JP2001004398A - Movement information detection method based on satellite sar image - Google Patents

Abstract

PROBLEM TO BE SOLVED: To automatically and accurately detect the speed and navigation direction of a ship according to a ship image being detected from the image of an earth surface being photographed by the synthetic aperture radar of an artificial satellite. SOLUTION: A ship is detected by a ship detection part 4 based on the image amplitude ratio of satellite SAR sea-surface image data 1 where a land part is eliminated in advance for treatment and a threshold, furthermore, the Hough transform of the upper and lower sides of the ship being detected by a track detection part 5 is used for detecting a track, and Doppler shift is calculated by a ship-side detection part 6 according to the amount of deviation between a navigation position and a ship position for detecting the speed of the ship, thus accurately detecting the speed and the navigation direction of the ship according to a satellite SAR image.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a moving object based on a satellite SAR image detecting movement information of the moving object photographed by a synthetic aperture radar (hereinafter, referred to as a satellite SAR) mounted on an artificial satellite. The moving information detecting method of the present invention detects a moving speed and a moving direction of a moving body, for example, a ship on the sea or a vehicle on land.

[0002]

2. Description of the Related Art A satellite SAR radiates a transmission wave from an antenna mounted on a satellite to the earth's surface, processes a reflected wave from the earth's surface by a receiver mounted on the satellite, and transmits data to a ground station. . The ground station receives the data from the satellite, performs necessary signal processing, and creates image data called a satellite SAR image. This satellite SAR image is
It is stored in a medium such as a CD-ROM or a magnetic tape by a satellite operating organization, or provided to users by data communication means.

[0003] Since the satellite SAR image is obtained as still image data, it is possible to detect the position at the time of imaging of a moving body such as a ship navigating at sea, but to detect the speed in the conventional technology. I couldn't do that.

In the prior art, studies have been made on estimating the speed of a ship from the length of the wake of the ship shown in the satellite SAR image. However, the length of the wake is not limited to the speed of the ship. Ship size, ship shape, ship propulsion, ocean current,
Depending on conditions such as water depth and ocean wind, the satellite SA
Since these conditions cannot be specified from the R image, the speed of the ship cannot be correctly estimated.

[0005] As a method other than the above, a satellite S of one scene is used.
Utilizing that an AR image is created as a superposition of multiple images with different observation timings called looks,
A method of comparing and evaluating the look images and detecting the speed and the traveling direction of a moving object such as a ship has been studied.However, this method has a small time difference between the looks, and a low speed difference of a fishing boat or the like during the time difference. There is a problem that the moving amount of the ship is smaller than the distance resolution of the satellite SAR image, and as a result, the speed is not detected.

[0006]

When performing a search on the sea for the purpose of monitoring and controlling the behavior of a ship in a wide sea area such as the high seas or 200 nautical mile water area, a search aircraft or a ship is not sufficient. Due to the inability to obtain a large search range, restrictions on weather conditions and flight / navigation range, an image of the sea surface is obtained by an imaging device mounted on a satellite, and the image is analyzed to The need to grasp movement has come. At that time, conventional technology can detect the position of the ship, but cannot detect the speed and navigation direction of the ship. Was.

SUMMARY OF THE INVENTION The present invention has been made to solve such a problem, and a moving object based on a satellite SAR image capable of correctly detecting moving information of the moving object from a satellite SAR image as a still image. It aims to provide an information detection method.

[0008]

A satellite SA according to the present invention
The method of detecting the movement information of the moving object based on the R image is performed by using the satellite SA
The moving trajectory of the moving object is detected by detecting the image of the moving object from the R image and analyzing the amplitude information of the peripheral pixels in the search range in consideration of the Doppler shift of the image of the moving object. The moving direction is detected based on the following.

Further, a Doppler shift and a range rate of the ship are calculated from a deviation between the detected starting point of the moving trajectory and the position of the moving body, and the moving speed of the moving body is detected based on the calculated Doppler shift and the direction of the detected moving track. It is characterized by the following.

Further, the moving body is a ship, and the trajectory of the ship is detected as the movement trajectory.

Further, the moving body is a vehicle, and detects a rut, a road, or a track as the moving trajectory.

[0012]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiment 1 FIG. 1 relates to Embodiment 1 of the present invention, and is an overall configuration diagram of an apparatus for executing a method for detecting a speed and a traveling direction of a ship on a satellite SAR image. In FIG. 1, reference numeral 1 denotes a satellite SAR image provided by a satellite operating organization on a medium such as a CD-ROM or a magnetic tape. In the present embodiment, it is assumed that the satellite SAR image to be processed is a sea surface image, and the satellite SAR image including land has been subjected to land land removal processing in advance. 2 reads the satellite SAR image data 1 from the medium, causes the display device 3 to display the image, and performs search / registration processing of the ship and wake database 7 necessary for processing of each part of the device. And an input / output control unit that outputs ship specification data such as the speed and navigation direction of the ship detected by the present apparatus.

Reference numeral 3 denotes a display device for displaying input satellite SAR image data, processing results, and the like. 4 is satellite SA
A ship detection unit that detects a ship reflected on the R image and specifies the position of the ship, and 5 is a satellite SAR image of a wake generated by each of the ships detected by the ship detection unit 4 when the ship sails. A wake detecting unit for detecting the position and direction of the starting point of the wake, and 6 is a ship detecting unit.
And a ship speed detection unit that detects a speed and a navigation report of each ship from the position of each ship specified in the above and the position and direction of the starting point of the corresponding wake detected by the wake detection unit 5. Reference numeral 7 registers and manages the processing result of the ship detecting unit 4 and the processing result of the wake detecting unit 5,
Further, a ship and wake database 8 for performing a search for each of the processes as necessary is an output of the processing results of the present apparatus, and various data such as the speed and traveling direction of each ship detected from the satellite SAR image. Original data.

Next, the operation of the above embodiment will be described with reference to the flowcharts shown in FIGS. FIG.
3 is a ship detection unit 4, FIG. 3 is a wake detection unit 5, FIG. 4 is a wake search process on the lower side of the ship which forms a part of the wake detection unit 5, and FIG. FIG. 6 is a flow chart showing the operation of the marine vessel speed detecting unit 6. FIG. 6 is a flow chart showing the operation of the marine vessel speed detecting unit 6.

The operation of the ship detector 4 will be described with reference to the flowchart shown in FIG. First, in the scan start position setting process 9, the upper left position of the satellite SAR sea surface image is set as the scan start position. Here, the coordinate system of the satellite SAR sea surface image includes a range direction 63 and an azimuth direction 64, as shown in FIG. Azimuth direction 64 is satellite 6
The range direction 63 is a direction orthogonal to the direction parallel to the traveling direction 66 of FIG. The antenna of the satellite SAR is mounted parallel to the azimuth direction, and performs transmission centering on the range direction.

The range measured with respect to the satellite is 0 at the satellite position. However, in the description of the processing of the present apparatus, the range at the upper left position and the azimuth coordinates of X_RNG = 1, Y_RNG =
In the following description, coordinate values are counted corresponding to the pixels of the satellite SAR image from top to bottom in the azimuth direction and from left to right in the range direction. Therefore, the increasing direction of the distance from the satellite is opposite to the increasing direction of the range in this coordinate system.

Next, the ship detection gate processing 10 shown in FIG.
In, the ship detection gate shown in FIG. 9 is set around the current scanning position on the satellite SAR image. In FIG. 9, the center of the target detection gate 67 is the current scanning position. The target detection gate is composed of an outer frame, an inner side, and two parts. In the ship detection gate processing 10, the average of the amplitudes of the pixels located in the outer frame (mean_Gate_
out) and the average of the amplitudes of the inner pixels (mean_Gat)
e_in) is calculated, and the ratio (Gate_Ra) is calculated.
tio) is calculated.

The ship detection gate applied to the satellite SAR image is a two-dimensional target detection circuit used in radar as shown in FIG. In the target detection circuit shown in FIG. 10, the range bin delay circuit 69 corresponds to the outer frame of the target detection gate, and the range bin 70 in the detection range corresponds to the inside of the target detection gate. In the radar target detection circuit, these ratios are calculated and compared with a threshold value.

Next, in the ship detection determination process 11 shown in FIG. 2, the threshold (Gate_Ratio) calculated in the ship detection gate process 10 is set in the same manner as the radar target detection circuit shown in FIG. (Threshold) is evaluated. Based on the result, it is determined whether or not a ship exists at the current scanning position.

If it is determined in this vessel determination processing 11 that a vessel exists at the current scanning position, the newly detected vessel registration processing 12 determines the range of the current scanning position,
The azimuth position is registered in the ship and wake database 7 as a newly detected ship position.

After the end of the newly detected vessel registration processing 12 shown in FIG. 2 or when no vessel is detected in the vessel detection determination processing 11, the current scanning position is changed to the range scanning in the range direction scanning completion confirmation processing 13. It is determined whether the end of the direction (right end) has been reached.

If it has not reached the right end, in the range direction scanning continuation process 14 shown in FIG. 2, the scanning position is advanced in the range direction, and then the transition to the ship detection gate process 10 is repeated.

On the other hand, if it has reached the right end, it is checked in the azimuth direction scanning end confirmation processing 15 shown in FIG. 2 whether the current scanning position has reached the end (lower end) in the azimuth direction.

If the lower end has not been reached, in the azimuth direction continuation process 16, the scanning position is advanced in the azimuth direction, and then the range direction is set to the initial position (left end).

Next, the operation of the track detection unit 5 will be described with reference to the flowchart shown in FIG. The wake detection unit 5 searches the sea surface around the ship detected by the ship detection unit 4 and detects the wake.

The satellite SAR image, as shown in FIG.
It consists of a range direction and an azimuth direction based on the traveling direction of the satellite. Since the satellite SAR is an existing satellite, detailed description of the principle of imaging is omitted, but the outline is as follows.

When processing and imaging a reflected wave from an object on the earth's surface, the satellite SAR performs positioning in the range direction based on the arrival time of the reflected wave from the object from the transmission of the radar wave, and performs Doppler reflection of the reflected wave. By determining the azimuth position of the object by the frequency, the image of the object is positioned on the satellite SAR image. Specifically, the position of the object in the azimuth direction is determined as a direction directly facing the position of the satellite at the time when the Doppler frequency of the reflected wave from the object becomes zero.

As shown in FIG. 11, a fixed target 71 on the earth's surface is a satellite SAR while the satellite 73 is in orbit.
Irradiation of the radar wave 72 in the range of the beam width of
The satellite SAR receives the reflected wave. The reflected wave receives the Doppler effect corresponding to the relative speed between the satellite and the target. Therefore, the Doppler frequency 74 of the received wave is a positive Doppler frequency while the satellite approaches the target, and a minus Doppler frequency when the satellite moves away. When the satellite approaches the fixed target again, that is, when the fixed target is located in the direction of the central axis of the satellite SAR antenna, the Doppler frequency becomes 0, and the target Azimuth position can be detected correctly.

On the other hand, as shown in FIG. 12, for a target 75 moving on the earth's surface, the target moves in addition to the Doppler frequency caused by the satellite's travel in the received wave Doppler frequency. The generated Doppler frequency is superimposed. The component that changes the Doppler frequency of the received wave due to the movement of the target is a Doppler shift 7
6.

As a result of the Doppler shift 76, the received wave Doppler frequency 77 before the influence of the Doppler shift is obtained.
With respect to the received wave Doppler frequency 7 after the Doppler shift.
8 results in the position of the 0 point being shifted by Δ _AZ 79. As a result, the moving target appears at a position shifted by _ΔAZ on the satellite SAR image.

The sign of Δ _AZ is determined depending on whether the moving direction of the target is a direction approaching the satellite or a direction moving away from the satellite. In the case of approaching, plus (downward of the satellite image),
In the case of distant, it becomes minus (upward in the satellite image).

_ΔAZ is calculated from the physical law of the Doppler effect as follows. First, the Doppler frequency of a reflected wave of a fixed target is given by the following equation. f ^SAT _d = 2 · V _SAT · sin θ _AZ (1) Here, f ^SAT _d is a Doppler frequency generated due to the motion of the satellite with respect to the received wave from the fixed target, and
_SAT is the speed at which the satellite is traveling, and θ _AZ is the azimuth angle of the fixed target as seen from the satellite, and changes with time as the satellite travels.

On the other hand, the Doppler frequency generated by the movement of the target is given by the following equation. f ^TGT _d = proviso _{(2 · V R · cosθ AZ} ) + (2 · V AZ · sinθ AZ) (2), V R is the range direction of the velocity of the moving target, V _AZ is a shows the speed of the azimuth direction I do.

[0034] In the position of the delta _AZ, since the Doppler frequency generated with the progress of the satellite with respect to the Doppler frequency and the fixed target shown in equation (2) due to the movement of the target is canceled,
By making the right side of equation (1) equal to the right side of equation (2), the following equation is obtained. 2 · V _SAT · sin θ _AZ = (2 · V _R · cos θ _AZ ) + (2 · V _AZ · sin θ _AZ ) (3)

When this is arranged, the following equation is obtained. In _{(V SAT -V AZ) · sinθ} AZ = V R · cosθ AZ (4) equation (4), by V _AZ is less than the _{V SAT, tanθ AZ = V R} / V SAT (5) where And the left side is the distance R between the satellite and the moving target,
There is the following relationship. _{tanθ AZ = ΔAz / R (6} ) Equation (5) and (6), .DELTA.AZ = a _{(V R · R) / V} SAT (7).

Here, the distance R between the satellite and the moving target can be calculated as follows. First, the respective distances from the satellite to the right and left ends in the range direction of the satellite SAR image are calculated from the orbit altitude of the satellite and the angle between the horizontal line and the radar beam axis of the SAR. Since this calculation has a known content, a detailed description thereof will be omitted. Next, R is obtained by interpolating the distance from the satellite to the rightmost position and the leftmost position corresponding to the target range on the satellite SAR image.

Since the influence of the Doppler effect does not change the arrival time of the reflected wave from the target, the range of the moving target appears at the original position.

The vessels navigating is, when the moving target in the above description, in proportion to the components results vessels Doppler shift perpendicular to the satellite orbit velocity sailing, the position shifted delta _AZ from the original position in the azimuth direction Appears as an image.
On the other hand, the wake created by this ship is almost stationary on the sea surface, and is hardly affected by the Doppler shift like a ship, so that an image appears almost at its original position.

The wake search azimuth range setting processing 17 shown in FIG. 3 of the wake detection unit 5 sets an azimuth range in which the wake generated by the marine vessel detected by the ship detection unit 4 should be searched according to the flowchart shown in FIG. Is what you do. Track is located at a position shifted delta _AZ with respect to the position of the vessel, the value of the delta _AZ is determined from each component of the formula (7), V _SAT and R is a parameter determined from the satellite orbit specifications since the range that can be taken of the delta _AZ is determined by the range of V _R. V _R is the range component of the speed at which the ship sails, and does not exceed the maximum speed of the ship.

Accordingly, the range in which the wake is searched depends on the maximum ship speed (assumed) of the ship type to be analyzed by the present apparatus.
What is necessary is just to set it as the range of the following formula. 0 ≦ Δ _AZ ≦ (V _MAX · R) / V _SAT = WIDTH (8) In addition, whether the ship sails in the direction approaching the satellite,
Depending on whether the ship is sailing away from the ship, the ship appears above or below the wake. Therefore, when searching for the wake around the ship, it is necessary to search the range of equation (8) for each of the upper and lower sides (azimuth direction). No. In the wake search azimuth range setting processing 17 shown in FIG. 3, the value of the wake search range WIDTH is set as the wake search azimuth range based on equation (8). Since the azimuth direction of the shift delta _AZ may occur in the vertical directions, the equation (8)
The range WIDTH for searching the wake determined by is actually one width of the search range of the wake.

The search for a wake is carried out for all ships registered in the wake database 7 by the ship detection unit 4 in accordance with the flowchart shown in FIG. In the wake search start process 18 shown in FIG. 3, one of the ships is arbitrarily selected, and the search for the wake is started.

The search for the wake of a ship is performed in the order of the wake search process 19 on the lower side of the ship and the wake search process 20 on the upper side of the ship. As shown in FIGS. 13 and 14, when the ship is moving away from the satellite orbit, the wake appears on the lower side of the ship, and when the ship approaches the satellite orbit, as shown in FIGS. If so, the wake appears above the ship. At the time of these processes, it is undetermined whether the ship is approaching or away from the satellite orbit, and the wake is searched for above and below the ship. Conversely, from the results of searching for the wake in these processes, it is determined whether the ship is approaching or away from the satellite wake. These contents will be described later with reference to FIGS.

In the wake search process 19 on the lower side of the ship and the wake search process 20 on the upper side of the ship, candidates regarded as the wakes of the corresponding ship are detected. The process of evaluating the candidates (plurality) and specifying the wake is the wake candidate selection process 21. This wake candidate selection process is described in FIG.
It will be described later with reference to FIG.

After the track of the currently selected vessel is determined as a result of the track candidate selection processing 21, in the track search end confirmation processing 22, the currently selected vessel is registered in the vessel and track database 7. Check if the ship is. If this is the last ship, the process of the wake detection unit ends. If this is not the last ship, the next ship is selected in the wake search continuation process 23 and the process 1
A series of processes of wake search after 9 are repeatedly executed.

Next, the wake search processing on the lower side of the ship by the wake detection unit 5 will be described with reference to the flowchart shown in FIG. A wake occurs below the ship when the ship is traveling in a direction away from the satellite orbit, as shown in FIG. At this time, the position on the satellite image of the ship and track, positional deviation of the azimuth direction corresponding to delta _AZ occurs. The range of this shift is, as determined by equation (8),
The range is from 0 to WIDTH.

First, in the track search start processing 24 on the lower side of the ship, the search for the track is started from the lowest point in the azimuth direction, and it is confirmed whether there is a track starting from the lowest point. Therefore, the range coordinates (RN) of the search start position
G_ORIGIN) and azimuth coordinates (AZ_ORIGI)
N) are set as follows. RNG_ORIGIN = SHIP_RNG AZ_ORIGIN = SHIP_AZ_WIDTH (9) Here, SHIP_AZ and SHIP_RNG are the azimuth position and the range position of the ship currently being searched for the wake.

Next, in the Hough transform processing 25, the satellite S
The Hough transform is performed for each search angle step from the wake search start point set on the AR image to the vicinity thereof over the search angle range, and linear component detection evaluation is performed. FIG. 15 shows the satellite S
It shows the positional relationship of the Hough transform performed on the AR image. The Hough transform is an operation of integrating the amplitude of the satellite SAR image from the wake search start point 82 to the direction of the wake search direction angle α over a range of the length of the Hough transform integration length L. α
The Hough transform result H (α) in the direction is given by the following equation (10).

[0048]

(Equation 1)

Here, A (X_RNG, Y_AZ) is
The amplitude of the (X_RNG, Y_AZ) position of the satellite SAR image is represented. L is an integral element from 0 to L. The conversion integral length L needs to be set based on the result of investigation of the track length appearing on the satellite SAR image, but is generally set in the range of several hundred meters to several kilometers. In the Hough transform process, a transform based on equation (10) is performed for the wake search direction angle α in the range of 0 ≦ α <2π. Note that the step of the wake search direction angle α is determined from the required resolution for specifying the voyage direction of the ship and the constraint on the calculation processing time. Further, since the vicinity direction is affected by the image of the ship, the processing is performed by excluding the coordinates (SHIP_RNG, SHIP_AZ) of the ship from the range of the Hough transform.

In the Hough transform statistical analysis result processing 26, a result H (α) (0 ≦ 0) obtained by performing a Hough transform on the current track search start point and its surroundings for each track search direction angle.
The average μ and the standard deviation σ of α <2π) are calculated.

FIG. 16 shows a bright and dark image of the ship and the wake on the satellite SAR image. Ship 85 is SAR compared to sea level
Since the radar wave has a high reflectance, the satellite SAR image appears brighter (whiter) than the surrounding sea surface. On the other hand, the wake is generally divided into two parts, a bright (white) wake 86 and a dark (black) wake 87. this is,
The wake is not completely stationary on the surface of the sea, but has a slight speed in the same direction as the ship's navigation direction. Therefore, the position of the wake shifts in the azimuth direction like the image of the ship.

As shown in FIG. 16, when the ship travels away from the satellite orbit and the ship image is shifted upward, the wake image is also shifted upward. Due to the shift of the wake image, the original wake position has a reduced amplitude and appears as a dark wake. Conversely, where the wake image is shifted, it appears as a bright wake because the amplitude of the original sea surface reflection and the amplitude of the wake image are superimposed. The above is the mechanism of the track image where the bright track and the dark track appear, and the satellite SAR
On the image, it is necessary to detect both. In addition,
From the above mechanism of generating the wake image, the original position of the wake is a dark wake, and the starting point of the dark wake is the original position of the ship.

The dark wake detection determination process 27 determines that the result of Hough transform performed on a dark wake shows a lower value (H (α)) than the result of Hough transform performed on a general sea surface. Use to detect dark wakes. The Hough transform is a conversion process that calculates the sum of the amplitudes of satellite SAR images on a line segment. Therefore, since the amplitude of each point is low on a dark wake, the sum is small compared to the general sea level. are doing.

In the Huff transform result statistical analysis processing 26 described above, the average μ and the standard deviation σ are calculated from the result of the Hough transform over the entire circumference of the current search start point. If there is a Hough transform result below -kσ, it can be concluded that the result has a specifically low value in the population. This is a candidate for a dark wake. Here, the value of k corresponds to a threshold for detecting a singular value, and is set in consideration of detection results, but is usually set to about 3.

The dark wake direction determination process 28 evaluates the detection status when a plurality of dark wake candidates are detected with respect to the same wake search start point in the dark wake detection determination process 27, and determines the most reliable. This is a process for determining a likely dark wake direction.

When only one dark wake candidate is detected in the dark wake detection determination process 27 described above, the dark wake direction determination process 28 immediately determines this as the result of the dark wake direction determination. On the other hand, if a plurality of dark wake candidates are detected in the dark wake detection determination process 27 described above, the dark wake direction determination process 28 distributes dark wake candidates most in the angle range of 0 ≦ α <2π. The direction is specified (evaluation of the frequency of the wake candidate at each angle step), firstly, candidates within the angular range are narrowed down, and the median value is determined as a dark wake direction determination result.

In the dark track registration processing 29, the track direction obtained as a result of the determination in the dark track direction determination processing 28 is registered in the ship and track candidate table of the track database 7. This table is configured for each ship. In this table, in addition to the dark track on the lower side of the vessel to be registered as a candidate in the dark track registration processing 29, the bright track registration processing 32 and the dark track registration processing 4 in the track search processing on the upper side of the vessel shown in FIG.
0, in the bright wake registration process 43, the wake candidates for the respective ships are registered, and the wake detection unit 5 shown in FIG.
Is a table for finally selecting a track from those candidates in the track candidate selection processing shown in FIG.

The information to be registered is: track candidate type: “dark track”, track candidate start point: current track search start point, track candidate direction: track search direction angle α obtained in dark track direction determination processing 28, track candidate score: | Huff transform result H in wake direction
(Α) −μ | / σ. Here, the track candidate score is used as a criterion for selecting one of a plurality of track candidates detected in the track candidate selection processing shown in FIG. 6 by the track detector 5, and the amplitude of the detected track candidate is used. Is an index indicating how much deviation is with respect to the amplitude of the surrounding sea level.

If no dark wake candidate is detected in the dark wake detection determination process 27 described above, then
In the bright wake detection determination processing 30, bright wake candidates are detected. The detection is performed in the same manner as in the dark wake detection determination processing 27, but in order to detect a direction in which the Hough transform result has a deviation on the bright side, if there is a Hough transform result exceeding μ + kσ, it is determined as a bright wake candidate.

Subsequently, the bright wake direction determination processing 31 and the bright wake registration processing 32 are the same processings as the dark wake direction determination processing 28 and the dark wake registration processing 29, respectively, and the description thereof will be omitted.

The wake search process on the upper side of the ship by the wake detector 5 shown in FIG. 5 is a process executed following the wake search process on the lower side of the ship in the flowchart by the wake detector 5 shown in FIG. The track search processing on the upper side of the ship is almost the same as the track search processing on the lower side of the ship shown in FIG.
In order to set the wake search range above the ship position, the azimuth position of the wake search start point is set in the wake search start process 35 on the upper side of the ship as in the following equation. AZ_ORIGIN = SHIP_AZ + WIDTH The content of the wake search process on the upper side of the ship shown in FIG. 5 is almost the same as that of the wake search process on the lower side of the ship in FIG. 4, and thus detailed description is omitted.

Next, the track candidate selection processing by the track detection unit 5 shown in FIG. 6 is performed by the track search processing detected in the ship track search processing on the lower side of the ship shown in FIG. 4 and the track search processing on the ship upper side shown in FIG. Is the process of selecting the most probable wake. The wake candidates include a dark wake and a bright wake, and the original wake position where the ship is generated is a dark wake position. Therefore, when a dark track can be detected, the dark track is prioritized as the track position. If no dark wake is detected, a bright wake is selected as the wake of the vessel.

When there are a plurality of dark tracks, and when there are only a plurality of bright tracks without dark tracks, it is necessary to select between tracks of the same kind. In each of the track search processing on the side and the track search processing on the upper side of the ship shown in FIG. 5, the score given to the track candidate (the evaluation result of the degree of deviation of the amplitude of the track candidate from the average amplitude of the sea surface) is referred to. The wake candidate with the highest score (the largest deviation from the average sea level amplitude) is selected.

The flow of the wake candidate selection processing is as described above, and FIG. 6 is a flowchart developed based on the wake candidate selection processing. By executing a series of processes based on the above flow, the most probable wake starting point position, wake direction, and wake type are specified.

First, in the track search processing on the lower side of the ship shown in FIG. 4 and the track search processing on the upper side of the ship shown in FIG. 5, processing 46 for listing up the dark track and the bright track registered in the track candidate table is performed. At 47, it is determined whether there is a dark wake. If there is a dark track, the highest score among the dark tracks is selected as the track of the current processing target vessel in the selection processing 48, and the current processing target is stored in the ship and track database 7 in the subsequent setting processing 49. The wake is set for the ship, and the value of the selected wake candidate is set in the wake start position, the wake direction, and the wake type.

On the other hand, when it is determined in the determination process 47 that there is no dark wake, it is determined in the determination process 50 whether there is a bright wake. If there is a bright wake,
In the selection process 51, the one with the highest score among the bright wakes is selected as the wake of the current target ship, and in the following setting process 52, the wake is set for the ship and the wake database for the current target ship in the wake database 7. The value of the selected track candidate is set in the track start point position, the track direction, and the track type.

If it is determined in the determination process 50 that there is no bright wake, in the process 53, it is determined that the current target ship has no wake, and in the subsequent setting process 54,
No wake is set for the current ship to be processed in the ship and wake database 7, and the wake start point position, wake direction, and wake type are not set. In this way, the most probable wake starting point position, wake direction, and wake type are specified.

Next, the operation of the boat speed detector 6 will be described with reference to the flowchart shown in FIG. The ship speed detection unit 6 is a process for detecting the speed and the traveling direction of the ship from the starting point position and the wake direction of the wake specified by the wake candidate selection process by the wake detection unit 5 shown in FIG. The ship speed detection unit 6 detects the ship and all the ships registered in the wake database 7 by the wake candidate selection process by the wake detection unit 5 shown in FIG. The speed of each is detected.

For this purpose, first, in the track speed detection start processing 55, one vessel is arbitrarily selected from the boat and the track database 7, and the processing is started. Next, for the selected ship, in the track presence / absence confirmation processing 56, it is confirmed whether or not a wake has been detected for the ship in the vessel detection processing by the track detection unit 5 shown in FIG.

As a result of the confirmation, if no wake is found, the estimated ship speed is set as undetectable in the ship speed undetectable determination processing 57, and the setting result is registered in the ship and wake database 7, A ship speed detection end confirmation process 61 is performed.

On the other hand, when the wake detection processing 56 confirms that a wake has been detected in the wake detection processing by the wake detection unit 4 shown in FIG.
An azimuth shift / range direction speed calculation process 58 is performed. Azimuth shift Δ _AZ
This is the distance that the image of the ship has moved in the azimuth direction from the original position of the ship, and the original position of the ship can be determined from the starting point of the wake. Therefore, the azimuth shift is calculated by the following equation. Δ _AZ = AZ_ORIGIN_SHIP_AZ (11) Here, AZ_ORIGIN is the azimuth coordinate of the wake starting point position of the ship currently being processed and read from the wake database 7. SHIP_AZ is also the azimuth coordinate of the ship.

Next, since the azimuth shift and the speed V _R of the ship in the range direction have the relationship of the equation (7), the following equation is obtained by solving the equation in reverse, and based on this, the following equation is obtained. to calculate the velocity V _R. V _R = (Δ _AZ · V _SAT ) / R (12) In the same manner as in equation (7), R is the distance between the satellite and the ship,
V _SAT is the speed at which the satellite is traveling.

Azimuth shift / range direction speed calculation processing
The speed V in the range direction of the ship obtained at 58_ROf the ship
The navigation speed vector is projected in the range direction of the satellite SAR image
Match the length you did. Therefore, the ship speed calculation process 59
From the range speed V _RIn the direction of the ship's navigation
By performing shadow conversion, the speed of the ship can be detected.
The calculation formula is given by the following formula. V_SHIP= V_R/ Cosα (13) where α is the navigation angle of the ship and V_SHIPIs the ship
The navigation speed.

In the following vessel speed / traveling direction registration processing 60, the traveling speed of the vessel detected in the vessel speed calculation processing 59 is registered in the vessel and wake database, and the vessel speed detection processing for the currently selected vessel is completed. .

In the vessel speed detection end confirmation processing 61, it is confirmed whether or not the processing has been completed for all vessels registered in the vessel and wake database. If the processing is completed, the processing of the boat speed detection unit is completed.

When there is an unprocessed ship, the next ship is selected in the ship speed detection continuation process 62, and the processes after the wake check process 56 are repeatedly executed for all the ships.

In the first embodiment, the moving body is described as an example of a ship and the trajectory is a ship wake. However, a vehicle is detected as a moving body, and a rut, a road, or a track is detected as a trajectory. Of course, the present invention can be applied to the case.

[0078]

As described above, according to the present invention, an image of a moving object is detected from a satellite SAR image, and amplitude information of peripheral pixels in a search range in consideration of a Doppler shift of the image of the moving object is analyzed. By doing so, the moving path of the moving object is detected, and the moving direction is detected based on the direction of the moving path, so that the moving direction of the moving object can be correctly detected from the satellite SAR image as a still image. .

Further, the Doppler shift and the range rate of the ship are calculated from the displacement between the detected starting point of the moving trajectory and the position of the moving body, and the moving speed of the moving body is detected based on the calculated Doppler shift and the direction of the detected moving trajectory. Thus, the moving speed of the moving object can be correctly detected from the satellite SAR image which is a still image.

[Brief description of the drawings]

FIG. 1 is an overall configuration diagram of an apparatus for detecting a speed of a ship on a satellite SAR ocean image according to a first embodiment of the present invention.

FIG. 2 is a diagram illustrating a processing flow of a ship detection unit in FIG. 1;

FIG. 3 is a diagram illustrating a processing flow of a wake detection unit in FIG. 1;

FIG. 4 is a diagram showing a processing flow of a wake search process on the lower side of the ship by the wake detection unit in FIG. 1;

FIG. 5 is a diagram showing a processing flow of a wake search process on the upper side of the ship by the wake detection unit in FIG. 1;

FIG. 6 is a diagram showing a processing flow of a track candidate selection process of the track detector of FIG. 1;

FIG. 7 is a diagram showing a processing flow of a ship speed detection unit in FIG. 1;

FIG. 8 is an explanatory diagram of a coordinate system of a satellite SAR image.

FIG. 9 is an explanatory diagram of a ship detection gate.

FIG. 10 is a diagram showing a target detection circuit used in a radar.

FIG. 11 is an explanatory diagram of the Doppler frequency of a reception wave from a fixed target.

FIG. 12 is an explanatory diagram of the Doppler frequency of a wave received from a moving target.

FIG. 13 is an explanatory diagram of a positional relationship on a satellite image of a ship navigating away from a satellite orbit and a wake.

FIG. 14 is an explanatory diagram of a positional relationship on a satellite image of a ship traveling in a direction approaching a satellite orbit and a wake.

FIG. 15 is a Huff-transformed positional relationship diagram for wake search.

FIG. 16 is a diagram showing a bright and dark image of a ship and a wake in a satellite SAR image.

[Description of Signs] 1 Satellite SAR sea surface image data, 2 Input / output control unit, 3
Display device, 4 ship detector, 5 track detector, 6 ship speed detector, 7 ship and track database, 8
Ship specification data, 9 Scan start position setting processing, 10
Ship detection gate processing, 11 Ship detection judgment processing, 12
New detection vessel registration processing, 13 range direction scanning end confirmation processing, 14 range direction scanning continuation processing, 15 azimuth direction scanning end confirmation processing, 16 azimuth direction scanning continuation processing, 17 wake search azimuth range setting processing, 18
Track search start processing, 19 Track search processing under the vessel, 2
0 Upper track search processing, 21 track candidate selection processing, 22 track search end confirmation processing, 23 track search continuation processing, 24 lower track search start processing, 25 Huff conversion processing, 26 Huff conversion result statistical analysis processing, 27 Dark wake detection judgment processing, 28 Dark wake direction judgment processing, 2
9 Dark track registration processing, 30 Bright track detection determination processing, 31 Bright track direction determination processing, 32 Bright track registration processing, 33 Track search end confirmation processing under the ship, 3
4 Wake search continuation processing on the lower side of the ship, 35 Wake search start processing on the upper side of the ship, 36 Huff conversion processing, 37 Huff conversion result statistical analysis processing, 38 Dark wake detection determination processing, 39
Dark track direction determination processing, 40 Dark track registration processing, 4
1 Bright wake detection determination processing, 42 Bright wake direction determination processing, 43 Bright wake registration processing, 44 Wake search end confirmation processing on the upper side of the ship, 45 Wake search continuation processing on the upper side of the ship, 46 Wake candidate list up processing, 47 Dark wake confirmation Processing, 48 dark track selection processing, 49 dark track registration processing, 50 bright track confirmation processing, 51 bright track selection processing, 52 bright track registration processing, 53 no track determination processing, 54 trackless registration processing, 55 track speed detection start processing , 56 Wake Presence Confirmation Process, 57 Vessel Speed Undetectable Determination Process, 58 Azimuth Shift / Range Speed Calculation Process, 59 Vessel Speed Calculation Process, 60 Vessel Speed / Navigation Direction Registration Process, 61 Vessel Speed Detection Completion Confirmation Process, 62 Vessel Speed detection continuation processing, 63 satellite SAR
Image range direction, 64 satellite SAR image azimuth direction,
65 satellite, 66 satellite traveling direction, 67 target detection gate, 68 gate scanning direction, 69 range bin delay circuit, 70 range bin in detection range, 71 fixed target on earth surface, 72 radar wave of satellite SAR (transmitted wave and reflected wave) , 73 satellite, 74 Doppler frequency of fixed target reflected wave, 75 moving target on the earth's surface, 76 Doppler shift due to target movement, 77 Doppler frequency of fixed target reflected wave, 78 Doppler frequency of moving target reflected wave, 79
Δ _Az (azimuth shift) due to Doppler shift,
80 Δ _Az (azimuth shift) on satellite image, 8
1 Range direction component of ship speed, 82 track search starting point,
83 Track search direction angle α, 84 Hough transform integral length L, 8
5 Light / dark image of ship, 86 Bright wake, 87
Dark wake.

 ────────────────────────────────────────────────── ─── Continued on the front page (72) Inventor Tsuyoshi Yamada 792 Kamimachiya, Kamakura City, Kanagawa Prefecture Mitsubishi Space Software Co., Ltd. Kamakura Division F term (reference) 2F029 AA04 AC02 AC12 5J070 AC01 AC06 AC11 AK22 BA01 5L096 BA02 CA18 DA02 FA19 FA24 FA32 FA67 FA69 GA17 GA51 HA04 LA10 9A001 BB02 BB03 CC05 FF03 GG04 HH23 HH28 JJ06 JJ78 KK56

Claims (4)

[Claims] A moving trajectory of a moving object is detected by detecting an image of the moving object from a satellite SAR image and analyzing amplitude information of peripheral pixels in a search range in consideration of a Doppler shift of the image of the moving object. SAR detecting a moving direction based on the direction of the moving trajectory
A moving information detection method for a moving object based on an image. 2. The method for detecting movement information of a moving object based on a satellite SAR image according to claim 1, wherein a Doppler shift and a range rate of the ship are calculated from a deviation between the detected starting point position of the moving trajectory and the position of the moving object. And detecting a moving speed of the moving object based on the satellite SAR image, based on the detected direction of the moving track. 3. The method for detecting movement information of a moving object based on a satellite SAR image according to claim 1 or 2, wherein the moving object is a ship, and a wake of the ship is detected as the movement trajectory. A method for detecting movement information of a moving object based on a SAR image. 4. The moving information detecting method for a moving object based on a satellite SAR image according to claim 1, wherein the moving object is a vehicle, and a rut, a road, or a track is detected as the moving trajectory. A method for detecting movement information of a moving object based on a satellite SAR image.