JP2018146354A - Sonar image processor, sonar image processing method, and sonar image processing program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a sonar image processing technique with which it is possible to discriminate a target with high accuracy while avoiding increase in the size of a device or the complexity of processing.SOLUTION: A sonar image processor of the present invention comprises: a search unit for extracting from a sonar image a plurality of partial images estimated to represent a target; a path difference calculation unit for calculating a path difference between a direct path signal reflected by the target and a sea surface reflected signal reflected by the target and search surface; a similarity determination unit for extracting, from the plurality of partial images, at least one set of the partial images having a relationship that corresponds to the path difference, and calculating the degree of similarity between the at least one set of extracted partial images to determine whether the at least one set of partial images are partial images representing the same target; and an identification unit for identifying the target on the basis of the determination of the degree of similarity from the at least one set of partial images having the relationship that corresponds to the path difference.SELECTED DRAWING: Figure 2

Description

  The present invention relates to a sonar image processing technique.

  As a technology to search for an underwater target, an acoustic beam is radiated from a sonar installed in a moving body such as an underwater towed vehicle or underwater vehicle, and a reflected wave (echo) reflected by the target is emitted. There is a technique for searching for a target by receiving and analyzing it. A technique for searching for a target using a sonar is disclosed in Patent Document 1, for example.

  For example, in the seafloor search using a side-scan sonar (a side-scan sonar, synthetic aperture sonar, etc.) that uses a sharp horizontal directivity acoustic beam, a target near the seabed based on echoes and shadows appearing in the image. The presence or absence of is determined.

  In the side scan type sonar, the transmission sound applied to the target near the sea bottom is scattered in the vertical direction. Therefore, the echoes of the target received by the receiving array include not only signals that return on the shortest straight path, but also signals that return after being reflected on the sea surface. Since the signal that arrives via the sea surface reflection arrives later in time than the direct reception signal, it appears at a position delayed in the time direction on the image, but the characteristics of both signals are similar, This may cause a false alarm when determining the presence or absence of the target.

  In general side scan sonars and synthetic aperture sonars, sonar sounds with a relatively high frequency are used. Therefore, in order to reduce the influence of such sea surface reflection, vertical directivity is formed in the receiving array. Yes. That is, by forming a sharp vertical directivity and directing the principal axis direction of the beam toward the sea bottom (lower side), the level of the signal coming from the sea surface (upper side) is suppressed so that it does not appear in the image. This is because a general side scan sonar, synthetic aperture sonar, etc. uses a sonar sound of a relatively high frequency, so that a receiving array having a vertical directivity can be manufactured in a small size and light weight. This is because it is easy to mount on an underwater vehicle.

  As another method, there is a method in which a signal arrival direction is separated into upper and lower by interferometry processing, and a signal coming from the upper side is determined as a false alarm. In the interferometry processing, the receiving array has two or more upper and lower structures, and the arrival direction of the signal or the seabed altitude of the target is calculated based on the phase difference between the upper and lower receiving arrays. A technique using interferometry is disclosed in Non-Patent Document 1, for example.

Patent Document 2 states that “When a receiver receives a direct wave and a multipath wave together, the phase becomes more unstable than when only a direct wave is used, and the accuracy of estimation of the aspect, azimuth, and overall length of the target deteriorates. For this reason, it is necessary to improve the estimation accuracy by eliminating multipath waves, but in this case, the multipath wave always arrives later than the direct wave, so the target and underwater target detection The propagation delay time is obtained by calculating the propagation path of the multipath waves mf and me based on the positional relationship with the apparatus, and the multipath wave is eliminated based on this propagation delay time ". .

JP 2014-132248 A JP 2012-103054 A

Author Oceanographic Society of Japan, Title of Publication Basics and Applications of Ocean Acoustics (Issue Naruyamado Shoten), Date of Publication April 28, 2004, Description pages 165-166

  The disclosure of the above document is incorporated herein by reference. The analysis of the related technology described above is given below.

  For example, when a sonar sound having a relatively low frequency is used in a side scan type sonar, it is difficult to suppress the signal of the sea surface reflection path due to the vertical directivity of the receiving array.

  The reason is that in order to form vertical directivity with a relatively low frequency receiving array, it is necessary to enlarge the array aperture in the vertical direction. As a result, the receiving array becomes large, and a small underwater towed vehicle or This is because it is difficult to mount the vehicle on an underwater vehicle.

  For example, when a sonar sound having a relatively low frequency is used in a side scan type sonar, it is difficult to separate and remove an automatic detection result of a sea surface reflection path signal by interferometry processing.

  The reason for this is that the interferometry requires that the receiving array and the receiver have two or more stages in the vertical direction. As a result, the receiving array and the receiver are increased in size, resulting in a small underwater towed vehicle and underwater navigation. This is because it becomes difficult to mount on a running body.

  The present invention has been made in view of the above problems, and one of its purposes is a sonar image processing apparatus and a sonar capable of accurately identifying a target while avoiding enlargement of the apparatus or complicated processing. An image processing method and a sonar image processing program are provided. Other objects and advantages will be apparent to those skilled in the art from this disclosure.

  According to one aspect of the present invention, a detection unit that extracts a plurality of partial images estimated to show a target from a sonar image, a direct path signal reflected by the target, and a reflection by the target and the sea surface. A path difference calculating unit that calculates a path difference with the sea surface reflection signal, and extracting at least one set of the partial images having a relationship corresponding to the path difference from the plurality of partial images, and extracting the at least one extracted A similarity determination unit that calculates similarity between a set of partial images and determines whether the at least one set of partial images is a partial image representing the same target, and corresponds to the path difference There is provided a sonar image processing apparatus comprising: an identification unit that identifies the target based on the similarity determination from the at least one set of partial images having a relationship.

  According to one aspect of the present invention, a plurality of partial images estimated to represent a target object are extracted from a sonar image, and a direct path signal reflected by the target object and a sea surface reflected by the target object and the sea surface are obtained. Calculating a path difference from the reflected signal, extracting at least one set of the partial images having a relationship corresponding to the path difference from the plurality of partial images, and similarity between the extracted at least one set of partial images A degree is calculated to determine whether the at least one set of partial images is a partial image representing the same target, and the similarity is obtained from the at least one set of partial images having a relationship corresponding to the path difference. A sonar image processing method is provided that identifies the target based on a degree determination.

  According to one aspect of the present invention, a function of extracting a plurality of partial images estimated to show a target from a sonar image, a direct path signal reflected by the target, and reflected by the target and the sea surface. A function of calculating a path difference with a sea surface reflection signal, and extracting at least one set of the partial images having a relationship corresponding to the path difference from each other from the plurality of partial images, and extracting the at least one set of parts A function of calculating similarity between images and determining whether the at least one set of partial images is a partial image representing the same target, and the at least one set having a relationship corresponding to the path difference And a sonar image processing program for causing a computer to execute a function of identifying the target based on the determination of the degree of similarity from the partial images. In addition, a computer-readable recording medium (for example, a HDD (Hard Disk Drive), a CD (Compact Disk) / DVD (Digital Versatile Disk), a non-transitory computer readable recording medium such as a semiconductor storage device) that stores the above program is provided. Is done.

  According to one aspect of the present invention, it is possible to accurately identify a target while avoiding enlargement of the apparatus or complicated processing.

It is explanatory drawing at the time of searching the sea bottom vicinity with a side scan type sonar, (a) is a figure which shows the relationship between the sonar seen from the upper part and the target near the sea bottom, (b) is progress It is a figure which shows the relationship between the sonar seen from the direction back, the target near the sea bottom, the sea surface, the direct path signal, and the sea surface reflection path. It is a block diagram of the sonar image processing apparatus of 1st Embodiment. It is a block diagram of the sonar image processing apparatus of 2nd Embodiment. It is a figure explaining operation | movement of the sonar image processing apparatus which concerns on 2nd Embodiment, (a) is an automatic detection result in case the direct path | route signal and sea surface reflection path | route signal of one target appear on the sonar image, It is a figure which shows the relationship of a partial image, (b) is a figure which shows the feature element extracted from a partial image. It is a figure explaining the basic concept of sonar image processing of one embodiment.

  A basic block configuration and processing flow of sonar image processing according to an embodiment will be described. Referring to FIG. 5, the sonar image processing apparatus includes a detection unit 3, a path difference calculation unit 4, a similarity determination unit 5, and an identification unit 6. Each of the units 3 to 6 can be configured by software, hardware, or a combination thereof, alone or in combination.

  The detection unit 3 extracts a plurality of partial images estimated to appear the target from the sonar image (S501).

  The path difference calculation unit 4 calculates the path difference between the direct path signal reflected by the target and the sea surface reflection signal reflected by the target and the sea surface (S502).

  The similarity determination unit 5 extracts at least one set of the partial images having a relationship corresponding to the path difference from the plurality of partial images, and calculates the similarity between the extracted at least one set of partial images. Then, it is determined whether or not the at least one set of partial images is a partial image representing the same target (S503).

  The identification unit 6 identifies the target based on the similarity determination from the at least one set of partial images having a relationship corresponding to the path difference (S504). Preferably, the identification unit 6 can falsely remove the automatic detection result based on the sea surface reflection path signal appearing in the sonar image and output only the detection result based on the direct path signal.

  According to such sonar image processing, it is possible to search for a target with high accuracy while avoiding enlargement of the apparatus or complication of processing by similarity determination using a sea surface reflection path signal. For example, without using a large receiving array or interferometry processing, the automatic detection result of the sea surface reflection path signal that appears in the sonar image is removed by false alarm, and only the detection result of the direct path is output. The target can be searched with high accuracy.

  Such a sonar image processing apparatus is preferably applied to a side-scan type sonar, and can perform a predetermined process on the sonar image and output an image area that is considered to be a target as an automatic detection result.

<Embodiment>
Exemplary embodiments of the present invention will be described with reference to the drawings.

  With reference to FIGS. 1 (a) and 1 (b), description will be given of a case where the vicinity of the sea bottom is searched by a side-scan sonar that does not have a sharp vertical directivity.

  When using a side-scan sonar, the echo from the target arrives at the receiver from both the direct path and the sea surface reflection path, as described in the section of the problem to be solved by the invention, and appears on the echo image. May show two automatic detection results (partial images) corresponding to the target.

  Examples of sonar images and automatic detection results in this case are shown in FIGS. FIG. 1A is a diagram showing a relationship between a sonar viewed from above and a target near the sea bottom. FIG. 1B is a diagram showing the relationship between the sonar, the target near the sea bottom, the sea surface, the direct path signal, and the sea surface reflection path as seen from the rear in the traveling direction.

  In FIG. 1B, when the sonar horizontal plane where the sonar is located is used as a reference, the direct path signal comes from the lower side and the sea surface reflection path signal comes from the upper side. Therefore, when a receiving array having a sharp vertical directivity is used, the signal level coming from the upper side of the sonar horizontal plane can be suppressed so that it does not appear on the image.

  Even when a receiving array without sharp vertical directivity is used, the arrival direction in the vertical direction of each signal is calculated using interferometry processing, and the signal arriving from the upper side of the horizontal plane is determined as the sea surface reflection path. In some cases, it can be excluded from the automatic detection result.

  However, in a sonar image processing device that does not have both sharp vertical directivity and interferometry processing functions, the automatic detection result of both from the sonar image in which the echo by the direct path signal and the echo by the sea surface reflection path signal are mixed. Or to identify the targets 1 and 2.

<First Embodiment>
Therefore, in the first embodiment or the like, by evaluating the relative distance difference and the image similarity with respect to any two detection results among the automatic detection results, a pair of detection results can be obtained as a direct path / It is determined that the path is a sea surface reflection path, and the detection result by the sea surface reflection path is removed from the automatic detection result.

  FIG. 2 shows a block diagram of the sonar image processing apparatus according to the first embodiment. The sonar image processing apparatus 10 is applied to a side-scan type sonar, and can perform a predetermined process on the sonar image and output an image area that seems to be a target as an automatic detection result.

  The sonar image processing apparatus 10 includes an image generation unit 20, an automatic detection unit 30, an underwater environment measurement unit 40, a relative distance difference calculation unit 50, a distance difference evaluation unit 60, and an image similarity evaluation unit 70.

  The automatic detection means 30 corresponds to or is included in the detection unit 3 in FIG. The relative distance difference calculation unit 50 and the distance difference evaluation unit 60 correspond to or are included in the path difference calculation unit 4 of FIG. The image similarity evaluation means 70 corresponds to or is included in the similarity determination unit 5 and the identification unit 6 in FIG.

  The image generation means 20 generates a sonar image from the sonar received signal acquired by the side scan method using general signal processing and image processing. For example, in the case of a side scan sonar, pixel information for one bin in the traveling direction is generated by forming a beam having a sharp horizontal directivity with a received signal obtained by one transmission / reception. A sonar image is obtained by repeating this as many times as the number of times of transmission and reception and arranging in the traveling direction.

  The automatic detection means 30 automatically extracts a partial area that seems to be an echo or shadow of the target from the sonar image. The extraction result is output as an automatic detection result 80. The number of automatic detection results is not limited to one for one image, and a large number of detection results (= D1, D2,... Dn) are generally obtained from one image.

  In the present embodiment, it is assumed that two detection results (DT1, DT2) are extracted for one target. Note that detection results obtained from objects other than the target and false alarms caused by environmental factors such as seabed reverberation and noise are also included in the automatic detection results 80 (= D1, D2,... Dn).

  The automatic detection means 30 can extract a partial region in an image using an automatic detection means used in a general sonar image processing apparatus. For example, in a side scan sonar and a synthetic aperture sonar, it is often determined whether or not there is a target in the region based on the strength or shape of a shadow generated on the far range side of the target. Further, not only the shadow but also the intensity and shape of the bright echo part may be referred to.

  The underwater environment measuring means 40 has an underwater sensor such as a seabed altimeter, a depth gauge, and a water temperature gauge used in a general side scan sonar. The underwater environment measuring unit 40 sequentially outputs the measurement result obtained from the underwater sensor to the relative distance difference calculating unit 50. Here, “sequential output” means to continuously measure and output the result during operation of the sonar, and it is sufficient that the output cycle is approximately equal to the transmission / reception cycle of the sonar.

  The relative distance difference calculation means 50 calculates the relative distance difference (path difference) for each azimuth position using the seabed altitude, depth, and underwater sound speed acquired at the position closest to the position in the image traveling direction (azimuth position). Do.

  For each range / azimuth position (coordinates R, X) on the image, the relative distance difference ΔR between the direct path signal and the sea surface reflection path signal can be calculated as follows.

GR = sqrt {(c ・ ΔT / 2) ² -h ² }
ΔR = sqrt {GR ² + (h + 2d) ² } -sqrt (GR ² + h ² )
GR: Ground range distance ΔT: Round trip propagation time of the target of ground range distance GR
c: Underwater sound speed measured at the time closest to azimuth position X
h: Seafloor height measured at the time closest to azimuth position X
d: Depth measured at time closest to azimuth position X

The distance difference evaluation means 60 assumes that one of the n detection results (Di) included in the automatic detection result 80 is a direct path signal, and its range azimuth position (coordinate R _i, X _i ) calculates the predicted position of the sea surface reflection path signal (coordinates R _i + ΔR, X _i ) and other detection results (D ₁ , D ₂ , ... D _i-1 , D _{i + 1} , ... D _n ) all relative distance differences L _ij (j = 1,2, ... i-1, j + 1, ... n) from the range / azimuth position are calculated.

And predicted position of the detection result D _i, detection result D _j (range azimuth coordinates = R _j, X _j) relative distance difference L _ij between is expressed as follows.

L _ij = sqrt {(R _i + ΔR-R _j ) ² + (X _i -X _j ) ² }

The distance difference evaluation means 60 compares the calculated L _ij with a preset distance difference determination threshold L _th , and if L _ij <L _th , the detection result D _i and the detection result D _j are Then, it is determined that there is a high possibility that the direct path signal and the sea surface reflection path signal (a pair of signals) related to the same target are output, and two detection results are output to the image similarity evaluation means 70. When L _ij ≧ L _th , the distance difference evaluation means 60 performs distance difference determination for the next detection result D _{i + 1} .

The image similarity evaluation unit 70 refers to the partial images of the two detection results D _i and D _j input from the distance difference evaluation unit 60 and calculates the cross-correlation coefficient C _ij of both partial images. The image similarity evaluation means 70 compares the calculated cross-correlation coefficient C _ij with a preset correlation determination threshold value C _th . When C _ij > C _th , the image similarity evaluation means 70 determines that the detection result D _i and the detection result D _j are based on the direct path signal and the sea surface reflection path signal from the same target T, and the detection result D the _j is removed from the automatic detection results 80, and registers the detection result D _i as an automatic detection result 81. When C _ij ≦ C _th , the image similarity evaluation means 70 evaluates the similarity for other combinations of detection results.

The sonar image processing apparatus 10 configured as described above has a detection result based on a direct path signal regarding the same target included in the automatic detection result 80 (= D ₁ , D ₂ ,... D _n ), and a sea surface reflection path signal. Among the detection results by, detection results by the sea surface reflection path signal can be extracted and excluded. By applying the processing by the distance difference evaluation means 60 and the image similarity evaluation means 70 to all of the automatic detection results 80, an automatic detection result 81 composed of detection results based on direct path signals is finally registered. .

  Therefore, the sonar image processing apparatus 10 eliminates the automatic detection result by the sea surface reflection path signal passing through both the target and the sea surface as a false alarm without using a sharp vertical directivity receiving array and interferometry processing. Thus, an automatic detection result based on a direct route signal from the target T can be obtained.

  In the present embodiment, the similarity is determined based on the cross-correlation coefficient between the two images. However, the similarity between the two images may be determined by other methods. In the present embodiment, as shown in FIG. 1B, the case where the target is placed on the sea bottom has been described. However, the target that is the search target of the sonar may not necessarily be on the sea bottom. It may float slightly from the bottom of the sea and be underwater, or it may be placed in a shallow and buried state under the bottom of the sea.

Second Embodiment
A second embodiment will be described with reference to FIG. FIG. 3 is a block diagram of the sonar image processing apparatus 100 according to this embodiment.

  Referring to FIG. 3, sonar image processing apparatus 100 includes image generation means 200, automatic detection means 300, underwater environment measurement means 400, relative distance difference calculation means 500, distance difference evaluation means 600, and image similarity evaluation means 700. Is provided.

  The automatic detection means 300 corresponds to or is included in the detection unit 3 in FIG. The relative distance difference calculation unit 500 and the distance difference evaluation unit 600 correspond to or are included in the path difference calculation unit 4 of FIG. The image similarity evaluation unit 700 corresponds to or is included in the similarity determination unit 5 and the identification unit 6 in FIG.

  Similar to the first embodiment, the image generation unit 200 generates a sonar image from the sonar received signal acquired by the side scan method using general signal processing and image processing. Further, based on the latitude / lightness information calculated by the navigation calculation means 210, the range / azimuth coordinate system (R, X) of the sonar image is converted into latitude / longitude coordinates (x, y) and output.

  The navigation calculation means 210 calculates the latitude / longitude at every moment corresponding to the trajectory of the sonar sailing and outputs it to the image generation means 200. As a specific method of the navigation calculation, a method used for a general side scan sonar, synthetic aperture sonar, or the like can be used. For example, in an underwater vehicle, an inertial navigation device is built in, the acceleration in each of the X-Y-Z directions of the vehicle is measured, and time-integrated to calculate the vehicle position every moment. For underwater towed vehicles, the towed vessel on the water is equipped with GPS to measure the latitude and longitude, and the underwater positioning device that measures the relative position of the underwater towed vessel and the towed vessel is used. The position can be calculated.

  As in the first embodiment, the automatic detection means 300 automatically extracts a partial area that seems to be an echo or shadow of the target from the sonar image. The extraction result is output as an automatic detection result 800. The automatic detection means 300 is different from the automatic detection means 30 of the first embodiment in that the position information of the detection result is expressed not by the range / azimuth coordinates (R, X) but by the latitude / longitude coordinates (x, y). Is a point.

  Similar to the first embodiment, the underwater environment measuring unit 400 includes a depth meter and an underwater sound speed meter used in a general side scan sonar and the like, and sequentially outputs the measurement results to the relative distance difference calculating unit 500. The difference between the underwater environment measuring unit 400 and the underwater environment measuring unit 40 of the first embodiment is that the seafloor altitude information is not measured by the sensor, but is replaced by the seabed landform reference unit 410.

  The submarine landform reference means 410 holds detailed depth information of the sea area where the sonar is operated in a latitude / longitude map format, and outputs the sea area depth at the designated latitude / longitude position. The latitude / longitude map of the sea area depth may be a general nautical chart database format widely used in ocean surveys, but is not limited thereto.

The relative distance difference calculation means 500 assumes that one detection result (D _i ) out of n detection results included in the automatic detection result 800 is a direct route signal, and its latitude / longitude position (x _i, the predicted position of the sea surface reflection path signal _{(x 'i, y' i} ) calculated in y _i), and outputs the result to the distance difference evaluation unit 600.

As a calculation method of the predicted position of the sea surface reflection path signal on the image, the relative position is once calculated in the range / azimuth coordinate system (R _i + ΔR, X _i ), as in the relative distance difference calculation means 50 of the first embodiment. A method for obtaining and then converting to latitude / longitude coordinates in the same manner as the image generation means 200 can be used, but is not limited thereto.

The distance difference evaluation unit 600 includes the predicted position (x ′ _i , y ′ _i ) of the sea surface reflection path signal of the detection result (D _i ) selected by the relative distance difference calculation unit 500 and other detection results 800 included in the automatic detection result 800. Referring to the latitude and longitude positions of n-1 detection results (D ₁ , D ₂ , ... D _i-1 , D _{i + 1} , ... D _n ), their relative distances L _ij (j = 1,2, ... i-1, j + 1, ... n)

And predicted position of the detection result D _i, detection result D _j (latitude and longitude position coordinates = x _j, y _j) relative distance difference L _ij between is expressed as follows.

L _ij = sqrt {(x ' _i -x _j ) ² + (y' _i -y _j ) ² }

The distance difference evaluation means 600 compares L _ij with a preset distance difference determination threshold L _th , and when L _ij <L _th , the detection result D _i and the detection result D _j are the direct paths related to the same target. It is determined that there is a high possibility that the detection result is a pair of detection results composed of the signal and the detection result of the sea surface reflection path signal, and the two (pair) detection results are output to the image similarity evaluation means 700.

The image similarity evaluation unit 700 refers to the partial images of the pair of detection results D _i and D _j input from the distance difference evaluation unit 600, and determines k feature elements (F1, F2,. .Fk). Here, feature elements obtained from the partial image of the detection result Di are F1 _i , F2 _i , ... Fk _i , and feature elements obtained from the partial image of the detection result D _j are F1 _j , F2 _j , ... Fk _j .

  There are various types of image feature elements. For example, various types of information such as the level, size, and shape of parts that appear to be echoes and shadows of the target in the image, and images that have been machine-learned with the target image acquired in advance. The degree of coincidence with the pattern, etc. can be cited as a feature element.

  Here, as a relatively simple example, a case where a target appears on an image as a line segment having a certain length will be described with reference to FIG. As specific feature elements in this example, there are two lengths and inclinations of line segments.

FIG. 4 (a), detection results range R _i by the direct path signal of the target, detection determined by the sea surface reflection path signal appears in the vicinity range R _j, respectively detection result D _i, the partial image P _i as D _j, An example in which P _j is extracted is shown.

FIG. 4B illustrates an example of partial images P _i , P _j and feature elements of the detection results D _i , D _j , and two specific feature elements, a line segment length F1 and a slope F2, are shown. Has been extracted. In other words, the line segment length of the detection result D _i is F1 _i , the slope is F2 _i , the line segment length of the detection result D _j is F1 _j , and the slope is F2 _j , and specific values are read from the partial images, respectively. It is done.

The image similarity evaluation unit 700 calculates the difference between the feature elements of the detection results D _i and D _j and compares the difference with a preset feature difference threshold to determine the similarity between the two. In this example, the image similarity evaluation unit 700 determines that the degree of similarity between the line segment length F1 and the inclination F2 is high when the following expression is satisfied.

(| F1 _i -F1 _j | <F1 _th ) & (| F2 _i -F2 _j | <F2 _th )
F1 _th : Threshold value for line length F1
F2 _th : Threshold value for slope F2

When it is determined that the similarity between the detection results D _i and D _j is high, the image similarity evaluation unit 700 deletes the detection result D _j from the automatic detection result 800.

  For the automatic detection result 800 output by the automatic detection means 300, the underwater environment measurement means 400, the seabed topography reference means 410, the relative distance difference calculation means 500, the distance difference evaluation means 600, and the image similarity evaluation means 700 are described above. By repeatedly applying the series of processes, it is possible to obtain an automatic detection result 810 from which all detection results by the sea surface reflection path signal are excluded.

  In the present embodiment, the image similarity evaluation unit 700 determines the similarity using the line segment length and the inclination as the feature elements. However, the image similarity may be determined using other feature elements. Further, the number of feature elements is not limited to two, and determination may be made using one or three or more feature elements.

In this embodiment, the image similarity evaluation unit 700 extracts feature elements from the automatic detection result 800 obtained by the automatic detection unit 300. The feature elements are included in the automatic detection method of the automatic detection unit 300. Extraction may be incorporated. The automatic detection result 800 is output in a form including the extraction result of predetermined feature element information. Therefore, the image similarity evaluation means 700 can perform only the threshold determination of the feature element between the detection results D _i and D _j without extracting the feature element again.

  In this embodiment, the sonar image is defined by latitude / longitude coordinates using the navigation calculation means 210, but may be defined by other coordinate systems. For example, the position at the start of operation of the sonar may be defined as the origin, the true north direction may be defined as the X axis, and the true east direction may be defined as the Y axis.

  In the first and second embodiments, the underwater sound speed measured by the underwater environment measuring means 40 and 400 is used for the range distance calculation. However, the underwater sound speed previously measured by another means may be used. In this case, there is a possibility that the actual underwater sound speed in the actual environment does not necessarily match, and the error included in the range distance calculation result increases.

  It should be noted that the disclosures of the above patent documents are incorporated herein by reference. Within the scope of the entire disclosure (including claims) of the present invention, the embodiments and examples can be changed and adjusted based on the basic technical concept. Various combinations or selections of various disclosed elements (including each element of each claim, each element of each embodiment, each element of each drawing, etc.) are possible within the scope of the claims of the present invention. . That is, the present invention of course includes various variations and modifications that could be made by those skilled in the art according to the entire disclosure including the claims and the technical idea.

Although not particularly limited, the embodiment, the modification, and the like are appended as follows.
(Appendix 1)
An automatic detection means that automatically extracts an image area that seems to be a target with reference to the generated sonar image, a sensor that accurately measures the seabed altitude, depth, underwater sound velocity, etc. of the sonar, and the sonar seabed by the sensor. Based on measurement results such as altitude, depth and underwater sound speed, a time difference between the direct path and the sea surface reflection path is calculated, and a relative distance difference between the direct path signal and the sea surface reflection path signal appearing on the sonar image is obtained; Distance difference evaluation means for determining whether the positional relationship between any two detection results matches the relative distance difference with reference to the automatic detection result obtained from the sonar image, and the positional relationship matches the relative distance difference Similarity evaluation means for calculating the similarity of partial images of two arbitrary detection results, and based on the relative distance difference and similarity obtained by the distance difference evaluation means and the similarity evaluation means Of the same as the determined result of detection, sonar image processing apparatus comprising: means for deleting from the automatic detection result false alarm one corresponding to the sea surface reflection path, a.
(Appendix 2)
This method removes only the detection result by the sea surface reflection path signal from the automatic detection result obtained by automatically extracting the image area that seems to be the target with reference to the generated sonar image. The time difference between the direct path and the sea surface reflection path is calculated based on the measurement results such as the underwater sound velocity, and the relative distance difference between the direct path signal and the sea surface reflection path signal appearing on the sonar image is obtained. A pair of detection results whose relative positional relationship matches the relative distance difference is extracted, the similarity of the partial images of the pair of detection results is calculated to determine the identity, and when it is recognized as the same, the sea surface reflection path A sonar image processing method for outputting an automatic detection result after deleting one corresponding to the false alarm from the automatic detection result.
(Appendix 3)
A sonar image processing program that can be executed by a computer of a sonar image processing apparatus that automatically extracts an image area that is considered to be a target with reference to a generated sonar image, such as the sonar seabed altitude, depth, underwater sound speed, etc. Based on the measurement results, the time difference between the direct path and the sea surface reflection path is calculated, the relative distance difference between the direct path signal and the sea surface reflection path signal that appears on the sonar image, and the relative position of the automatic detection results A function of extracting a pair of detection results that match the relative distance difference, a function of calculating the similarity of partial images of the pair of detection results and determining the identity, and if it is recognized as the same, corresponds to a sea surface reflection path A sonar image processing program for causing a computer to execute a function of outputting an automatic detection result after deleting one of the two as a false alarm from the automatic detection result.

DESCRIPTION OF SYMBOLS 1 Target object 2 Target object 3 Detection part 4 Path difference calculation part 5 Similarity determination part 6 Identification part 10 Sonar image processing apparatus 20 Image generation means 30 Automatic detection means 40 Underwater environment measurement means 50 Relative distance difference calculation means 60 Distance difference evaluation Means 70 Image similarity evaluation means 80 Automatic detection result (including false alarm by sea surface reflection path signal)
81 Automatic detection results (not including false alarms due to sea surface reflection path signals)
DESCRIPTION OF SYMBOLS 100 Sonar image processing apparatus 200 Image generation means 210 Navigation calculation means 300 Automatic detection means 400 Underwater environment measurement means 410 Seabed topography reference means 500 Relative distance difference calculation means 600 Distance difference evaluation means 700 Image similarity evaluation means 800 Automatic detection result (sea surface (Including false alarms due to reflection path signals)
810 Automatic detection results (not including false alarms due to sea surface reflection path signals)

Claims (10)

A detection unit that extracts a plurality of partial images estimated to show the target from the sonar image;
A path difference calculation unit for calculating a path difference between the direct path signal reflected by the target and the sea surface reflection signal reflected by the target and the sea surface;
Extracting at least one set of the partial images having a relationship corresponding to the path difference from each other from the plurality of partial images, calculating a similarity between the extracted at least one set of partial images, and calculating the at least one set of the partial images; A similarity determination unit that determines whether or not partial images are partial images representing the same target object;
A sonar image processing apparatus comprising: an identification unit that identifies the target based on the determination of the similarity from the at least one set of partial images having a relationship corresponding to the path difference.   The sonar image processing apparatus according to claim 1, wherein the identification unit is configured to remove the sea surface reflection signal from a detection result of the sonar.   The identifying unit is configured to delete at least one partial image by the sea surface reflection signal based on at least the similarity from at least one set of the partial images having a relationship corresponding to the path difference. The sonar image processing apparatus according to claim 1 or 2, characterized in that:   In order to extract at least one set of the partial images having a relationship corresponding to the path difference, the positional relationship between at least two of the partial images corresponds to the path difference between the direct path signal and the sea surface reflection signal. The sonar image processing apparatus according to claim 1, further comprising a path difference evaluation unit that determines whether or not.   The route difference calculation unit is configured to calculate the route difference based on an underwater environment condition including at least one of sea bottom altitude, depth, and underwater sound speed. The sonar image processing apparatus according to one.   6. The route difference calculation unit according to claim 1, wherein the route difference calculation unit includes bottom altitude information of a sonar operation sea area in a map format, and is configured to calculate the route difference based on at least the sea bottom height information. The sonar image processing apparatus according to any one of the above.   The similarity determination unit calculates a correlation coefficient between at least one set of the partial images, and determines the similarity based on a magnitude relationship between the correlation coefficient and a predetermined threshold value. The sonar image processing apparatus according to claim 1, wherein the sonar image processing apparatus is configured.   The similarity determination unit is configured to extract feature elements of at least one set of the partial images, and to determine the similarity based on a magnitude relationship between a difference between the extracted feature elements and the threshold value. The sonar image processing apparatus according to any one of claims 1 to 7. Extract multiple partial images estimated to reveal the target from the sonar image,
Calculating the path difference between the direct path signal reflected by the target and the sea surface reflection signal reflected by the target and the sea surface;
Extracting at least one set of the partial images having a relationship corresponding to the path difference from each other from the plurality of partial images, calculating a similarity between the extracted at least one set of partial images, and calculating the at least one set of the partial images; Determining whether a partial image is a partial image representing the same target,
A sonar image processing method, wherein the target is identified from the at least one set of partial images having a relationship corresponding to the path difference based on the similarity determination. A function to extract a plurality of partial images estimated to show the target from the sonar image;
A function of calculating a path difference between a direct path signal reflected by the target and a sea surface reflection signal reflected by the target and the sea surface;
Extracting at least one set of the partial images having a relationship corresponding to the path difference from each other from the plurality of partial images, calculating a similarity between the extracted at least one set of partial images, and calculating the at least one set of the partial images; A function of determining whether the partial images are partial images representing the same target;
A sonar image processing program that causes a computer to execute a function of identifying the target based on the determination of the similarity from the at least one set of partial images having a relationship corresponding to the path difference.

Images