation point (Step S209). The hypothesis reliability β _i is the likelihood of the peak position g _fe for the parent prediction reliability β _{p (i)} calculated in the previous observation and the integrated prediction value x _{k (−)} calculated in step S207. Based on the above, for example, it is calculated by the following equations (7) and (8).

Here, γ _i is the hypothesis reliability before normalization at a certain observation point. L is the number of all hypotheses in this observation. N _pk is the number of peak positions in the current observation correlated with the existing peak trajectory in the parent hypothesis. N _PT is the number of existing peak trajectories in the parent hypothesis. g _{f, s} is a value representing the probability that the peak position corresponds to the peak of the spectrum. g _fe is the likelihood of the peak position with respect to the existing peak trajectory. P _G is a capture probability (parameter).

g _{f, s} is, for example, a probability density function based on a theoretical formula such as a swirling case, and is a value to be calculated with S / N as an input. g _fe is calculated from the integrated residual covariance matrix S calculated from the integrated prediction value x _{k (−)} and the integrated prediction error covariance matrix P _{k (−)} based on the same premise as Expression (4).

  Note that the hypothesis evaluation unit 109 may generate a new peak locus for a peak position that has no correlation with the peak locus, and create a hypothesis including the peak locus, following the past observation. Specifically, for example, the hypothesis evaluation unit 109, for a peak position that has no correlation with the peak trajectory, goes back to the past observation, and the correlation with the peak trajectory is within a certain range set in advance by the same processing as described above. It is determined whether or not there are no peak positions for a certain period. When it is determined that there are peak positions that have no correlation with the peak locus for a certain period or longer, the hypothesis evaluation unit 109 generates a peak locus based on these peak positions, and stores the hypothesis including the peak locus as hypothesis information to the hypothesis memory 105. Remember me.

  When steps S203 to 209 are executed for all hypotheses (step S202; loop B) and steps S202 to 209 are executed for all observation points (step S201; loop A), the hypothesis evaluation unit 109 determines the hypothesis reliability. The calculation process (step S105) ends.

Returning to FIG. 5, the hypothesis evaluation unit 109 determines the best hypothesis that is the most reliable hypothesis for each observation point based on the hypothesis reliability based on the calculated hypothesis reliability (step S106). Specifically, the hypothesis evaluation unit 109 determines the hypothesis corresponding to the maximum hypothesis reliability β _i as the best hypothesis for each observation point. Then, the hypothesis evaluation unit 109 outputs best hypothesis information including the best hypothesis at each observation point and its hypothesis reliability.

  The primary scattering estimation unit 110 acquires the best hypothesis information from the hypothesis evaluation unit 109. Based on the acquired best hypothesis information, the primary scattering estimation unit 110 calculates a Doppler velocity, which is a visual line direction velocity, of the Doppler frequency corresponding to the peak position of the primary scattering at each observation point (step S107). Then, the primary scattering estimation unit 110 outputs output data including the Doppler velocity at each observation point and the hypothesis reliability of the best hypothesis to the display 103.

  When the display 103 acquires the output data from the primary scattering estimation unit 110, the display 103 displays an image that associates the Doppler velocity and hypothesis reliability of each observation point based on the output data (step S108).

  As described above, the radar data processing apparatus 100 according to the present embodiment holds the peak locus hypothesis for all observation points for each observation for a certain period, and selects the best hypothesis for each observation point. Then, a peak locus corresponding to the primary scattering is selected from among them.

  Thus, in the Doppler spectrum, when the peak corresponding to the unwanted signal is close to the peak corresponding to the primary scattering, the peak corresponding to the primary scattering corresponds to the primary scattering even when there is a sudden change. Peaks can be selected appropriately.

  The radar data processing apparatus 100 repeats calculating Doppler velocities at all observation points based on the selected peak locus. Therefore, it is possible to suppress the influence of the unnecessary signal on the observation result and correctly measure the Doppler velocity of the surface flow.

  Conventionally, when the peak position corresponding to the first-order scattering cannot be clearly determined, the observation results of the entire coverage area are interpolated by interpolating the observation results at the adjacent observation points with the observation point data missing. You may get In this case, the extrapolated data is inferior in reliability to the data obtained by actual observation, but the user cannot grasp it. In the present embodiment, the display 103 displays the Doppler speed together with a value indicating its reliability. As a result, the user can quantitatively grasp the reliability of the Doppler speed obtained as a measurement result.

  In the change model memory 104, change model data in which an observation point and a change model are associated with each other is stored, so that a change model can be set for each observation point. Therefore, when the basic temporal change pattern of Doppler velocity differs for each unspecified area due to the influence of topography, environment, weather, etc., a change model suitable for each area is applied to each observation point belonging to each area. It becomes possible to do.

  As mentioned above, although Embodiment 1 of this invention was demonstrated, this Embodiment is not restricted to this, You may deform | transform as follows.

  For example, in the first embodiment, the radar data processing apparatus 100 has been described as being provided in a marine radar system whose surface current is an observation target. However, the radar data processing apparatus 100 is provided in a weather radar system whose observation target is raindrops in the atmosphere. Also good. In this case, the Doppler velocity of raindrops at each observation point is continuously calculated based on the Doppler spectrum as in the first embodiment. Further, when equipped in a weather radar system, typically, an aircraft or the like is a factor of unnecessary signals.

Modification 1
The primary scattering estimation unit 110 further stores the peak trajectory adopted as the previous primary scattering for each observation point, whereby the primary scattering peak trajectory is transferred to another peak trajectory in the current observation. You may detect that.

  For example, for each observation in which the peak trajectory used as the primary scatter in the past was adopted, the best hypothesis that the peak trajectory as the primary scatter exists in this observation is selected again, and the peak trajectory is observed as the primary scatter. Recalculate the results. At the same time, it is displayed on the display device 103 to warn that the past observation result has been recalculated.

  There is actually only one peak due to the first-order scattering of Bragg scattering at a certain observation point. Therefore, it is possible to improve the observation accuracy by reviewing the observation result in which the peak locus that has been assumed to be the primary scattering in the past has been adopted. Further, by displaying a warning, the user can grasp the reliability of observation.

Modification 2
In the Doppler spectrum of ocean radar, waves are composed of both approaching and moving away waves. For this reason, at an observation point with a good S / N, peaks corresponding to the first-order scattering of Bragg scattering appear symmetrically in the positive frequency region and the negative frequency region, and the Doppler shift of the surface flow is similar to the approaching wave. Acts the same for waves that move away. By utilizing this for the identification of the first-order scattering, it becomes possible to accurately measure the Doppler velocity of the observation target.

  The primary scattering estimation unit according to the present modification first determines a peak locus corresponding to the primary scattering as in the first embodiment.

  Next, the primary scattering estimation unit according to this modification determines whether or not there is a correlation between the determined peak trajectory and all peak trajectories existing in the opposite frequency region. In the determination of the presence or absence of this correlation, the primary scattering estimation unit, for example, by using the following equation (9), as a peak trajectory correlation degree indicating the degree of correlation between peak trajectories existing in the frequency regions on the opposite side, A Mahalanobis distance ε between peak trajectories is calculated.

  Here, d is a residual between peak trajectories. D is a residual covariance matrix between peak trajectories.

  The primary scattering estimation unit according to the present modification compares the calculated peak locus correlation with a predetermined threshold E. Based on the comparison result and the peak trajectory corresponding to the determined primary scattering, the primary scattering estimation unit sets the reliability of each observation point in descending order of reliability, for example, a high reliability observation point, Classify into medium and low reliability stations.

  Here, with reference to FIG. 7 showing an example of observation of each reliability, the reliability classification processing by the primary scattering estimation unit according to the present modification will be described. In the figure, the Doppler spectrum at this observation point is shown in the upper row, and the time change of the peak trajectory until this observation is shown in the lower row. In addition, in order from the left, the Doppler spectrum and the peak trajectory of the high reliability observation point, the medium reliability observation point, and the low reliability observation point are shown.

  Referring to the Doppler spectrum shown in the upper left of the figure, f0 is the Bragg scattering frequency of the sea surface, and since this observation point has a good S / N, it is roughly in each of the positive frequency region and the negative frequency region. A peak appears symmetrically. Here, the flow of the sea surface, that is, the Doppler shift fd due to the surface current is generated as a difference from f0. Corresponding changes in time of the peak trajectories PT1 and PT2 up to the present observation follow the same trajectory as shown in the lower left of the figure. Therefore, the Mahalanobis distance ε of the peak trajectories PT1 and PT2 is larger than the threshold value E. As a result, it is determined that there is a correlation between the peak trajectories of the positive frequency region and the negative frequency region, and this observation point is a highly reliable observation point in the sense that the reliability of the determined primary scattering is high. Classified.

  Referring to the Doppler spectrum shown in the upper middle of the figure, peaks appear in each of the positive frequency region and the negative frequency region, but they do not appear symmetrically. Corresponding changes over time in the peak trajectories PT3 and PT4 up to the present observation are not similar to the trajectories as shown in the lower middle of the figure. Follow a similar trajectory. Therefore, the Mahalanobis distance ε of the peak trajectories PT1 and PT2 is equal to or less than the threshold value E. As a result, it is determined that there is no correlation between the peak trajectories of the positive frequency region and the negative frequency region, and this observation point is a medium in the sense that the primary scattering is determined but the reliability is not high. Classified as a reliability observation point.

  Referring to the Doppler spectrum shown in the upper right of the figure, the peak appears only in one frequency region (negative frequency region in the figure), and the other frequency region peak does not appear. The peak locus PT5 up to this observation corresponding to this has a short existence period as shown in the lower right of the figure. Therefore, this observation point is classified as a low-reliability observation point because there is no peak locus that is the target of primary scattering.

  By displaying the classification of reliability of each observation point determined in this way on the display 103, the user relates to the Doppler velocity obtained as a measurement result, for example, the reliability including the status of the received signal by the marine radar. It becomes possible to grasp the sex.

  Furthermore, for observation points in a bad environment where there are many peak trajectories, it is recommended that hypotheses that consider all combinations are considered for the peak track in the positive frequency domain and the peak trajectory in the negative frequency domain. . If the pair of peak trajectories, which is the best hypothesis, does not change over a certain observation period, this can be selected as primary scattering. Since it is unlikely that the unnecessary signal exhibits the same behavior as the peak locus in the positive frequency region and the negative frequency region, the system is more resistant to the unnecessary signal. As a result, it is possible to accurately measure the Doppler velocity of the observation target.

Embodiment 2. FIG.
Unlike the radar data processing apparatus 100 according to the first embodiment, the radar processing apparatus according to the present embodiment refers to unnecessary signal information regarding an aircraft, a ship, or the like that causes an unnecessary signal. As a result, the peak trajectory correlated with the peak trajectory of the unnecessary signal is excluded from the peak trajectory candidates corresponding to the primary scattering. As a result, it is possible to provide a radar processing apparatus that is robust against unwanted waves.

  The radar data processing device 200 according to the second embodiment of the present invention is connected to an external monitoring system 211 as shown in FIG. 8 in addition to the configuration provided in the radar data processing device 100 according to the first embodiment. Functionally, an unnecessary signal peak locus generation unit 212 is provided. The radar data processing apparatus 200 includes a primary scattering estimation unit 210 that replaces the primary scattering estimation unit 110.

  The external monitoring system 211 outputs unnecessary signal information including one or more of the position and speed information of the aircraft, ship, etc. that cause the unnecessary signal to the radar data processing device 200. The unnecessary signal information may include error information determined based on the detected sensor specifications.

  When the unnecessary signal peak trajectory generating unit 212 acquires the unnecessary signal information from the external monitoring system 211, the unnecessary signal peak trajectory generating unit 212 converts the unnecessary signal information into a peak trajectory (unnecessary signal peak trajectory) of the unnecessary signal at each observation point. The unnecessary signal peak locus generation unit 212 generates unnecessary signal peak locus information including unnecessary signal peak locus obtained by the conversion.

  In this conversion process, the unnecessary signal peak trajectory generation unit 212 may estimate a range that is assumed to be affected by the unnecessary signal to a large extent, and generate an unnecessary signal peak trajectory for observation points that exist within the range.

  In general, depending on the size and shape of the target that is the source of the unwanted signal, due to the influence of the side lobe, etc., not only the observation point that includes the unwanted signal position, but also the neighboring observation points are considered as unwanted signals as Doppler. It may appear on the spectrum. By estimating a range that is assumed to be affected by unnecessary signals to a large extent, it is possible to improve the robustness of the radar data processing apparatus 200 against unnecessary signals.

  In addition, when the RCS (Radar Cross Section) can be estimated in advance, the unnecessary signal peak trajectory generation unit 212 generates an unnecessary signal peak trajectory for a range in which the reflected wave is assumed to be a peak from the radar equation. May be. In addition, the range of influence may be taken into account depending on the distance to the target and the size of the target. These also can improve the robustness of the radar data processing apparatus 200 against unnecessary signals.

  The primary scattering estimation unit 210 calculates a value indicating the degree of correlation between the peak locus and the unnecessary signal peak locus (peak locus−unnecessary signal peak locus correlation) for each observation. When the calculated peak locus-unnecessary signal peak locus correlation degree indicates that the correlation is larger than a predetermined threshold, the primary scattering estimation unit 210 determines the peak locus as a peak locus corresponding to the primary scattering. do not do. Even if there is a correlation with the unnecessary signal peak locus, the peak locus existing before the unnecessary signal peak locus may be determined as a peak locus corresponding to the primary scattering as an exception.

  In this way, the unnecessary signal peak locus obtained based on the unnecessary signal information can be used to exclude the primary scattering peak locus from the candidates. This makes it possible to improve the robustness of the radar data processing device 200 against unwanted waves.

Embodiment 3 FIG.
In the present embodiment, an abrupt change in the motion of the observation target in an arbitrary region of the observation target is detected and notified to the user.

  As shown in FIG. 9, the laser data processing apparatus 300 according to the third embodiment of the present invention functionally has a change model reliability memory in addition to the configuration included in the radar data processing apparatus 100 according to the first embodiment. 313 and a change model transition detection unit 314.

  The change model reliability memory 313 stores change model reliability information indicating the reliability of the change model of the peak locus adopted for the primary scattering at each observation point. The change model reliability information is stored in the change model reliability memory 313 by the primary scattering estimation unit 110.

  The change model transition detection unit 314 refers to the change model reliability information stored in the change model reliability memory 313 as needed. The change model transition detection unit 314 monitors the reliability of each observation point for each change model.

  The change model transition detection unit 314 calculates a rate at which the change model has transitioned within a predetermined time among observation points belonging to a predetermined region to be observed. The change model transition detection unit 314 determines that a sudden movement change has occurred in a predetermined region to be observed when the calculated ratio is equal to or greater than a predetermined threshold. Thereby, the change model transition detection unit 314 can detect a sudden change in motion of a predetermined region to be observed.

  When the change model transition detection unit 314 detects a sudden movement change, the change model transition detection unit 314 outputs warning information indicating the change to the display 103.

  When the warning information indicating that a sudden movement change has been detected is acquired from the change model transition detection unit 314, the display device 103 displays this.

  In this way, by monitoring the temporal change in the reliability of the change model, it is possible to detect a change in motion in an arbitrary region to be observed. In addition, this can be notified to the user. As a result, the user can grasp the observation reliability.

Modification 3
When disturbance information regarding earthquakes, typhoons, and the like affecting the observation target can be obtained from an external system or the like, the radar data processing apparatus 100, 200, or 300 may further include a disturbance information acquisition unit that acquires disturbance information from the external system. . Each processing unit of the radar data processing device 100, 200, or 300 may change the change model applied to the calculation of the predicted value based on the disturbance information, and the transition between the change models in the above equation (2). The probability setting (setting value related to the applied change model) may be changed.

  By changing the setting of the change model or the transition probability, it is possible to improve the degree of matching between the observation result and the change model.

  As mentioned above, although embodiment of this invention was described, this invention is not limited to each embodiment, The form which combined each embodiment and modification suitably, and the form which added the various change to it In addition, the technical scope equivalent to these forms is also included.

  The radar data processing apparatus 100, 200 or 300 according to the present invention includes, for example, one or a plurality of processors, a RAM (Random Access Memory) serving as a work area of the processor, a ROM (Read It may be composed of only memory) and HDD (hard disk drive). In this case, all or part of the functions of the radar data processing apparatus 100, 200, or 300 may be realized by, for example, executing a computer program in which the processor is installed in advance or as a function of the processor itself.

  Further, the present invention may be a computer program that causes a computer to execute the processing described in each embodiment and the like by being installed in a general-purpose computer. This computer program may be stored and distributed in various storage media, or may be distributed via a communication line such as the Internet.

  100, 200, 300 Radar data processing device, 101 receiver, 102 signal processor, 103 display, 104 change model memory, 105 hypothesis memory, 106 peak information extraction unit, 107 correlation determination unit, 108 hypothesis creation unit, 109 hypothesis Evaluation unit, 110, 210 Primary scattering estimation unit, 211 External monitoring system, 212 Unnecessary signal peak locus generation unit, 313 Change model reliability memory, 314 Change model transition detection unit.

Claims (10)

A peak information extraction unit that acquires data indicating a Doppler spectrum for each observation point and extracts peak information including an observation value that is a frequency corresponding to a peak in the Doppler spectrum;
Correlation determination that determines whether or not there is a correlation between a predicted value that is a predicted peak position by applying a predetermined change model to a peak trajectory obtained by connecting the peaks of past observations in time series, and the observed value And
A hypothesis memory storing hypothesis information indicating a hypothesis created by a combination of the peak trajectories;
The peak trajectory included in the hypothesis indicated by the hypothesis information includes a consistent combination of peak trajectories obtained by connecting the current observed value or predicted value in time series, and the correlation determination unit determines that there is a correlation. A hypothesis creation unit that creates a hypothesis including a peak locus in which the observed value determined to have a correlation with the predicted value is connected to the peak locus corresponding to the prediction value in time series;
The likelihood of the change model is calculated based on the predicted value and the observed value, and based on the likelihood of the change model applied to a peak locus included in a hypothesis created by the hypothesis creating unit, A hypothesis evaluation unit that calculates a hypothesis reliability that is a reliability of the hypothesis, and selects a best hypothesis showing the best value of the temporary reliability;
A primary scattering estimator for determining a peak trajectory corresponding to the primary scattering of Bragg scattering from the peak trajectories included in the best hypothesis based on the specifications of the peak trajectory, and calculating a Doppler velocity of an observation target; A radar data processing apparatus comprising: The primary scattering estimation unit further outputs the reliability of the best hypothesis determined as a peak locus corresponding to the primary scattering as reliability at each observation point. The radar data processing device described in 1. The primary scattering estimation unit calculates a peak locus correlation indicating a degree of correlation between a peak locus corresponding to the primary scattering and a peak locus appearing in the opposite frequency region, and based on the peak locus correlation. The radar data processing apparatus according to claim 1, wherein the reliability of the Doppler velocity at the observation point is classified, and the classified result is output as the reliability at each observation point. The first-order scattering estimation unit classifies observation points that have a peak locus determined to correspond to the first-order scattering and whose peak locus correlation is equal to or higher than a predetermined threshold as observation points with high reliability. An observation point having a peak locus determined to correspond to the first-order scattering and having a peak locus correlation less than the threshold is classified as an observation point having a medium reliability, and the first-order scattering is performed. The radar data processing apparatus according to claim 3, wherein observation points that do not have a peak locus determined to correspond to are classified into observation points with low reliability. The primary scattering estimator includes, for each observation adopted in the peak trajectory previously assumed to correspond to the primary scattering, the best including the peak trajectory corresponding to the primary scattering in the current observation. The radar data processing apparatus according to any one of claims 1 to 4, wherein a hypothesis is re-selected, and the observation result is recalculated using the peak locus as primary scattering. When unnecessary signal information related to unnecessary signals is acquired from the outside, an unnecessary signal peak locus generating unit that generates unnecessary signal peak locus information indicating unnecessary signal peak locus based on the unnecessary signal information is further provided,
The primary scattering estimation unit excludes, from the peak trajectory candidates corresponding to the primary scattering of Bragg scattering, the unnecessary signal peak trajectory indicated by the unnecessary signal peak trajectory information and the peak trajectory indicating a degree of correlation equal to or higher than a predetermined threshold. The radar data processing device according to claim 1, wherein the radar data processing device is a radar data processing device. A change model reliability memory storing change model reliability information indicating the reliability of the change model of the peak trajectory adopted for the first-order scattering at each observation point for each observation;
Based on the change model reliability information, by calculating the proportion of observation points belonging to the predetermined area to be observed and the change model changing within a predetermined time, The radar data processing apparatus according to claim 1, further comprising: a change model transition detection unit that detects whether or not a change in motion has occurred. A disturbance information acquisition unit for acquiring disturbance information about an element affecting the observation target from an external system;
The radar data processing apparatus according to claim 1, wherein a change value to be applied or a setting value related to the change model to be applied is changed based on the disturbance information. Obtain data showing the Doppler spectrum for each observation point,
Extracting peak information including an observed value that is a frequency corresponding to a peak in the Doppler spectrum;
Determine the presence or absence of a correlation between the predicted value, which is a predicted peak position by applying a predetermined change model to the peak trajectory that connects the peaks of past observations in time series, and the observed value,
Including a consistent combination of peak trajectories obtained by concatenating the current observed value or predicted value in time series to the peak trajectory included in the hypothesis indicated by the hypothesis information indicating the hypothesis created by the combination of the peak trajectories, and A peak locus obtained by connecting the observed value determined to be correlated with the predicted value determined to be correlated by the determination of the presence or absence of the correlation in time series to the peak locus corresponding to the predicted value. Create a hypothesis,
Calculating the likelihood of the change model based on the predicted value and the observed value;
Based on the likelihood of the change model applied to the peak trajectory included in the created hypothesis, calculate a hypothesis reliability that is the reliability of the hypothesis,
Selecting the best hypothesis showing the best value for the temporary reliability,
A radar which determines a Doppler velocity of an observation object by determining a peak locus corresponding to the primary scattering of Bragg scattering from the peak locus included in the best hypothesis based on the specifications of the peak locus. Data processing method. On the computer,
Obtain data showing the Doppler spectrum for each observation point,
Extracting peak information including an observed value that is a frequency corresponding to a peak in the Doppler spectrum;
Determine the presence or absence of a correlation between the predicted value, which is a predicted peak position by applying a predetermined change model to the peak trajectory that connects the peaks of past observations in time series, and the observed value,
Including a consistent combination of peak trajectories obtained by concatenating the current observed value or predicted value in time series to the peak trajectory included in the hypothesis indicated by the hypothesis information indicating the hypothesis created by the combination of the peak trajectories, and A peak locus obtained by connecting the observed value determined to be correlated with the predicted value determined to be correlated by the determination of the presence or absence of the correlation in time series to the peak locus corresponding to the predicted value. Create a hypothesis,
Calculating the likelihood of the change model based on the predicted value and the observed value;
Based on the likelihood of the change model applied to the peak trajectory included in the created hypothesis, calculate a hypothesis reliability that is the reliability of the hypothesis,
Selecting the best hypothesis showing the best value for the temporary reliability,
For determining the peak trajectory corresponding to the first-order scattering of Bragg scattering from the peak trajectories included in the best hypothesis and calculating the Doppler velocity of the observation target program.

Images