CN114019469B

CN114019469B - Sea surface target detection method and equipment based on concave packet judgment, medium and product

Info

Publication number: CN114019469B
Application number: CN202111272230.XA
Authority: CN
Inventors: 关键; 伍僖杰; 丁昊; 董云龙; 刘宁波; 王国庆; 黄勇; 于恒力
Original assignee: School Of Aeronautical Combat Service Naval Aeronautical University Of Pla
Current assignee: School Of Aeronautical Combat Service Naval Aeronautical University Of Pla
Priority date: 2021-10-29
Filing date: 2021-10-29
Publication date: 2024-06-21
Anticipated expiration: 2041-10-29
Also published as: CN114019469A

Abstract

The invention provides a sea surface target detection method based on concave packet judgment, equipment, medium and product, wherein the sea surface target detection method based on concave packet judgment comprises the following steps: obtaining a convex hull judgment space corresponding to the characteristic points of the historical sea clutter samples; the convex hull judgment space characterizes that the characteristic points are distributed in the interior and each vertex of a mesh topological structure formed by triangulation planes; performing preset convex-concave conversion treatment on the convex hull judgment space to obtain a concave hull judgment space; the preset convex-concave conversion method comprises the steps of removing the longest edge in the convex hull judgment space and not removing the characteristic points; and determining the feature points to be detected as sea surface small targets according to the relation between the feature points to be detected and the concave packet judgment space. The method can realize the purpose of uniformly distributing the sea clutter samples in the gathering area, thereby improving the detection efficiency and the detection probability of small targets on the sea surface.

Description

Sea surface target detection method and equipment based on concave packet judgment, medium and product

Technical Field

The invention relates to the technical field of sea surface target detection, in particular to a sea surface target detection method and equipment based on concave packet judgment, a medium and a product.

Background

With the improvement of radar resolution and the increase of target types to be observed, the military and civil fields both put forward higher requirements on the capability of radar to detect targets, especially for small target detection under the ocean background, the space-time variation characteristics of sea clutter are complex, and the small targets have weak echoes, so that the signal clutter is smaller, and the difficulty is increased for accurately detecting the small targets on the sea surface. Therefore, how to increase the detection probability of small targets on the sea surface becomes a critical issue to be solved.

In the existing sea surface small target detection method, the sea surface small target detection problem can be generally classified into a binary hypothesis test problem based on whether a unit to be detected comprises sea clutter and noise signals, and a convex hull judgment space is designed by using a sea clutter sample set and a convex hull algorithm when the sea clutter samples are uniformly distributed, if the characteristic points of the unit to be detected fall into the convex hull judgment space, the unit to be detected is regarded as sea clutter, and the assumption H ₀ is established; otherwise, the sea surface is regarded as a small target, and H ₁ is assumed to be true.

However, as the sea clutter samples are not uniformly distributed in practice, and the separability between the target samples and the sea clutter samples is not outstanding when the signal clutter ratio is small, a convex hull judgment space designed by adopting a convex hull algorithm can have an obvious blank area, so that the sea surface small target detection result obtained by using the convex hull judgment space is inaccurate, and the detection probability of the sea surface small target is also not high.

Disclosure of Invention

The invention provides a sea surface target detection method, equipment, medium and product based on concave packet judgment, which are used for solving the defects of inaccurate sea surface small target detection result and low detection probability caused by the fact that sea clutter samples are required to be uniformly distributed and the separability between the target samples and the sea clutter samples is not prominent when the sea clutter samples are smaller in the prior art, and achieving the purpose of reliably and accurately detecting the sea surface small target on the premise that the sea clutter samples are not required to be uniformly distributed.

The invention provides a sea surface target detection method based on concave packet judgment, which comprises the following steps:

Obtaining a convex hull judgment space corresponding to the characteristic points of the historical sea clutter samples; the convex hull judgment space characterizes that the characteristic points are distributed in the interior and each vertex of a mesh topological structure formed by triangulation planes;

Performing preset convex-concave conversion treatment on the convex hull judgment space to obtain a concave hull judgment space; the preset convex-concave conversion method comprises the steps of removing the longest edge in the convex hull judgment space and not removing the characteristic points;

and determining the feature points to be detected as sea surface small targets according to the relation between the feature points to be detected and the concave packet judgment space.

According to the sea surface target detection method based on concave-convex judging, the preset convex-concave conversion treatment is carried out on the convex-convex judging space to obtain the concave-convex judging space, and the sea surface target detection method comprises the following steps:

Determining a first internal feature point set of the convex hull decision space; the first internal feature point set comprises feature points distributed in the convex hull judgment space;

Based on the first internal feature point set, performing preset convex hull inner section operation on the convex hull judgment space to obtain an inner section judgment space;

Selecting a second internal feature point set which is removed in the process of executing the preset convex hull inner section operation from the first internal feature point set;

And based on the second internal characteristic point set, performing preset concave inclusion external compensation operation on the internal profile judgment space to obtain a concave inclusion judgment space.

According to the sea surface target detection method based on concave hull judgment provided by the invention, based on the first internal feature point set, the convex hull judgment space is subjected to preset convex hull inner section operation to obtain an inner section judgment space, and the method comprises the following steps:

Determining a first triangulation surface and a second triangulation surface of the convex hull decision space; the first triangulation surface and the second triangulation surface share a target longest edge;

determining a target internal feature point closest to the midpoint of the longest target edge in the first internal feature point set;

determining J new triangulation planes formed by four vertexes of the first triangulation plane, the second triangulation plane and the target internal feature points; wherein none of the J new triangulation planes contains the target longest edge;

and removing the first triangulation surface and the second triangulation surface from the convex hull judgment space, and generating an inner dissection judgment space after adding the J new triangulation surfaces.

According to the sea surface target detection method based on concave packet judgment provided by the invention, based on the second internal characteristic point set, the concave packet external compensation operation is performed on the internal profile judgment space to obtain a concave packet judgment space, and the method comprises the following steps:

determining J distances between second internal feature points in the second internal feature point set and the J new triangulation planes respectively for the new triangulation planes in the J new triangulation planes;

when the distances between the second internal feature point and the new triangulation surface are smaller than the J distances, determining the second internal feature point as an external compensation feature point to be concave;

And removing a third triangulation surface closest to the external complement feature point to be concave from the internal subdivision judgment space, and generating a concave subdivision judgment space after adding a fourth triangulation surface consisting of the external complement feature point to be concave and the vertex of the third triangulation surface.

According to the sea surface target detection method based on concave hull judgment, the determining of the first triangulation surface and the second triangulation surface of the convex hull judgment space comprises the following steps:

Determining the sum of the side lengths of all triangulation surfaces in the convex hull judgment space, and calculating a judgment threshold based on the sum of the side lengths;

when the sum of the side lengths is larger than the judgment threshold, selecting a first triangulation surface with the sum of the three side lengths as the maximum value from the convex hull judgment space;

a target longest edge of the first triangulation surface is determined, and a second triangulation surface comprising the target longest edge is determined.

According to the sea surface target detection method based on concave hull judgment provided by the invention, the second internal feature point set which is removed in the process of executing the preset convex hull inner section operation is selected from the first internal feature point set, and the method comprises the following steps:

Acquiring a plurality of tetrahedrons formed by four vertexes of the first triangulation surface, the second triangulation surface and the target internal feature points;

A second set of internal feature points of the first set of internal feature points distributed among the plurality of tetrahedrons is determined.

According to the sea surface target detection method based on concave-convex hull judgment, the convex hull judgment space corresponding to the characteristic points of the obtained historical sea clutter samples comprises the following steps:

acquiring a characteristic point set of a historical sea clutter sample, and calculating a preset false alarm number according to the number of the characteristic points in the characteristic point set and a preset false alarm rate;

Determining the gravity centers of the feature point sets, and calculating the Euclidean distance between each feature point in the feature point sets and the gravity centers;

Based on the Euclidean distance, eliminating L characteristic points meeting a preset distance relation between the characteristic points and the gravity center from the characteristic point set to obtain a target characteristic point set; wherein L is the preset false alarm number;

And performing triangulation operation on the target characteristic point set to obtain a convex hull judgment space.

The invention also provides an electronic device comprising a memory, a processor and a computer program stored on the memory and operable on the processor, wherein the processor implements the steps of the sea surface target detection method based on the concave packet judgment as described in any one of the above when executing the program.

The invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of a method for sea surface target detection based on a dishing decision as described in any of the above.

The invention also provides a computer program product comprising a computer program which when executed by a processor implements the steps of a method for sea surface target detection based on a dishing decision as described in any of the above.

According to the sea surface target detection method based on concave bag judgment, a convex bag judgment space corresponding to the characteristic points of historical sea clutter samples is firstly obtained, then preset convex-concave conversion is carried out on the convex bag judgment space to obtain a concave bag judgment space, and finally the characteristic points to be detected are determined to be sea surface small targets according to the relation between the characteristic points to be detected and the concave bag judgment space. Because the characteristic points of the convex hull judgment space representation history sea clutter samples are distributed in the interior and each vertex of a mesh topological graph formed by triangulation planes, the preset convex-concave conversion processing of removing the longest edge operation in the convex hull judgment space and not eliminating the characteristic points is performed, the obtained concave hull judgment space can meet the purpose that the sea clutter samples are uniformly distributed in an aggregation area, and therefore the accuracy and the reliability of determining the characteristic points to be detected as sea surface small targets by adopting the concave hull judgment space are greatly improved, the detection efficiency is improved, and the detection probability of the sea surface small targets is also improved.

Drawings

In order to more clearly illustrate the invention or the technical solutions of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.

FIG. 1 is a schematic flow chart of a sea surface target detection method based on concave packet judgment;

FIG. 2 is a second flow chart of the method for detecting sea surface targets based on concave packet judgment;

FIG. 3 is a schematic diagram showing the comparison of the effects of the convex hull algorithm and the concave hull method provided by the invention;

FIG. 4 is a diagram showing the comparison of the detection results of the convex hull judgment space and the concave hull judgment space provided by the invention;

FIG. 5 is a schematic diagram of the structure of the sea surface target detection device based on the concave packet judgment;

Fig. 6 is a schematic structural diagram of an electronic device provided by the present invention.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

For sea-surface small target detection, it is assumed that the marine radar transmits a series of coherent pulses within one beam and receives an echo time sequence x (K) of length K at each range bin. If the echo information is not affected by the target, the unit to be detected should only contain sea clutter and noise signals c (k), otherwise, the target signals s (k) should be mixed. Based on this, the target detection in sea clutter can be attributed to a binary hypothesis testing problem as in equation (1).

Because the number of the target samples which can be obtained in the actual sea surface environment is far less than that of sea clutter samples, and the distribution characteristics of different targets in the feature space are different, the problem of sea surface small target detection is often converted into the problem of single-class classifier design under an abnormal detection frame in the actual environment, namely, the feature vector belonging to a sea clutter unit is regarded as conventional observation, the feature vector belonging to the target unit is regarded as abnormal observation, and the distribution area of the existing sea clutter samples is divided according to a given false alarm rate P _F, so that a judgment space is formed. If the characteristic points of the unit to be detected fall into the judgment space, the unit to be detected is regarded as a conventional observation point, and the assumption H ₀ is established; otherwise, it is regarded as an abnormal observation point, and H ₁ is assumed to be true.

The key to the design of a single class classifier is the formation of the decision space Ω. When the probability distribution p (ζ|h ₁)、p(ξ|H₀) of the target samples and the sea clutter samples in the feature space is known, the formation of the decision space can be described as the following problem according to the niemann-pearson criterion:

In the formula (2), P _D represents the detection probability, and P _F represents the false alarm rate.

Then, assuming that the total distribution space of the sea clutter samples and the target samples is Θ, and the sea clutter samples and the target samples are uniformly distributed in the respective aggregation areas, the detection probability P _D is equivalent to 1- |Ω|/|Θ|, the formula (2) can be simplified after ignoring the unknown factor P (ζ|h ₁):

min{|Ω|},s.t.∫∫∫_Ωp(ξ|H₀)dξ＝1-P_F (3)

in formula (3), i·| represents the volume of space.

In order to solve the problem of equation (3), the decision space Ω can be limited to a limited convex hull space, and it is noted that only sea clutter samples are available for actual detection, and the limitation in equation (3) can be replaced by equation (4) when the number I is sufficiently large.

In equation (4), # { i, ζ _i εΩ } represents the number of sea clutter samples that fall into the decision space Ω. Equation (2) can be further simplified to equation (5).

In the formula (5), C is a set of all convex hull spaces in the feature space.

However, the simplification of the problem of equation (2) is based on the uniform distribution of sea clutter samples, and this assumption is not necessarily fully true in practical detection. When the signal ratio is small, the separability between the target sample and the sea clutter sample is not prominent, and the distribution of the sea clutter sample may have a certain geometric shape. At this time, a significant neutral region appears in the sample distribution space obtained by adopting the convex hull algorithm, which is obviously unreasonable. Therefore, the detection result of the small sea surface target obtained by using the convex hull judgment space is inaccurate, and the detection probability of the small sea surface target is also low.

Therefore, based on the above-mentioned problems, the present invention provides a sea surface target detection method based on concave packet judgment, and the implementation main body of the sea surface target detection method based on concave packet judgment may be a sea surface target detection device based on concave packet judgment, and the sea surface target detection device based on concave packet judgment may be implemented in a mode of software, hardware or a combination of software and hardware to become part or all of terminal equipment. Alternatively, the terminal device may be a personal computer (Persodal Computer, PC), a portable device, a notebook computer, a smart phone, a tablet computer, a portable wearable device, or other electronic devices, such as a tablet computer, a mobile phone, or the like. The invention is not limited to the specific form of the terminal device.

It should be noted that, the execution body of the method embodiment described below may be part or all of the terminal device described above. The following method embodiments are described taking an execution body as a terminal device as an example.

Fig. 1 is a schematic flow chart of a sea surface target detection method based on concave packet judgment, as shown in fig. 1, and the sea surface target detection method based on concave packet judgment comprises the following steps:

step S110, obtaining convex hull judgment spaces corresponding to characteristic points of historical sea clutter samples; the convex hull decision space characterizes the characteristic points to be distributed in the interior of a mesh topological structure formed by triangulation planes and each vertex.

In particular, since the sea clutter refers to backward scattering echoes on the sea surface after the radar beam irradiates the sea surface, the characteristics of the sea clutter are much more complex than those of the ground clutter and the weather clutter, and the existence of the sea clutter has serious influence on the target detection and positioning tracking performance of the radar. Therefore, when the judgment space generated for the sea clutter data is accurate enough and reliable enough, the detection probability of the subsequent sea surface small target can be greatly increased.

In addition, for the feature point set S ₀ of the obtained historical sea clutter sample, the historical sea clutter sample of the historical sea clutter data may be generated by using MATLAB software, then the feature point set S ₀ of the historical sea clutter sample may be extracted based on the existing feature extraction method, and the feature point set S ₀ includes W feature points, and each feature point may include at least one of echo fluctuation and modulation spectrum characteristics, direct example features, bispectral features, spectral domain features, center distance features, scattering center features, linear prediction coding features, and other features. And then generating a convex hull judgment space corresponding to the characteristic points of the historical sea clutter samples based on the existing convex hull generation algorithm, so that the characteristic points of the historical sea clutter samples are distributed at the vertexes of all triangulation planes in the convex hull judgment space and wrapped in the convex hull judgment space.

And step S120, performing preset convex-concave conversion treatment on the convex hull judgment space to obtain a concave hull judgment space.

The preset convex-concave conversion method comprises the operation of removing the longest edge in the convex hull judgment space and not removing the characteristic points.

Specifically, the terminal device may perform multiple longest edge removal operations and feature point external spatial compensation operations for the convex hull decision space, for example, each long edge in the convex hull decision space that may cause that sea clutter samples cannot be uniformly distributed in the aggregation area may be set first, then perform the deletion operations for each triangulation surface sharing each long edge in the convex hull decision space, then further generate a new triangulation surface based on feature points related to each deleted triangulation surface, and finally add the generated new triangulation surface to the convex hull decision space after performing the deletion operations, thereby generating a concave hull decision space.

And step S130, determining the feature points to be detected as sea surface small targets according to the relation between the feature points to be detected and the concave packet judgment space.

Specifically, the terminal device may obtain a radar echo signal to be detected, extract a plurality of feature points to be detected from the radar echo signal to be detected according to the existing feature extraction method, then judge the relationship between each feature point to be detected and the concave packet judgment space, when the feature point to be detected is in the concave packet judgment space, indicate that the feature point to be detected is sea clutter, and when the feature point to be detected is out of the concave packet judgment space, indicate that the feature point to be detected is a small sea surface target. Therefore, the purpose of detecting the sea surface small targets of the plurality of feature points to be detected is achieved.

According to the sea surface target detection method based on concave bag judgment, firstly, convex bag judgment spaces corresponding to characteristic points of historical sea clutter samples are obtained, then preset convex bag judgment spaces are subjected to convex-concave conversion treatment to obtain concave bag judgment spaces, and finally, according to the relation between the characteristic points to be detected and the concave bag judgment spaces, the characteristic points to be detected are determined to be sea surface small targets. Because the characteristic points of the convex hull judgment space representation history sea clutter samples are distributed in the interior and each vertex of a mesh topological graph formed by triangulation planes, the preset convex-concave conversion processing of removing the longest edge operation in the convex hull judgment space and not eliminating the characteristic points is performed, the obtained concave hull judgment space can meet the purpose that the sea clutter samples are uniformly distributed in an aggregation area, and therefore the accuracy and the reliability of determining the characteristic points to be detected as sea surface small targets by adopting the concave hull judgment space are greatly improved, the detection efficiency is improved, and the detection probability of the sea surface small targets is also improved.

Optionally, the implementation process of step S120 includes: firstly, determining a first internal characteristic point set of the convex hull judgment space; the first internal feature point set comprises feature points distributed in the convex hull judgment space; then, based on the first internal feature point set, performing preset convex hull inner section operation on the convex hull judgment space to obtain an inner section judgment space; further, selecting a second internal feature point set which is removed in the process of executing the preset convex hull inner section operation from the first internal feature point set; and finally, based on the second internal characteristic point set, performing a preset concave inclusion external compensation operation on the internal profile judgment space to obtain a concave inclusion judgment space.

Specifically, for the first internal feature point set of the convex hull decision space, since the convex hull decision space is a mesh topology structure formed by triangulation planes, and feature points of historical sea clutter samples are distributed in the interior and each vertex of the mesh topology structure, namely, the feature points of the historical sea clutter samples are wrapped in the interior of the convex hull decision space except for the vertices of each triangulation plane in the convex hull decision space. Therefore, a set formed by the residual characteristic points after the characteristic points at the vertexes of each triangulation surface in the convex hull judgment space are removed can be determined as a first internal characteristic point set IPS; and a set formed by the residual characteristic points of the characteristic points which are correspondingly deleted and distributed at the vertexes of each triangulation surface in the convex hull judgment space in the characteristic points of the historical sea clutter samples can be determined to be a first internal characteristic point set IPS.

And then, determining long sides to be removed and triangulation surfaces to be removed which share the long sides to be removed in the convex hull judgment space according to the relation between the triangulation surfaces to be removed and a preset threshold in the convex hull judgment space, namely removing the long sides to be removed, namely removing the triangulation surfaces to be removed, adding a new triangulation surface which is generated based on the vertexes of the triangulation surfaces to be removed and does not contain the long sides to be removed while removing the triangulation surfaces to be removed in the convex hull judgment space, and generating an inner profile judgment space, wherein the process of generating the inner profile judgment space by the convex hull judgment space is preset convex hull inner profile operation.

Still further, considering that the feature points of some historical sea clutter samples may be excluded from the decision space when the convex hull inner-cut operation is performed, so that the effect of constant false alarm cannot be achieved, the obtained inner-cut decision space cannot improve the detection probability of small sea targets, and the inner-cut decision space needs to be subjected to the preset concave-package outer-cut operation, that is, a third inner feature point set FPS closer to the new triangulation plane is determined in the second inner feature point set RIP,Then removing the nearest triangulation surface S _nearest closest to the third internal feature point in the third internal feature point set FPS from the internal profile judgment space, and generating a concave packet judgment space after adding the latest triangulation surface consisting of the third internal feature point and the vertex of the nearest triangulation surface S _nearest; the process of generating the concave packet judgment space by the inner section judgment space is the preset concave packet outer compensation operation.

According to the sea surface target detection method based on concave hull judgment, the first internal characteristic point set of the convex hull judgment space corresponding to the characteristic points of the historical sea clutter sample is firstly determined, then the constant false alarm is achieved by means of carrying out preset convex hull inner hull operation on the convex hull judgment space based on the first internal characteristic point set and carrying out preset concave hull outer complement operation on the inner hull judgment space obtained after the preset inner hull operation based on the second internal characteristic point set removed in the process of executing the preset convex hull inner hull operation, so that the purpose of effectively preventing the characteristic points containing the historical sea clutter sample from being removed in the preset convex hull inner hull process is achieved while the algorithm efficiency is effectively improved, and the reliability and the accuracy of the concave hull judgment space are improved.

Optionally, the performing, based on the first internal feature point set, a preset convex hull inner section operation on the convex hull decision space to obtain an inner section decision space includes:

Firstly, determining a first triangulation surface and a second triangulation surface of the convex hull judgment space; the first triangulation surface and the second triangulation surface share a target longest edge; then determining a target internal characteristic point cp closest to the midpoint of the longest target edge in the first internal characteristic point set IPS; further determining J new triangulation planes consisting of four vertices of the first triangulation plane, the second triangulation plane and the target internal feature point; wherein none of the J new triangulation planes contains the target longest edge; and finally, removing the first triangulation surface and the second triangulation surface from the convex hull judgment space, and generating an inner dissection judgment space after adding the J new triangulation surfaces.

Specifically, the terminal device first determines a first triangulation plane and a second triangulation plane sharing the longest target edge, then determines a target internal feature point cp closest to a midpoint of the longest target edge from the first internal feature point set IPS, if the longest target edge is marked as a ridge, the first triangulation plane and the second triangulation plane sharing the longest target edge include 4 vertices and the 4 vertices are marked as ①、②、③、④, the ridge is formed by vertices ② and ③ to the longest target edge, at this time, the added J new triangulation planes may be 4 and include a 1 st new triangulation plane formed by using cp and ①、② as vertices, a 2 nd new triangulation plane formed by using cp and ①、③ as vertices, a 3 rd new triangulation plane formed by using cp and ④、② as vertices, and a4 th new triangulation plane formed by using cp and ④、③ as vertices, that is, the internal triangulation space is a space generated by removing the first triangulation plane and the second triangulation plane in the convex hull decision space and adding the 4 new triangulation planes.

According to the sea surface target detection method based on concave hull judgment, the first triangulation surface and the second triangulation surface which share the longest edge of the target are removed from the convex hull judgment space corresponding to the characteristic points of the historical sea clutter samples, and the new triangulation surface formed by the vertexes of the first triangulation surface and the second triangulation surface and the target internal characteristic points is added, so that the purpose that the concave hull judgment space is obtained after the convex hull judgment space is subjected to preset convex hull internal section operation is achieved, the defect that an obvious neutral gear area exists in a sample distribution space obtained by using a convex hull algorithm in the traditional method is effectively overcome, and a foundation is laid for obtaining a more reliable and accurate concave hull judgment space in the follow-up process.

Optionally, the performing, based on the second internal feature point set RIP, a preset concave-bag external compensation operation on the internal profile decision space to obtain a concave-bag decision space, including:

Determining J distances between second internal feature points in the second internal feature point set and the J new triangulation planes respectively for the new triangulation planes in the J new triangulation planes; when the distances between the second internal feature point and the new triangulation surface are smaller than the J distances, determining the second internal feature point as an external compensation feature point to be concave; and removing a third triangulation surface closest to the external complement feature point to be concave from the internal subdivision judgment space, and generating a concave subdivision judgment space after adding a fourth triangulation surface consisting of the external complement feature point to be concave and the vertex of the third triangulation surface.

Specifically, for the t new triangulation surface in the J new triangulation surfaces, firstly calculating J distances between the f second internal feature points in the second internal feature point set RIP and the J new triangulation surfaces respectively, then judging the magnitude relation between the distances between the f second internal feature points and the t new triangulation surfaces and the J distances, if the distances between the f second internal feature points and the t new triangulation surfaces are smaller than the J distances, determining the f second internal feature points as the m external compensation feature points to be recessed, then adding 1 to the values of m and f respectively, and returning to execute the step of calculating the J distances between the f second internal feature points in the second internal feature point set RIP and the J new triangulation surfaces respectively; otherwise, determining that the f second internal feature point does not need to be subjected to concave inclusion external compensation, adding 1 to the value of f, and returning to execute the step of calculating J distances between the f second internal feature point in the second internal feature point set RIP and J new triangulation planes respectively. And obtaining K to-be-concave external compensation characteristic points corresponding to the t new triangulation surface. Wherein, the initial values of t, F and m are all 1, f=1, 2, 3, …, F, k=1, 2, 3, …, J, m=1, 2, 3, …, K is less than or equal to F, F is the total number of the second internal feature points contained in the second internal feature point set, and J is the total number of the new triangulation surface.

And then, sorting the K external complement feature points to be concave from near to far or from far to near according to the distance between the K external complement feature points to be concave and the t new triangulation plane, and marking the sorted K external complement feature points to be concave as a t third internal feature point set FPS _t,FPS_t＝[P_t1,P_t2,…,P_tr,…,P_tK, wherein r=1, 2, 3, … and K.

Further, the 1 third triangulation plane S ^tr _nearest closest to the r third internal feature point P _tr is removed from the internal subdivision decision space, then a fourth triangulation plane composed of the r third internal feature point P _tr and the vertex of the third triangulation plane S ^tr _nearest is added, an internal subdivision decision space after the outer complement of the r concave packet is generated, then the value of r is added with 1, and the step of removing the 1 third triangulation plane S ^tr _nearest closest to the r third internal feature point P _tr from the internal subdivision decision space, and then adding the fourth triangulation plane composed of the r third internal feature point P _kr and the vertex of the third triangulation plane S ^tr _nearest is executed is performed. And generating an inner profile judgment space after the outer complement of the Kth concave bag, and determining the inner profile judgment space after the outer complement of the Kth concave bag as an inner profile judgment space after the outer complement of the kth new triangulation surface concave bag. Then, the value of t is added with 1, and the values of f and m are respectively initialized to 1, and the step of calculating J distances between the f second internal feature points in the second internal feature point set RIP and J new triangulation planes is carried out. And generating an inner dissection judgment space after the outer dissection of the concave bag of the J-th new triangulation surface.

Finally, judging whether the sum of the side lengths of all the triangulation surfaces in the inner dissection judgment space after the outer interpolation of the concave bag of the J-th new triangulation surface is larger than a judgment threshold, if the sum of the side lengths of all the triangulation surfaces in the inner dissection judgment space after the outer interpolation of the concave bag of the J-th new triangulation surface is larger than the judgment threshold, taking the inner dissection judgment space after the outer interpolation of the concave bag of the J-th new triangulation surface as a new convex bag judgment space again, returning to the step S120, and executing the preset convex-concave conversion treatment on the convex bag judgment space again to obtain a concave bag judgment space; otherwise, if the sum of the side lengths of all the triangulation surfaces in the inner dissection judgment space after the outer interpolation of the concave bag of the J-th new triangulation surface is smaller than or equal to a judgment threshold, determining the inner dissection judgment space after the outer interpolation of the concave bag of the J-th new triangulation surface as a concave bag judgment space.

According to the sea surface target detection method based on concave bag judgment, the purpose of generating a concave bag judgment space by an inner bag judgment space is achieved by determining the outer bag compensation characteristic point from the second inner characteristic point set removed in the process of the inner bag preset operation, removing the third triangulation surface closest to the outer bag compensation characteristic point from the inner bag judgment space, and adding the fourth triangulation surface formed by the outer bag compensation characteristic point and the top point of the third triangulation surface.

Optionally, the determining the first triangulation surface and the second triangulation surface of the convex hull decision space includes:

Determining the sum of the side lengths of all triangulation surfaces in the convex hull judgment space, and calculating a judgment threshold based on the sum of the side lengths; when the sum of the side lengths is larger than the judgment threshold, selecting a first triangulation surface with the sum of the three side lengths as the maximum value from the convex hull judgment space; a target longest edge of the first triangulation surface is determined, and a second triangulation surface comprising the target longest edge is determined.

Specifically, for the convex hull judgment space corresponding to the characteristic points of the historical sea clutter samples, the side length sum of all triangulation planes in the convex hull judgment space can be calculated,Length (Δ _i) is the sum of the side lengths of the triangulation planes in the ith convex hull decision space, i=1, 2, 3, …, M is the total number of triangulation planes in the convex hull decision space; then calculating a decision threshold based on the side length sum length, wherein threshold=mean (length), and mean is an average operation; further, a first triangulation plane delta _max with the sum of three side lengths as the maximum value is selected from the convex hull decision space, the longest side of the first triangulation plane delta _max is determined to be the target longest side ridge, and a second third triangulation plane delta _share containing the target longest side ridge, that is, the first triangulation plane delta _max and the second triangulation plane delta _share share the target longest side ridge, is determined.

According to the sea surface target detection method based on concave hull judgment, the judgment threshold is calculated based on the sum of the side lengths of all the triangular surfaces in the convex hull judgment space, when the sum of the side lengths is larger than the judgment threshold, the first triangular surface with the sum of the three side lengths as the maximum value is selected from the convex hull judgment space, and the second triangular surface with the longest target side is shared by the first triangular surface, so that a powerful basis is provided for carrying out preset convex hull inner section on the convex hull judgment space in the follow-up process, and the algorithm efficiency is effectively improved.

Optionally, the selecting, from the first set of internal feature points, a second set of internal feature points that is removed during the process of executing the preset convex hull inner profile operation includes:

acquiring a plurality of tetrahedrons formed by four vertexes of the first triangulation surface, the second triangulation surface and the target internal feature points; a second set of internal feature points of the first set of internal feature points distributed among the plurality of tetrahedrons is determined.

Specifically, when the terminal device determines that the first triangulation plane Δ _max and the second third triangulation plane Δ _share of the target longest edge bridge are shared and determines, from the first internal feature point set IPS, the target internal feature point cp closest to the midpoint of the target longest edge bridge, if 4 vertices included in the first triangulation plane Δ _max and the second triangulation plane Δ _share of the target longest edge bridge are marked as ①、②、③、④, the 1 st tetrahedron formed by cp and ①、②、③ as vertices and the 2 nd tetrahedron formed by cp and ②、③、④ as vertices may be obtained, and then all internal feature points distributed in the 1 st tetrahedron and the 2 nd tetrahedron in the first internal feature point set IPS are determined as the second internal feature point set RIP in a collective manner.

According to the sea surface target detection method based on concave inclusion judgment, four vertexes of the first triangulation surface and the second triangulation surface sharing the longest edge of a target and a plurality of tetrahedrons formed by target internal feature points determined from the first internal feature point set are obtained, and then the second internal feature point set distributed in the tetrahedrons in the first internal feature point set is further determined, so that a powerful basis is provided for the follow-up preset concave inclusion external compensation operation on an internal profile judgment space, and the constant false alarm purpose is achieved.

Optionally, the implementation process of step S110 includes:

acquiring a characteristic point set of a historical sea clutter sample, and calculating a preset false alarm number according to the number of the characteristic points in the characteristic point set and a preset false alarm rate; determining the gravity centers of the feature point sets, and calculating the Euclidean distance between each feature point in the feature point sets and the gravity centers; based on the Euclidean distance, eliminating L characteristic points meeting a preset distance relation between the characteristic points and the gravity center from the characteristic point set to obtain a target characteristic point set; wherein L is the preset false alarm number; and performing triangulation operation on the target characteristic point set to obtain a convex hull judgment space.

Specifically, the terminal device firstly acquires a characteristic point set S of a historical sea clutter sample, and calculates a preset false alarm number L, l=w·p _F according to W characteristic points in the characteristic point set S and a preset false alarm rate p _F; then, determining the gravity center of the feature point set S, namely determining the average value of W feature points in the feature point set S ₀ as the gravity center of the feature point set S, and calculating the Euclidean distance between each feature point in the feature point set S and the gravity center according to the existing Euclidean distance calculation formula to obtain W Euclidean distances; further, the W Euclidean distances are ordered according to the order from small to large or from large to small, and the L Euclidean distances after the order from small to large are selected or the L Euclidean distances before the order from large to small are selected; then, removing L characteristic points corresponding to the selected L Euclidean distances from the characteristic point set S ₀ to obtain a target characteristic point set comprising W-L characteristic points; finally, performing triangulation operation on the target feature point set based on Convhull functions to obtain a convex hull judgment space, wherein the convex hull judgment space can be expressed as a convex hull surface set S _n,S_n＝{Δ₁,Δ₂,…,Δ_i,…,Δ_m},Δ_i to represent an ith triangulation surface, and each triangulation surface can be a delaunay triangulation surface.

According to the sea surface target detection method based on concave-convex hull judgment, firstly, according to the Euclidean distance between each characteristic point in the characteristic point set of the historical sea clutter sample and the gravity center of the characteristic point set, the L characteristic points meeting the preset distance relation between the L characteristic points and the gravity center are deleted from the characteristic point set based on the Euclidean distance to obtain the target characteristic point set, and further, after the triangulation operation is carried out on the target characteristic point set, the convex hull judgment space corresponding to the characteristic points of the historical sea clutter sample is obtained, so that the aim of eliminating the false alarm points in the characteristic points of the historical sea clutter sample is fulfilled, and the reliability and pertinence of the follow-up execution of convex-concave hull theory operation are improved.

Optionally, the process of step S120 is described in detail with reference to fig. 2, and includes the following sub-steps:

Step S2.1, initializing: the convex hull judgment space corresponding to the characteristic points of the historical sea clutter samples is an initial convex hull judgment space omega ₁; the initial values of t, F and m are 1, f=1, 2, 3, … and F _n,t＝1、2、3、…、J_n,m＝1、2、3、…、K_nt,K_nt≤F_n,F_n are the total number of second internal feature points contained in an nth second internal feature point set RIP _n, J _n is the total number of new triangulation planes formed during the inner section operation of an nth preset convex hull, and K _nt is the total number of third internal feature points contained in a t third internal feature point set FPS _nt determined during the outer section operation of an nth preset concave hull; the initial value of n is 1, and the number of times of preset convex hull inner section operation is the same as the number of times of preset concave hull outer compensation operation.

Step S2.2 determines that the set of the remaining feature points after the feature points at the vertex of each triangulation surface in the nth convex hull decision space Ω _n are removed is the nth first internal feature point set IPS _n.

Step S2.3 calculates the sum of side lengths length _n of all triangulation planes in the nth convex hull decision space Ω _n, and calculates the nth decision threshold _n.

Step S2.4, judging whether the side length sum length _n is larger than a judgment threshold _n, if so, executing step S2.5; otherwise, if length _n≤threshold_n, step S2.22 is performed.

Step S2.5 selecting the nth first triangulation surface with the sum of three side lengths as the maximum value from the nth convex hull decision space omega _n And determining the nth first triangulation plane/>Is the nth target longest edge ridge _n, and determines the nth second third split face/>, which contains the nth target longest edge ridge _n Step S2.6 is then performed.

Step S2.6 determining the nth target internal feature Point cp _n closest to the midpoint of the nth target longest edge ridge _n from the nth first internal feature Point set IPS _n, then determining the first triangulation plane with the nthNth second third split face/>P _n new triangulation surfaces formed by the fourth vertex of the (n) target internal feature point cp _n and P _n new triangulation surfaces do not contain the n-th target longest edge ridge _n, and then removing the n-th first triangulation surface/>, from the n-th convex hull decision space Ω _n And an nth second third split face/>And after P _n new triangulation planes are added, generating an nth inner dissection judgment space/>Step S2.7 is then performed. Preferably, P _n has a value of 4.

Step S2.7 acquiring the nth first triangulation planeNth second third split face/>Two tetrahedrons composed of the nth target internal feature point cp _n and the fourth vertex are used as vertices, all internal feature points distributed in the two tetrahedrons in the nth first internal feature point set IPS _n are determined as an nth second internal feature point set RIP _n in a set manner, and then step S2.8 is performed.

Step S2.8 calculates, for the t-th new triangulation plane of the P _n new triangulation planes, P _n distances between the f-th second internal feature points in the n-th second internal feature point set RIP _n and the P _n new triangulation planes, respectively, and then performs step S2.9.

Step S2.9, judging the size relation between the distance L _fn between the f second internal feature point and the t new triangulation surface and the distance P _n, if the distance L _fn between the f second internal feature point and the t new triangulation surface is smaller than the distance P _n, determining the f second internal feature point as the m external compensation feature point to be concave, and executing step S2.10; otherwise, if the distance between the f second internal feature point and the t new triangulation plane is greater than at least one of the distances P _n, determining that the f second internal feature point does not need the concave-packed external complement, and executing step S2.12.

Step S2.10 judges whether F is greater than F _n, if F > F _n, step S2.14 is executed; if F is less than or equal to F _n, step S2.11 is performed.

Step S2.11 adds 1 to the values of m and f, respectively, and returns to step S2.8.

Step S2.12 judges whether F is greater than F _n, if F > F _n, step S2.14 is executed; if F is less than or equal to F _n, step S2.13 is performed.

Step S2.13 adds 1 to the value of f, and returns to step S2.8.

Step S2.14 indicates that K _nt to-be-concave external complement feature points corresponding to the t th new triangulation surface determined when the n-th preset concave external complement operation is obtained currently when F > F _n, then the K _nt to-be-concave external complement feature points are sorted from near to far or from far to near according to the distance between the K _nt to-be-concave external complement feature points and the t new triangulation surface respectively, the sorted K _nt to-be-concave external complement feature points are determined to be the t third internal feature point set FPS _nt,FPS_nt＝[P_nt(1),P_nt(2),…,P_nt(r),…,P_nt(Knt)],r＝1、2、3、…、K_nt when the n-th preset concave external complement operation, the initial value of r is 1, and then step S2.15 is executed.

Step S2.15 from the nth inner section decision spaceAfter eliminating the 1 third triangulation plane S ^nt(r) _nearest nearest to the r third internal feature point P _nt(r), and adding a fourth triangulation plane consisting of the r third internal feature point P _nt(r) and the vertex of the third triangulation plane S ^nt(r) _nearest, generating an n-th inner-profile judgment space and performing inner-profile judgment space/>, after the r-th concave inclusion outer compensationStep S2.16 is then performed.

Step S2.16 judges whether the value of r is larger than K _nt, if r is smaller than or equal to K _nt, step S2.17 is executed; otherwise, if r > K _nt, step S2.18 is performed.

Step S2.17 adds 1 to the value of r, and returns to step S2.15.

Step S2.18 shows that the inner section judgment space after the outer compensation of the K _nt concave bag is obtained currently when r > K _nt And the inner section judgment space/>, after the outer compensation of the K _nt concave bag, is carried outDetermining an inner dissection judgment space/>, which is obtained after the outer interpolation of the concave bag of the t new triangulation surface and corresponds to the inner dissection operation of the n-th preset convex hullStep S2.19 is then performed.

Step S2.19 judges whether the value of t is larger than J _n, if t is smaller than or equal to J _n, step S2.20 is executed; otherwise, if t > J _n, then step S2.21 is performed.

Step S2.20 adds 1 to the value of t, and initializes the values of f and m to 1, respectively, returning to step S2.8.

Step S2.21 determines that an inner profile judgment space after the outer complement of the concave profile of the J _n th new triangulation surface corresponding to the inner profile operation of the n-th preset convex hull is obtained when t > J _n Step S2.22 is then performed. /(I)

Step S2.22, judging an inner dissection judgment space after the concave bag outer interpolation of the J _n new triangulation surfaceWhether the sum of the side lengths of all triangulation planes in the triangular plane is larger than a judgment threshold or not, if so, the inner-section judgment space/>If the sum length _n of the side lengths of all the triangulation planes is larger than the judgment threshold, executing the step S2.23; conversely, if the inner section decides space/>And if the sum of the side lengths of all the triangulation planes is smaller than or equal to the decision threshold, executing the step S2.24.

Step S2.23, adding 1 to the value of n, and adding the J _n new triangulation surface concave bag to the internal dissection judgment space after external interpolationAnd determining the convex hull decision space omega _n as an nth convex hull decision space, respectively initializing t, f and m to be 1, and returning to the step S2.2.

And S2.24, determining that the corresponding judgment space is a concave judgment space when the sum of the side lengths is smaller than or equal to the judgment threshold.

It should be noted that, the implementation process in steps S2.1 to S2.24 may be referred to in correspondence with the above-described sea surface target detection method based on the concave packet determination, which is not described herein again.

According to the sea surface target detection method based on concave-convex hull judgment, firstly, the longest side is found based on the convex hull judgment space corresponding to the characteristic points of the historical sea clutter samples, the preset convex hull inner section is carried out once from the longest side into the convex hull, meanwhile, in order to ensure that no sample points are removed and the preset concave hull outer compensation is required, if the side length of a certain convex hull surface is larger than the judgment threshold after the preset concave hull outer compensation, the preset convex hull inner section operation and the preset concave hull outer compensation operation are repeatedly executed until the side lengths of all the convex hull surfaces are smaller than the threshold, the aim of effectively improving algorithm efficiency can be achieved through the preset convex hull inner section mode, the aim of preventing the characteristic points of the sea clutter samples from being removed when the preset convex hull inner section is achieved through the preset concave hull outer compensation operation mode is achieved, and therefore the detection precision and the detection accuracy of small sea surface targets are improved.

In the actual processing process, in order to verify the effectiveness and advantages of the method, the algorithm verification is firstly carried out on the traditional convex hull algorithm and the concave hull algorithm provided by the invention, the verification result is shown in fig. 3, (a) is a simulation effect diagram obtained by using the traditional convex hull algorithm, and (b) is a simulation effect diagram obtained by using the concave hull method provided by the invention. As can be seen from fig. 3, the concave-convex method provided by the invention can realize the purpose that the characteristic points of the sea clutter sample are uniformly distributed in the aggregation area, thereby avoiding the defects that the characteristic points of the sea clutter sample are respectively uneven and obvious blank areas appear after a convex-convex algorithm is used, and effectively improving the algorithm efficiency and the algorithm precision.

Further, for the same feature sample set, the invention uses the traditional convex hull judgment space and the concave hull judgment space provided by the invention to detect the small target on the sea surface, the detection result is shown in fig. 4, (a) is a detection result diagram obtained by using the traditional convex hull judgment space, and (b) is a detection result diagram obtained by using the concave hull judgment space provided by the invention. As can be seen from fig. 4, the concave packet judgment space provided by the invention fully utilizes the characteristic points of the sea clutter samples, so that the accuracy and reliability of the subsequent sea small target detection using the concave packet judgment space are higher, and the effect is better.

The sea surface target detection device based on the concave packet judgment provided by the invention is described below, and the sea surface target detection device based on the concave packet judgment described below and the sea surface target detection method based on the concave packet judgment described above can be correspondingly referred to each other.

Fig. 5 illustrates a sea surface target detection apparatus based on a dishing decision, as shown in fig. 5, the sea surface target detection apparatus 500 based on a dishing decision includes: the obtaining module 510 is configured to obtain a convex hull decision space corresponding to a feature point of the historical sea clutter sample; the convex hull judgment space characterizes that the characteristic points are distributed in the interior and each vertex of a mesh topological structure formed by triangulation planes; the processing module 520 is configured to perform a preset convex-to-concave processing on the convex hull decision space to obtain a concave hull decision space; the preset convex-concave conversion method comprises the steps of removing the longest edge in the convex hull judgment space and not removing the characteristic points; the determining module 530 is configured to determine, according to a relationship between the feature point to be detected and the concave packet judgment space, that the feature point to be detected is a small sea surface target.

Optionally, the processing module 520 may be specifically configured to determine a first set of internal feature points of the convex hull decision space; the first internal feature point set comprises feature points distributed in the convex hull judgment space; based on the first internal feature point set, performing preset convex hull inner section operation on the convex hull judgment space to obtain an inner section judgment space; selecting a second internal feature point set which is removed in the process of executing the preset convex hull inner section operation from the first internal feature point set; and based on the second internal characteristic point set, performing preset concave inclusion external compensation operation on the internal profile judgment space to obtain a concave inclusion judgment space.

Optionally, the processing module 520 may be further configured to determine a first triangulation plane and a second triangulation plane of the convex hull decision space; the first triangulation surface and the second triangulation surface share a target longest edge; determining a target internal feature point closest to the midpoint of the longest target edge in the first internal feature point set; determining J new triangulation planes formed by four vertexes of the first triangulation plane, the second triangulation plane and the target internal feature points; wherein none of the J new triangulation planes contains the target longest edge; and removing the first triangulation surface and the second triangulation surface from the convex hull judgment space, and generating an inner dissection judgment space after adding the J new triangulation surfaces.

Optionally, the processing module 520 may be further specifically configured to determine J distances between the second internal feature points in the second internal feature point set and the J new triangulation planes, for the new triangulation planes of the J new triangulation planes, respectively; when the distances between the second internal feature point and the new triangulation surface are smaller than the J distances, determining the second internal feature point as an external compensation feature point to be concave; and removing a third triangulation surface closest to the external complement feature point to be concave from the internal subdivision judgment space, and generating a concave subdivision judgment space after adding a fourth triangulation surface consisting of the external complement feature point to be concave and the vertex of the third triangulation surface.

Optionally, the processing module 520 may be further specifically configured to determine a side length sum of all triangulation planes in the convex hull decision space, and calculate a decision threshold based on the side length sum; when the sum of the side lengths is larger than the judgment threshold, selecting a first triangulation surface with the sum of the three side lengths as the maximum value from the convex hull judgment space; a target longest edge of the first triangulation surface is determined, and a second triangulation surface comprising the target longest edge is determined.

Optionally, the processing module 520 may be further specifically configured to obtain a plurality of tetrahedrons formed by four vertices of the first triangulation surface, the second triangulation surface and the target internal feature point; a second set of internal feature points of the first set of internal feature points distributed among the plurality of tetrahedrons is determined.

Optionally, the processing module 510 may be specifically configured to obtain a feature point set of the historical sea clutter sample, and calculate a preset false alarm number according to the number of feature points in the feature point set and a preset false alarm rate; determining the gravity centers of the feature point sets, and calculating the Euclidean distance between each feature point in the feature point sets and the gravity centers; based on the Euclidean distance, eliminating L characteristic points meeting a preset distance relation between the characteristic points and the gravity center from the characteristic point set to obtain a target characteristic point set; wherein L is the preset false alarm number; and performing triangulation operation on the target characteristic point set to obtain a convex hull judgment space.

Fig. 6 illustrates a physical schematic diagram of an electronic device, as shown in fig. 6, which may include: processor 610, communication interface (Communications Interface) 620, memory 630, and communication bus 640, wherein processor 610, communication interface 620, memory 630 communicate with each other via communication bus 640. The processor 610 may invoke logic instructions in the memory 630 to perform a sea surface target detection method based on the dishing decisions, the method comprising: obtaining a convex hull judgment space corresponding to the characteristic points of the historical sea clutter samples; the convex hull judgment space characterizes that the characteristic points are distributed in the interior and each vertex of a mesh topological structure formed by triangulation planes; performing preset convex-concave conversion treatment on the convex hull judgment space to obtain a concave hull judgment space; the preset convex-concave conversion method comprises the steps of removing the longest edge in the convex hull judgment space and not removing the characteristic points; and determining the feature points to be detected as sea surface small targets according to the relation between the feature points to be detected and the concave packet judgment space.

Further, the logic instructions in the memory 630 may be implemented in the form of software functional units and stored in a computer-readable storage medium when sold or used as a stand-alone product. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a usb disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.

In another aspect, the present invention also provides a computer program product, the computer program product comprising a computer program, the computer program being storable on a non-transitory computer readable storage medium, the computer program, when executed by a processor, being capable of performing the method for detecting a sea surface target based on a concave packet decision provided by the above methods, the method comprising: obtaining a convex hull judgment space corresponding to the characteristic points of the historical sea clutter samples; the convex hull judgment space characterizes that the characteristic points are distributed in the interior and each vertex of a mesh topological structure formed by triangulation planes; performing preset convex-concave conversion treatment on the convex hull judgment space to obtain a concave hull judgment space; the preset convex-concave conversion method comprises the steps of removing the longest edge in the convex hull judgment space and not removing the characteristic points; and determining the feature points to be detected as sea surface small targets according to the relation between the feature points to be detected and the concave packet judgment space.

In yet another aspect, the present invention further provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, is implemented to perform the method for detecting a sea surface target based on a concave packet decision provided by the above methods, the method comprising: obtaining a convex hull judgment space corresponding to the characteristic points of the historical sea clutter samples; the convex hull judgment space characterizes that the characteristic points are distributed in the interior and each vertex of a mesh topological structure formed by triangulation planes; performing preset convex-concave conversion treatment on the convex hull judgment space to obtain a concave hull judgment space; the preset convex-concave conversion method comprises the steps of removing the longest edge in the convex hull judgment space and not removing the characteristic points; and determining the feature points to be detected as sea surface small targets according to the relation between the feature points to be detected and the concave packet judgment space.

The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.

From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on this understanding, the foregoing technical solution may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective embodiments or some parts of the embodiments.

Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. The sea surface target detection method based on the concave packet judgment is characterized by comprising the following steps of:

determining a target internal feature point closest to the midpoint of the longest target edge in the first internal feature point set; determining J new triangulation planes formed by four vertexes of the first triangulation plane, the second triangulation plane and the target internal feature points; wherein none of the J new triangulation planes contains the target longest edge;

Removing the first triangulation surface and the second triangulation surface from the convex hull judgment space, and generating an inner dissection judgment space after adding the J new triangulation surfaces;

selecting a second internal characteristic point set which is removed in the process of executing the preset convex hull inner section operation from the first internal characteristic point set;

Removing a third triangulation surface closest to the external complement feature point to be concave from the internal subdivision judgment space, and generating a concave judgment space after adding a fourth triangulation surface consisting of the external complement feature point to be concave and the vertex of the third triangulation surface;

2. The method of hull decision based sea surface target detection according to claim 1, wherein said determining a first triangulation surface and a second triangulation surface of said hull decision space comprises:

3. The method for detecting a sea surface target based on concave hull decision according to claim 1, wherein selecting a second set of internal feature points that are removed in the process of performing the preset convex hull inner hull operation from the first set of internal feature points includes:

4. A method for detecting a sea surface target based on concave hull decision according to any one of claims 1 to 3, wherein the convex hull decision space corresponding to the feature point of the acquired historical sea clutter sample includes:

5. Sea surface target detection device based on concave packet judgement, characterized by comprising:

The acquisition module is used for acquiring a convex hull judgment space corresponding to the characteristic points of the historical sea clutter samples; the convex hull judgment space characterizes that the characteristic points are distributed in the interior and each vertex of a mesh topological structure formed by triangulation planes;

The processing module is used for determining a first internal characteristic point set of the convex hull judgment space; the first internal feature point set comprises feature points distributed in the convex hull judgment space; determining a first triangulation surface and a second triangulation surface of the convex hull decision space; the first triangulation surface and the second triangulation surface share a target longest edge; determining a target internal feature point closest to the midpoint of the longest target edge in the first internal feature point set; determining J new triangulation planes formed by four vertexes of the first triangulation plane, the second triangulation plane and the target internal feature points; wherein none of the J new triangulation planes contains the target longest edge; removing the first triangulation surface and the second triangulation surface from the convex hull judgment space, and generating an inner dissection judgment space after adding the J new triangulation surfaces; selecting a second internal characteristic point set which is removed in the process of executing the preset convex hull inner section operation from the first internal characteristic point set; determining J distances between second internal feature points in the second internal feature point set and the J new triangulation planes respectively for the new triangulation planes in the J new triangulation planes; when the distances between the second internal feature point and the new triangulation surface are smaller than the J distances, determining the second internal feature point as an external compensation feature point to be concave; removing a third triangulation surface closest to the external complement feature point to be concave from the internal subdivision judgment space, and generating a concave judgment space after adding a fourth triangulation surface consisting of the external complement feature point to be concave and the vertex of the third triangulation surface;

and the determining module is used for determining the feature points to be detected as sea surface small targets according to the relation between the feature points to be detected and the concave packet judgment space.

6. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor performs the steps of the method for sea surface target detection based on a dishing decision according to any one of claims 1 to 4 when the program is executed.

7. A non-transitory computer readable storage medium having stored thereon a computer program, which when executed by a processor, implements the steps of the method for sea surface target detection based on dishing decisions according to any of claims 1 to 4.

8. A computer program product comprising a computer program which, when executed by a processor, implements the steps of the method for sea surface target detection based on dishing decisions according to any one of claims 1 to 4.