CN115828108B - Shape matching-based moving object track similar segment extraction method - Google Patents
Shape matching-based moving object track similar segment extraction method Download PDFInfo
- Publication number
- CN115828108B CN115828108B CN202211432189.2A CN202211432189A CN115828108B CN 115828108 B CN115828108 B CN 115828108B CN 202211432189 A CN202211432189 A CN 202211432189A CN 115828108 B CN115828108 B CN 115828108B
- Authority
- CN
- China
- Prior art keywords
- track
- sub
- seg
- vertex
- particle
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000605 extraction Methods 0.000 title claims abstract description 14
- 239000002245 particle Substances 0.000 claims abstract description 209
- 238000000034 method Methods 0.000 claims abstract description 70
- 238000005457 optimization Methods 0.000 claims abstract description 25
- 238000005070 sampling Methods 0.000 claims description 58
- 230000006870 function Effects 0.000 claims description 32
- 101100517648 Saccharomyces cerevisiae (strain ATCC 204508 / S288c) NUM1 gene Proteins 0.000 claims description 18
- 101100129590 Schizosaccharomyces pombe (strain 972 / ATCC 24843) mcp5 gene Proteins 0.000 claims description 18
- 238000010606 normalization Methods 0.000 claims description 18
- 239000013598 vector Substances 0.000 claims description 17
- 230000009466 transformation Effects 0.000 claims description 13
- 238000013519 translation Methods 0.000 claims description 13
- 238000012545 processing Methods 0.000 claims description 11
- 238000011282 treatment Methods 0.000 claims description 8
- 238000000354 decomposition reaction Methods 0.000 claims description 7
- 239000011159 matrix material Substances 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 7
- 239000012634 fragment Substances 0.000 abstract description 12
- 230000008901 benefit Effects 0.000 abstract description 5
- 238000010586 diagram Methods 0.000 description 7
- 238000004458 analytical method Methods 0.000 description 6
- 238000012952 Resampling Methods 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 5
- 230000006399 behavior Effects 0.000 description 3
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005065 mining Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
Landscapes
- Image Analysis (AREA)
Abstract
The invention provides a method for extracting a similar segment of a moving target track based on shape matching, which comprises the following steps: acquiring a first track and a second track to be subjected to shape matching, acquiring a corresponding first track segment set and a second track segment set, and defining a coding mode of particles in a particle swarm; performing shape description on the first sub-track and the second sub-track corresponding to each particle based on the shape descriptor of the symbol centroid distance to respectively obtain corresponding shape feature sequences; constructing an objective function based on the shape feature sequence, and calculating an objective function value of each particle; based on a distributed neighbor search strategy, iterative optimization is carried out on the particle swarm, and a first sub-track and a second sub-track corresponding to the particle with the largest objective function value in the particle swarm are output as similar fragments. According to the scheme of the invention, the efficient extraction of the similar shape fragments among the tracks is realized, and compared with the existing method, the method has advantages in both track shape description and similar fragment retrieval.
Description
Technical Field
The invention relates to the field of data processing, in particular to a method for extracting a similar segment of a moving target track based on shape matching.
Background
With the continuous improvement of the technical level of sensing positioning and mobile communication, the source of the data of the position of the mobile target is also becoming wider. A series of position data sets with time stamps of a moving target in a certain time period can be described as a target motion track, and a track similarity analysis method is used for mining a behavior mode and a movement rule of target commonalities from the track set, so that the method has important value for understanding different field problems such as human activities, traffic conditions, animal migration, ecological environment and the like. Behavior patterns mined from a target track can be generally classified into frequent patterns, companion patterns, aggregate patterns, and exception patterns. Shape-based templates are a special target motion pattern, which in a broad sense belongs to a frequent pattern, i.e. a special shape frequently appears in the trajectory of the target.
Trajectory similarity analysis is the basis for target motion pattern mining, and similarity includes a variety of connotations: the temporal similarity of the targets may be analyzed based on the time or order in which the targets appear; the spatial similarity of the target can be analyzed based on the location area of the target motion; the similarity of the motion parameters of the target can be analyzed based on the characteristics and the changes of the speed, the acceleration, the motion direction and the like of the target; the similarity analysis results of the time, space and motion parameters are integrated to analyze the similarity of the behavior patterns of the targets, so that regular target activities such as hot spot areas, frequent paths, aggregation conditions, accompanying features and the like can be found. Although the trajectory similarity analysis has a difference between the target object and the result requirement, the distance calculation is still an essential link of most of the trajectory similarity analysis tasks. The existing distance calculation methods are many and can be divided into a plurality of types based on curves, real distances, editing distances, time perception, segmentation and the like, and each type of calculation model has the characteristics of each type of calculation model, but most of the distance calculation methods are designed based on space-time similarity among tracks. However, there may be no spatiotemporal correlation of motion trajectories or segments of similar shape, and the similar shape trajectories or segments have transformation characteristics such as rotation, translation, and scaling, so the existing methods cannot be directly used for trajectory shape similarity metrics and trajectory similar shape segment extraction.
Disclosure of Invention
In order to solve the technical problems, the invention provides a method for extracting similar track fragments of a moving target based on shape matching, which is used for solving the technical problems that track shape similarity is difficult to measure in the existing track similarity analysis technology, and then similar track fragments with non-space-time correlation or geometric transformation are difficult to extract effectively.
According to a first aspect of the present invention, there is provided a method for extracting a moving object trajectory similarity fragment based on shape matching, the method comprising the steps of:
step S1: acquiring a first track TR to be subjected to shape matching A And a second track TR B Obtain and respectively and TR A 、TR B Corresponding first track segment set TotalSeg A A second track segment set TotalSeg B ，TotalSeg A ＝{Seg Anum1 }，TotalSeg B ＝{Seg Bnum2 }, wherein Seg Anum1 For the first track segment set TotalSeg A Is not less than 1 NUM1 and not more than NUM1, NUM1 being the first track segment set TotalSeg A Number of middle track segments, seg Bnum2 For the second track segment set TotalSeg B Is not less than 1 and not more than 2 and is not more than 2, wherein NUM2 is a second track segment set TotalSeg B The number of mid-track segments; based on the first track segment set TotalSeg A The method comprises the steps of determining an original track length of a first track and a length of a first sub-track formed by a plurality of continuous track segments in the first track segment set; based on the second track segment set TotalSeg B The original track length of the second track and the length of a second sub-track formed by a plurality of continuous track segments in the second track segment set can be determined;
defining the coding mode of the particles in the particle swarm, so that the particle Pos comprises the starting track segment number id of the first sub-track AS End track segment number id AE Starting track segment number id of the second sub-track BS End track segment number id BE I.e. { id AS ,id AE ,id BS ,id BE }；
Initializing particlesParameters of a particle swarm algorithm comprise a preset number NP of particles, a maximum iterative optimization frequency N1, and a proportion limiting condition SLP= { delta of the first sub-track and the second sub-track corresponding to the particles respectively occupying the length of the corresponding original track, wherein the parameters of the particle swarm algorithm comprise the preset number NP of the particles, the maximum iterative optimization frequency N1 and the proportion limiting condition SLP= { delta of the first sub-track and the second sub-track corresponding to the particles respectively occupying the length of the corresponding original track A ,δ B }, satisfy0＜δ A ,δ B Not more than 1, wherein LSeg A 、LSeg B LTR is the length of the first sub-track and the second sub-track corresponding to the particle A 、LTR B The original track length of the first track and the original track length of the second track are respectively;
step S2: first sub-trajectories Seg corresponding to respective particles A Second sub-track Seg B Performing shape description based on the shape descriptor of the symbol centroid distance to obtain a shape feature sequence SDSeg respectively A And SDSeg B The method comprises the steps of carrying out a first treatment on the surface of the Based on shape feature sequence SDSeg A And SDSeg B Constructing an objective function for measuring a first sub-track Seg corresponding to the particle A And a second sub-track Seg B Similarity of (2); calculating objective function values of the particles;
step S3: based on a distributed neighbor search strategy, carrying out iterative optimization on a particle swarm until the iterative optimization times reach N1, and carrying out first sub-track Seg corresponding to the particle with the largest objective function value in the particle swarm A And a second sub-track Seg B As a similar segment output. Preferably, the step S2 includes:
step S21: for the first sub-track Seg A Second sub-track Seg B Selecting m points respectively, sampling m points equidistantly, and using the first sub-track Seg A M points of said second sub-track Seg B Respectively form a first sub-track Segs after sampling A And a second sub-track Segs after sampling B Wherein the first sub-track Seg A M points of (1) include Seg A The first point of the start track section and the end point of the end track section, the secondSub-track Seg B M points of (1) include Seg B A first segment point of the start track segment and a second segment point of the end track segment; for Segs A 、Segs B M-sided Polygon composed of sampling points A 、Polygon B Respectively performing polygon decomposition to respectively obtain m-2 triangles taking the segment head point of the starting point track segment as the vertex; calculating m-sided Polygon A 、Polygon B Centroid coordinates Conpolygon A Conpolygon B ；
Step S22: respectively for the first sub-track Segs after sampling A And a second sub-track Segs after sampling B Performing normalization processing, wherein the normalization processing comprises translation transformation of the mass center of the m-sided polygon corresponding to the sub-track and area normalization of the m-sided polygon; obtaining a normalized first sub-track Segsn A And a normalized second sub-track Segsn B ；
Step S23: acquiring a normalized first sub-track Segsn A Distance of each vertex of (2) to the corresponding centroid, segsn A The vertices of the sequence are ordered in a clockwise direction to obtain Segsn A A deflection angle of a vector formed by each vertex of the plurality of vertices and a point next to the vertex with respect to a vector formed by the origin of coordinates and the vertex; obtaining a normalized second sub-track Segsn B Distance of each vertex of (2) to the corresponding centroid, segsn B The vertices of the sequence are ordered in a clockwise direction to obtain Segsn B A deflection angle of a vector formed by each vertex of the plurality of vertices and a point next to the vertex with respect to a vector formed by the origin of coordinates and the vertex;
define centroid distance of vertex node as bd node Deflection angle of theta node Will deflect by an angle theta node Converted into positive sign or negative sign and then superimposed to the centroid distance to form the sign centroid distance shape descriptor sbd of the vertex node node I.e. the angle of rotation of the vertex satisfies θ node E [0, pi), the symbol centroid of vertex node is separated from shape descriptor sbd i ＝-bd node The rotation angle satisfies theta node E [ pi, 2 pi), the symbol centroid distance shape of vertex node is tracedThe symbol sbd node ＝bd node ；
Acquiring a normalized first sub-track Segsn A The symbol centroid distance shape descriptors of each vertex of the plurality of vertices are arranged to obtain a shape feature sequence SDSeg by arranging the symbol centroid distance shape descriptions Fu Anxu corresponding to each vertex A ；
Obtaining a normalized second sub-track Segsn B The symbol centroid distance shape descriptors of each vertex of the plurality of vertices are arranged to obtain a shape feature sequence SDSeg by arranging the symbol centroid distance shape descriptions Fu Anxu corresponding to each vertex B 。
Preferably, the sampling method of the m points in the step S2 includes:
step S211: calculating the track length of a sub-track, wherein the sub-track is a first sub-track or a second sub-track; the track length of the sub track isDividing L into m-1 parts, wherein the length lm of each part is equal-distance sampling step length, n is the number of original track points contained from the first point to the tail point of the sub track, the number of the first point is 1, the number of the tail point is n, and p j+1 The j+1st point from the first point has coordinate information; p is p j A j-th point from the first point, having coordinate information; the sum of the distances between two adjacent points is the length L of the sub-track;
step S212: setting the total count total of sampling points to be 0 by taking the first point of the starting track segment of the sub track as a starting point;
step S213: moving the equidistant sampling step lm each time along the direction of the sub-track from the starting point until the requirement is metAnd->Wherein k is the number of original sub-track points including the first point which moves along the direction of the sub-track according to the step lm; determiningDetermining coordinates of a new sampling point:
wherein xq is the abscissa of the new sampling point, yq is the ordinate of the new sampling point, and p k P is the last original track point of the new sampling point in the original sub-track k+1 For the next original track point of the new sampling point in the original sub-track, the new sampling point is located at p k And p k+1 Between xp k Is p k X p k+1 Is p k+1 Is the abscissa of yp k Is p k Yp, ordinate of (c) k+1 Is p k+1 Is the ordinate of (2); setting the total count total of sampling points as total+1;
step S214: if total is equal to m-2, ending the method; otherwise, the new sampling point is taken as a starting point, and the process proceeds to step S213.
Preferably, the distributed neighbor search strategy refers to constructing a neighbor matrix M PD Matrix element M PD (ii, jj) is a particle Pos ii ,Pos jj The distance between the two particles, ii, jj are respectively the particle numbers, 1 is less than or equal to ii, jj is less than or equal to NP, an update mark is set for each particle, and an initial value is 0, so that the particle does not evolve update currently; for each particle, the following operations are performed in numbered order: and acquiring a particle number, searching particles which are not evolutionarily updated around the particles corresponding to the number and are nearest to the particles corresponding to the number as nearest neighbor particles, forming a particle pair by the particles corresponding to the number and the nearest neighbor particles, completing evolutionary update of the particle pair, and assigning two particle update marks of the particle pair as 1 after the evolutionary update.
Preferably, the basis of the evolution update is to determine an evolution update mode of a particle which is not subjected to the evolution update in the particle pair based on the history optimum of each particle in the particle pair and the current optimum of two particles in the particle pair; the basis for judging the optimal history and the optimal current is the function value corresponding to the objective function.
According to a second aspect of the present invention, there is provided a moving object trajectory similarity piece extraction device based on shape matching, the device comprising:
An initialization module: configured to obtain a first trajectory TR to be shape matched A And a second track TR B Obtain and respectively and TR A 、TR B Corresponding first track segment set TotalSeg A Second track set TotalSeg B ，TotalSeg A ＝{Seg Anum1 }，TotalSeg B ＝{Seg Bnum2 }, wherein Seg Anum1 For the first track segment set TotalSeg A Is not less than 1 NUM1 and not more than NUM1, NUM1 being the first track segment set TotalSeg A Number of middle track segments, seg Bnum2 For track segment set TotalSeg B Is not less than 1 and not more than 2 and is not more than 2, wherein NUM2 is a second track segment set TotalSeg B The number of mid-track segments; based on the first track segment set TotalSeg A The method comprises the steps of determining an original track length of a first track and a length of a first sub-track formed by a plurality of continuous track segments in the first track segment set; based on the second track segment set TotalSeg B The original track length of the second track and the length of a second sub-track formed by a plurality of continuous track segments in the second track segment set can be determined;
defining the coding mode of the particles in the particle swarm, so that the particle Pos comprises the starting track segment number id of the first sub-track AS End track segment number id AE Starting track segment number id of the second sub-track BS End track segment number id BE I.e. { id AS ,id AE ,id BS ,id BE }；
Initializing parameters of a particle swarm algorithm and randomly initializing NP particles, wherein the parameters of the particle swarm algorithm comprise a preset number NP of particles, a maximum iteration optimizing frequency N1, and a proportion limiting condition SLP= { delta of a first sub-track and a second sub-track corresponding to the particles respectively occupying the length of the corresponding original track A ,δ B }, satisfy0＜δ A ，δ B Not more than 1, wherein LSeg A 、LSeg B LTR is the length of the first sub-track and the second sub-track corresponding to the particle A 、LTR B The original track length of the first track and the original track length of the second track are respectively;
the shape description module: a first sub-track Seg configured to correspond to each particle A Second sub-track Seg B Performing shape description based on the shape descriptor of the symbol centroid distance to obtain a shape feature sequence SDSeg respectively A And SDSeg B The method comprises the steps of carrying out a first treatment on the surface of the Based on shape feature sequence SDSeg A And SDSeg B Constructing an objective function for measuring a first sub-track Seg corresponding to the particle A And a second sub-track Seg B Similarity of (2); calculating objective function values of the particles;
and an optimization module: the method comprises the steps of performing iterative optimization on a particle swarm based on a distributed neighbor search strategy until the iterative optimization times reach N1, and performing first sub-track Seg corresponding to the particle with the largest objective function value in the particle swarm A And a second sub-track Seg B As a similar segment output.
According to a third aspect of the present invention, there is provided a moving object trajectory similarity segment extraction system based on shape matching, comprising:
a processor for executing a plurality of instructions;
a memory for storing a plurality of instructions;
wherein the plurality of instructions are for storage by the memory and loading and executing by the processor the method as described above.
According to a fourth aspect of the present invention, there is provided a computer-readable storage medium having stored therein a plurality of instructions; the plurality of instructions are for loading and executing by the processor the method as described above.
According to the scheme of the invention, the problem of extracting the target similar fragments is converted into the problem of matching the local shape, and the method comprises two key links: firstly, an excellent shape descriptor is established, and the shape descriptor is required to have the characteristics of integrity and accuracy of shape expression, invariance of rotation and translation scaling, compactness and the like; secondly, a high-efficiency matching method is designed, and the method has good similar-shape track segment searching efficiency. Aiming at the first link, the invention provides a novel shape descriptor based on the symbol centroid distance, and the track data is accurately converted into a one-dimensional shape feature sequence by carrying out equidistant resampling, polygonal decomposition, normalization, unique description and other treatments on the track, so that the invention has the characteristics of low complexity, unchanged geometric transformation, noise resistance and the like, and the efficiency of matching and searching of subsequent similar shape fragments is greatly improved. Aiming at the second link, the invention designs a similar shape track fragment retrieval scheme based on particle swarm optimization, defines a track shape similarity evaluation function, and better solves the problem that a standard particle swarm algorithm is easy to fall into local optimization in multi-dimensional and multi-peak value optimization by applying a distributed neighbor search strategy, thereby improving the accuracy of global optimal similar fragment retrieval. The method realizes the efficient extraction of the similar shape fragments among the tracks, and has advantages in both track shape description and similar fragment retrieval compared with the existing method.
The foregoing description is only an overview of the present invention, and is intended to provide a better understanding of the present invention, as it is embodied in the following description, with reference to the preferred embodiments of the present invention and the accompanying drawings.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention, illustrate the invention and together with the description serve to explain the invention. In the drawings:
FIG. 1 is a schematic diagram of an overall frame of a method for extracting similar segments of a moving target track based on shape matching;
FIG. 2 is a schematic flow chart of a method for extracting similar segments of a moving target track based on shape matching;
FIG. 3 is a schematic diagram of equidistant resampling and polygon decomposition in a symbol centroid distance shape descriptor in accordance with the present invention;
FIG. 4 is a diagram illustrating a unique description of a shape descriptor of a symbol centroid distance according to the present invention;
FIG. 5 is a schematic diagram of a particle swarm distributed neighbor search strategy according to the present invention;
FIG. 6 is a diagram showing an example of extraction of similar segments of a moving object trajectory based on shape matching according to an embodiment of the present invention;
FIG. 7 is a graph comparing search results of similar shape track segments based on a distributed neighbor search particle swarm method and a standard particle swarm method according to an embodiment of the present invention;
Fig. 8 is a schematic structural diagram of a moving object track similar segment extraction device based on shape matching.
Detailed Description
In order to make 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 specific embodiments of the present invention and corresponding drawings. It will be apparent that the described embodiments are only some, but not all, embodiments of the invention. 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.
First, a method for extracting a similar segment of a moving object track based on shape matching according to an embodiment of the present invention will be described with reference to fig. 1 to 2. As shown in fig. 1-2, the method comprises the steps of:
step S1: acquiring a first track TR to be subjected to shape matching A And a second track TR B Obtain and respectively and TR A 、TR B Corresponding first track segment set TotalSeg A Second track set TotalSeg B ，TotalSeg A ＝{Seg Anum1 }，TotalSeg B ＝{Seg Bnum2 }, wherein Seg Anum1 For the first track segment set TotalSeg A Is not less than 1 NUM1 and not more than NUM1, NUM1 being the first track segment set TotalSeg A Number of middle track segments, seg Bnum2 For a set of track segmentsSynthesis of TotalSeg B Is not less than 1 and not more than 2 and is not more than 2, wherein NUM2 is a second track segment set TotalSeg B The number of mid-track segments; based on the first track segment set TotalSeg A The method comprises the steps of determining an original track length of a first track and a length of a first sub-track formed by a plurality of continuous track segments in the first track segment set; based on the second track segment set TotalSeg B The original track length of the second track and the length of a second sub-track formed by a plurality of continuous track segments in the second track segment set can be determined;
defining the coding mode of the particles in the particle swarm, so that the particle Pos comprises the starting track segment number id of the first sub-track AS End track segment number id AE Starting track segment number id of the second sub-track BS End track segment number id BE I.e. { id AS ,id AE ,id BS ,id BE }；
Initializing parameters of a particle swarm algorithm and randomly initializing NP particles, wherein the parameters of the particle swarm algorithm comprise a preset number NP of particles, a maximum iteration optimizing frequency N1, and a proportion limiting condition SLP= { delta of a first sub-track and a second sub-track corresponding to the particles respectively occupying the length of the corresponding original track A ,δ B }, satisfy0＜δ A ,δ B Not more than 1, wherein LSeg A 、LSeg B LTR is the length of the first sub-track and the second sub-track corresponding to the particle A 、LTR B The original track length of the first track and the original track length of the second track are respectively;
step S2: first sub-trajectories Seg corresponding to respective particles A Second sub-track Seg B Performing shape description based on the shape descriptor of the symbol centroid distance to obtain a shape feature sequence SDSeg respectively A And SDSeg B The method comprises the steps of carrying out a first treatment on the surface of the Based on shape feature sequence SDSeg A And SDSeg B Constructing an objective function for measuring a first sub-track corresponding to the particleTrace Seg A And a second sub-track Seg B Similarity of (2); calculating objective function values of the particles;
step S3: based on a distributed neighbor search strategy, carrying out iterative optimization on a particle swarm until the iterative optimization times reach N1, and carrying out first sub-track Seg corresponding to the particle with the largest objective function value in the particle swarm A And a second sub-track Seg B As a similar segment output.
In this embodiment, the first track TR A And a second track TR B The two driving tracks of the same object or different objects can be adopted. The track segment set comprises a first track segment set and a second track segment set, each track segment is a straight line segment formed by two adjacent track points in a track, and each track point has time information and position information (such as position coordinates). All track segments in each track segment set are spliced in sequence, and corresponding tracks can be corresponding.
Step S2 of the present invention is further described with reference to fig. 3-4, including:
step S21: for the first sub-track Seg A Second sub-track Seg B Selecting m points respectively, sampling m points equidistantly, and using the first sub-track Seg A M points of said second sub-track Seg B Respectively form a first sub-track Segs after sampling A And a second sub-track Segs after sampling B Wherein the first sub-track Seg A M points of (1) include Seg A The first sub-track Seg is a first sub-track segment, and the second sub-track Seg is a second sub-track segment B M points of (1) include Seg B A first segment point of the start track segment and a second segment point of the end track segment; for Segs A 、Segs B M-sided Polygon composed of sampling points A 、Polygon B Respectively performing polygon decomposition to respectively obtain m-2 triangles taking the segment head point of the starting point track segment as the vertex; calculating m-sided Polygon A 、Polygon B Centroid coordinates Conpolygon A Conpolygon B 。
As shown in fig. 3, in order to ensure that the resampled track fits the original track as much as possible, the first and last points of the resampled track coincide with the first and last points of the original track, while the other track points still need to be on the original track.
Further, the sampling method of m points includes:
step S211: calculating the track length of a sub-track, wherein the sub-track is a first sub-track or a second sub-track; the track length of the sub track is Dividing L into m-1 parts, wherein the length lm of each part is equal-distance sampling step length, n is the number of original track points contained from the first point to the tail point of the sub track, the number of the first point is 1, the number of the tail point is n, and p j+1 The j+1st point from the first point has coordinate information; p is p j A j-th point from the first point, having coordinate information; the sum of the distances between two adjacent points is the length L of the sub-track;
step S212: setting the total count total of sampling points to be 0 by taking the first point of the starting track segment of the sub track as a starting point;
step S213: moving the equidistant sampling step lm each time along the direction of the sub-track from the starting point until the requirement is metAnd->Wherein k is the number of original sub-track points including the first point which moves along the direction of the sub-track according to the step lm; determining coordinates of a new sampling point:
wherein xq is the abscissa of the new sampling point, yq is the ordinate of the new sampling point, and p k P is the last original track point of the new sampling point in the original sub-track k+1 To be the instituteThe new sampling point is positioned at p k And p k+1 Between xp k Is p k X p k+1 Is p k+1 Is the abscissa of yp k Is p k Yp, ordinate of (c) k+1 Is p k+1 Is the ordinate of (2); setting the total count total of sampling points as total+1;
step S214: if total is equal to m-2, ending the method; otherwise, the new sampling point is taken as a starting point, and the process proceeds to step S213.
The m-sided Polygon is Polygon A 、Polygon B Respectively decomposing into m-2 triangles, wherein for any m-polygon, the m-2 triangles decomposed by the m-2 triangles have a common vertex q 1 . Based on this, the polygon area and centroid coordinates are calculated. Solving the area of each triangle by using a sea-land formula, and further solving the total area of the polygonsWherein u is l Is half perimeter of the first triangle, a l 、b l 、c l Is of side length. Then, centroid q of polygon PG bc Coordinates are
Step S22: respectively for the first sub-track Segs after sampling A And a second sub-track Segs after sampling B Performing normalization processing, wherein the normalization processing comprises translation transformation of the mass center of the m-sided polygon corresponding to the sub-track and area normalization of the m-sided polygon; obtaining a normalized first sub-track Segsn A And a normalized second sub-track Segsn B 。
In this embodiment, for m-sided Polygon A 、Polygon B Translation transformation of centroid of (c) and respectively for m-sided Polygon A 、Polygon B Is a function of the area normalization of the (c).
The translation transformation of the centroid of the m-sided polygon refers to transforming the m-sided deformation corresponding to the sub-track into a two-dimensional rectangular coordinate system taking the centroid of the m-sided polygon as an origin; the area normalization of m polygons is to scale the coordinates of m vertices so that trajectories of different sizes and the same shape can be coincident in the same coordinate system after processing.
In this embodiment, for Segs A And Segs B Normalization is carried out, including translation transformation based on polygon centroid and normalization based on polygon area, translation invariance and scaling invariance of track segment shape description can be respectively realized, and normalized track segment Segsn is obtained A And Segsn B . Translation transformation based on polygon centroid is to transform the track to centroid q bc In a new two-dimensional rectangular coordinate system as the origin, then the locus point q i Transformed point q' i Is the coordinates of (a)Then normalizing based on polygonal area, and obtaining the track point q' i Normalized point q i Is +.>The above procedure can be seen as a translation and a scaling of the trajectory, which obviously does not change the shape of the trajectory.
Step S23: acquiring a normalized first sub-track Segsn A Distance of each vertex of (2) to the corresponding centroid, segsn A The vertices of the sequence are ordered in a clockwise direction to obtain Segsn A A deflection angle of a vector formed by each vertex of the plurality of vertices and a point next to the vertex with respect to a vector formed by the origin of coordinates and the vertex; obtaining a normalized second sub-track Segsn B Distance of each vertex of (2) to the corresponding centroid, segsn B The vertices of the sequence are ordered in a clockwise direction to obtain Segsn B A deflection angle of a vector formed by each vertex of the plurality of vertices and a point next to the vertex with respect to a vector formed by the origin of coordinates and the vertex;
define centroid distance of vertex node as bd node Deflection angle of theta node Will deflect by an angle theta node Converted into positive sign or negative sign and then superimposed to the centroid distance to form the sign centroid distance shape descriptor sbd of the vertex node node I.e. the angle of rotation of the vertex satisfies θ node E [0, pi), the symbol centroid of vertex node is separated from shape descriptor sbd i ＝-bd node The rotation angle satisfies theta node E [ pi, 2 pi), the symbol centroid of vertex node is from shape descriptor sbd node ＝bd node ；
Acquiring a normalized first sub-track Segsn A The symbol centroid distance shape descriptors of each vertex of the plurality of vertices are arranged to obtain a shape feature sequence SDSeg by arranging the symbol centroid distance shape descriptions Fu Anxu corresponding to each vertex A ；
Obtaining a normalized second sub-track Segsn B The symbol centroid distance shape descriptors of each vertex of the plurality of vertices are arranged to obtain a shape feature sequence SDSeg by arranging the symbol centroid distance shape descriptions Fu Anxu corresponding to each vertex B 。
In this embodiment, for Segsn A And Segsn B Performing uniqueness description, and calculating distance bd between each point and centroid i It is the normalized trace points q i Distance to centroid, which is located at origin of coordinates, thenFrom the equidistant resampling process, the points of the resampling track can also be regarded as being sequentially obtained by forward pushing from the first point, wherein the forward pushing distance is +.>Using bd i And fd i The shape of the track is described but this does not ensure the uniqueness of the shape as shown in fig. 4. The left graph shows the result of q 1 Forward pushing q 2 Process of q 2 In q 1 As the center of a circle with fd 2 On a circle of radius, q 2 At the same time take O as the center of a circle and bd 2 On a circle of radius, a false point wq' may be generated 2 The method comprises the steps of carrying out a first treatment on the surface of the Also, the process of the present invention is,the right graph shows the result of q 2 Forward pushing q 3 May generate uq 3 、vq″ 3 、wq″ 3 Three false points, thereby generating 4 tracks including the real track. By analogy, push forward to the point q m Together, generate 2 m-1 And (3) a track. Therefore, additional information is necessary to ensure track uniqueness, considering the use of corners to solve the problem, i.e., vectorsRelative to vector->Angle of deflection θ in the clockwise direction i E [0,2 pi ], 2.ltoreq.i.ltoreq.m and θ 1 =0. Since the intersection points of circles have symmetry, the sum of the angle values formed by the two intersection points is 2pi, thus θ in the figure 2 ＝θ 2 ' and theta 3 ＝θ 3 '. According to the definition of the rotation angle, the rotation angle satisfies theta i E [0, pi) the symbol centroid distance sbd i ＝-bd i While the rotation angle satisfies theta i E [ pi, 2 pi) the symbol centroid distance sbd i ＝bd i Thus, the shape of the track can be uniquely described. Calculating sbd 1 When the track is not pushed forward, the sbd can be set 1 And sbd 2 The symbols are the same.
Step S24: based on shape feature sequence SDSeg A And SDSeg B Constructing an objective function for measuring a first sub-track Seg corresponding to the particle A And a second sub-track Seg B Similarity of (2); calculating objective function values of the particles; the objective function is:
sbd i for the shape feature sequence SDSeg of the first sub-track A Is the shape feature sequence SDSeg of the second sub-track B I sbd i |、|sbd′ i The I is the absolute value of the centroid distance of the symbol corresponding to the ith vertex of the first sub-track and the second sub-track, namely the centroid distance.
In the invention, the symbol centroid distance shape descriptor is the uniqueness expression of the track shape through equidistant resampling, polygon decomposition, normalization, uniqueness description and other processes. The shape description is carried out by the shape descriptor of the symbol centroid distance, so that the invariance of the obtained shape feature sequence to geometric transformations such as translation, rotation, scaling and the like of the same shape track is realized, and further, different track sections can be registered accurately and undistorted.
The step S3, wherein:
the distributed neighbor search strategy is to construct a neighbor matrix M PD Matrix element M PD (ii, jj) is a particle Pos ii ,Pos jj The distance between the two particles, ii, jj are respectively the particle numbers, 1 is less than or equal to ii, jj is less than or equal to NP, an update mark is set for each particle, and an initial value is 0, so that the particle does not evolve update currently; for each particle, the following operations are performed in numbered order: and acquiring a particle number, searching particles which are not evolutionarily updated around the particles corresponding to the number and are nearest to the particles corresponding to the number as nearest neighbor particles, forming a particle pair by the particles corresponding to the number and the nearest neighbor particles, completing evolutionary update of the particle pair, and assigning two particle update marks of the particle pair as 1 after the evolutionary update. The distributed neighbor search strategy can finish pairing and evolutionary updating of all particles in the particle swarm, namely, the update mark of all particles is assigned to be 1. The evolution updating method is characterized in that the basis of evolution updating is to determine an evolution updating mode of particles which are not subjected to evolution updating in the particle pair based on the optimal history of each particle in the particle pair and the current better particles of two particles in the particle pair; the basis for judging the optimal history and the optimal current is the function value corresponding to the objective function.
Further, the particle evolution speed V in the particle swarm k ′＝C 1 ·V k +C 2 ·(Pb k -P k )+C 3 ·(Pnb-P k ) Wherein V is k Is the particle velocity of the last iteration, C 1 Is the inertial weight, C 2 Is a self-learning factor, C 3 Is a neighbor learning factor, pb k Is the individual history optimum, pnb is the neighboring better particle.
Further, for V k ' after boundary checking, the position of the particle is updated, and the updated particle isWherein P is k Is the state before particle update, P k (l) Is P k V is the first element of (1) k ' (l) is V k ' first element, when V k ' when (l) > 0, the particle state is rounded up and updated, and when V k And (l) when the value of the sum is less than 0, rounding down to update the particle state. Taking the updated particles as the current state of the particles, and evolutionarily updating the particles, namely setting P k ＝P k ′,V k ＝V k ' k=ii, jj completes particle evolution update, checks the numerical boundary of the particle, and sets P if the lower limit of the track segment length scale is met i ,P j The update flag of (1).
In the present embodiment, matrix element M PD (ii, jj) is a particle Pos ii ,Pos jj The distance formula between them is
The function of step S3 is described with reference to fig. 5. Based on a distributed neighbor search strategy, iterative optimization is carried out on particle swarm, and the problem that when complex multivariable optimization problems of a plurality of peaks exist in an objective function are faced, a standard particle swarm algorithm is easy to fall into local optimization is solved. The distributed neighbor search strategy is to use a global optimal solution in a particle neighbor high-quality solution substitution standard particle swarm algorithm and combine with a historical optimal solution of the particle, so as to jointly drive iterative evolution of the particle, thereby the overall particle swarm presents a distributed optimizing search situation. The strategy has the advantages that the essential capability of particle swarm evolution optimization is reserved, and the particle swarm can obtain the optimal solution of each local area through distributed neighbor searching, so that the probability of obtaining the global optimal solution is greatly increased.
As shown in fig. 6, an example of moving object trajectory similar segment extraction based on shape matching is given. The typical shape segments provided by the MPEG-7 dataset are combined with the maneuver object simulated trajectory data for similar shape segment extraction. From fig. 6, it can be intuitively seen that the method of the present invention can accurately extract similar shape track segments from the track which is non-relevant in space-time and has geometric transformation, and the 6 searched good shape pairs have good shape similarity.
As shown in fig. 7, a comparison graph of the search results of the similar shape track segments based on the distributed neighbor search particle swarm method and the standard particle swarm method is given. The parameter setting method based on the distributed neighbor search particle swarm comprises the following steps: the iteration number ni=100, and the proportion of the matching track segment to the original track length limits slp= { δ A ＝0.4,δ B =0.4 }, inertial weight C 1 =0.8, self-learning factor C 2 =1.5, neighbor learning factor C 3 =1.5, maximum and minimum update speed V max ＝5,V min = -5. The parameter setting of the standard particle swarm method comprises the following steps: iteration number ni=100, inertial weight C 1 =0.8, self-learning factor C 2 =1.5, population learning factor C 4 =1.5, maximum and minimum update speed V max ＝5,V min = -5. As can be intuitively seen from fig. 7, in 100 searches, as the number of particles increases, the advantages of the distributed neighbor search are continuously highlighted, and the optimal solution obtained based on the distributed neighbor search particle swarm method and the number of times of the optimal solution in the vicinity of the maximum and minimum update speeds are superior to those of the standard particle swarm method.
Fig. 8 is a schematic structural diagram of a moving object trajectory similarity segment extraction device based on shape matching according to an embodiment of the present invention, as shown in fig. 8, the device includes:
an initialization module: configured to obtain a first trajectory TR to be shape matched A And a second track TR B Obtain and respectively and TR A 、TR B Corresponding first track segment set TotalSeg A A second track segment set TotalSeg B ，TotalSeg A ＝{Seg Anum1 }，TotalSeg B ＝{Seg Bnum2 }, wherein Seg Anum1 For the first track segment set TotalSeg A Is not less than 1 NUM1 and not more than NUM1, NUM1 being the first track segment set TotalSeg A Number of middle track segments, seg Bnum2 For the second track segment set TotalSeg B Is not less than 1 and not more than 2 and is not more than 2, wherein NUM2 is a second track segment set TotalSeg B The number of mid-track segments; based on the first track segment set TotalSeg A The method comprises the steps of determining an original track length of a first track and a length of a first sub-track formed by a plurality of continuous track segments in the first track segment set; based on the second track segment set TotalSeg B The original track length of the second track and the length of a second sub-track formed by a plurality of continuous track segments in the second track segment set can be determined;
defining the coding mode of the particles in the particle swarm, so that the particle Pos comprises the starting track segment number id of the first sub-track AS End track segment number id AE Starting track segment number id of the second sub-track BS End track segment number id BE I.e. { id AS ,id AE ,id BS ,id BE }；
Initializing parameters of a particle swarm algorithm and randomly initializing NP particles, wherein the parameters of the particle swarm algorithm comprise a preset number NP of particles, a maximum iteration optimizing frequency N1, and a proportion limiting condition SLP= { delta of a first sub-track and a second sub-track corresponding to the particles respectively occupying the length of the corresponding original track A ,δ B }, satisfy0＜δ A ,δ B Not more than 1, wherein LSeg A 、LSeg B LTR is the length of the first sub-track and the second sub-track corresponding to the particle A 、LTR B Respectively are firstThe original track length of the track and the original track length of the second track;
the shape description module: a first sub-track Seg configured to correspond to each particle A Second sub-track Seg B Performing shape description based on the shape descriptor of the symbol centroid distance to obtain a shape feature sequence SDSeg respectively A And SDSeg B The method comprises the steps of carrying out a first treatment on the surface of the Based on shape feature sequence SDSeg A And SDSeg B Constructing an objective function for measuring a first sub-track Seg corresponding to the particle A And a second sub-track Seg B Similarity of (2); calculating objective function values of the particles;
and an optimization module: the method comprises the steps of performing iterative optimization on a particle swarm based on a distributed neighbor search strategy until the iterative optimization times reach N1, and performing first sub-track Seg corresponding to the particle with the largest objective function value in the particle swarm A And a second sub-track Seg B As a similar segment output.
The embodiment of the invention further provides a moving target track similar segment extraction system based on shape matching, which comprises the following steps:
a processor for executing a plurality of instructions;
a memory for storing a plurality of instructions;
wherein the plurality of instructions are for storage by the memory and loading and executing by the processor the method as described above.
The embodiment of the invention further provides a computer readable storage medium, wherein a plurality of instructions are stored in the storage medium; the plurality of instructions are for loading and executing by the processor the method as described above.
It should be noted that, without conflict, the embodiments of the present invention and features of the embodiments may be combined with each other.
In the several embodiments provided in the present invention, it should be understood that the disclosed systems, devices, and methods may be implemented in other manners. For example, the apparatus embodiments described above are merely illustrative, e.g., the division of the elements is merely a logical function division, and there may be additional divisions when actually implemented, e.g., multiple elements or components may be combined or integrated into another system, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices or units, which may be in electrical, mechanical or other form.
The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in hardware plus software functional units.
The integrated units implemented in the form of software functional units described above may be stored in a computer readable storage medium. The software functional unit is stored in a storage medium, and includes several instructions for making a computer device (which may be a personal computer, a physical machine Server, or a network cloud Server, etc., and need to install a Windows or Windows Server operating system) execute part of the steps of the methods described in the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
The above description is only of the preferred embodiments of the present invention, and is not intended to limit the present invention in any way, but any simple modification, equivalent variation and modification made to the above embodiments according to the technical substance of the present invention still fall within the scope of the technical solution of the present invention.
Claims (7)
1. The method for extracting the similar segments of the moving target track based on the shape matching is characterized by comprising the following steps:
step S1: acquiring a first track TR to be subjected to shape matching A And a second track TR B Obtain and respectively and TR A 、TR B Corresponding first track segment set TotalSeg A A second track segment set TotalSeg B ，TotalSeg A ＝{Seg Anum1 }，TotalSeg B ＝{Seg Bnum2 }, wherein Seg Anum1 For the first track segment set TotalSeg A Is not less than 1 NUM1 and not more than NUM1, NUM1 being the first track segment set TotalSeg A Number of middle track segments, seg Bnum2 For the second track segment set TotalSeg B Is not less than 1 and not more than 2 and is not more than 2, wherein NUM2 is a second track segment set TotalSeg B The number of mid-track segments; based on the first track segment set TotalSeg A The method comprises the steps of determining an original track length of a first track and a length of a first sub-track formed by a plurality of continuous track segments in the first track segment set; based on the second track segment set TotalSeg B The original track length of the second track and the length of a second sub-track formed by a plurality of continuous track segments in the second track segment set can be determined;
defining the coding mode of the particles in the particle swarm, so that the particle Pos comprises the starting track segment number id of the first sub-track AS End track segment number id AE Starting track segment number id of the second sub-track BS End track segment number id BE I.e. { id AS ,id AE ,id BS ,id BE }；
Initializing parameters of a particle swarm algorithm and randomly initializing NP particles, wherein the parameters of the particle swarm algorithm comprise a preset number NP of particles, a maximum iteration optimizing frequency N1, and a proportion limiting condition SLP= { delta of a first sub-track and a second sub-track corresponding to the particles respectively occupying the length of the corresponding original track A ,δ B }, satisfyWherein LSeg A 、LSeg B LTR is the length of the first sub-track and the second sub-track corresponding to the particle A 、LTR B The original track length of the first track and the original track length of the second track are respectively;
step S2: first sub-trajectories Seg corresponding to respective particles A Second sub-track Seg B Performing shape description based on the shape descriptor of the symbol centroid distance to obtain a shape feature sequence SDSeg respectively A And SDSeg B The method comprises the steps of carrying out a first treatment on the surface of the Based on shape feature sequence SDSeg A And SDSeg B Constructing an objective function for measuring a first sub-track Seg corresponding to the particle A And a second sub-track Seg B Similarity of (2); calculating objective function values of the particles;
step S3: based on a distributed neighbor search strategy, carrying out iterative optimization on a particle swarm until the iterative optimization times reach N1, and carrying out first sub-track Seg corresponding to the particle with the largest objective function value in the particle swarm A And a second sub-track Seg B Output as a similar segment;
the step S2 includes:
step S21: for the first sub-track Seg A Second sub-track Seg B Selecting m points respectively, sampling m points equidistantly, and using the first sub-track Seg A M points of said second sub-track Seg B Respectively form a first sub-track Segs after sampling A And a second sub-track Segs after sampling B Wherein the first sub-track Seg A M points of (1) include Seg A The first sub-track Seg is a first sub-track segment, and the second sub-track Seg is a second sub-track segment B M points of (1) include Seg B A first segment point of the start track segment and a second segment point of the end track segment; for Segs A 、Segs B M-sided Polygon composed of sampling points A 、Polygon B Respectively performing polygon decomposition to obtainM-2 triangles taking the segment head point of the starting point track segment as the vertex; calculating m-sided Polygon A 、Polygon B Centroid coordinates Conpolygon A Conpolygon B ；
Step S22: respectively for the first sub-track Segs after sampling A And a second sub-track Segs after sampling B Performing normalization processing, wherein the normalization processing comprises translation transformation of the mass center of the m-sided polygon corresponding to the sub-track and area normalization of the m-sided polygon; obtaining a normalized first sub-track Segsn A And a normalized second sub-track Segsn B ；
Step S23: acquiring a normalized first sub-track Segsn A Distance of each vertex of (2) to the corresponding centroid, segsn A The vertices of the sequence are ordered in a clockwise direction to obtain Segsn A A deflection angle of a vector formed by each vertex of the plurality of vertices and a point next to the vertex with respect to a vector formed by the origin of coordinates and the vertex; obtaining a normalized second sub-track Segsn B Distance of each vertex of (2) to the corresponding centroid, segsn B The vertices of the sequence are ordered in a clockwise direction to obtain Segsn B A deflection angle of a vector formed by each vertex of the plurality of vertices and a point next to the vertex with respect to a vector formed by the origin of coordinates and the vertex;
define centroid distance of vertex node as bd node Deflection angle of theta node Will deflect by an angle theta node Converted into positive sign or negative sign and then superimposed to the centroid distance to form the sign centroid distance shape descriptor sbd of the vertex node node I.e. the angle of rotation of the vertex satisfies θ node E [0, pi), the symbol centroid of vertex node is separated from shape descriptor sbd i ＝-bd node The rotation angle satisfies theta node E [ pi, 2 pi), the symbol centroid of vertex node is from shape descriptor sbd node ＝bd node ；
Acquiring a normalized first sub-track Segsn A The symbol centroid distance shape descriptors of each vertex, the symbol centroid distance shape descriptions Fu Anxu corresponding to each vertex are arranged to obtain the shape featureCharacterization sequence SDSeg A ；
Obtaining a normalized second sub-track Segsn B The symbol centroid distance shape descriptors of each vertex of the plurality of vertices are arranged to obtain a shape feature sequence SDSeg by arranging the symbol centroid distance shape descriptions Fu Anxu corresponding to each vertex B 。
2. The method of claim 1, wherein the equidistant sampling method of the m points in the step S2 includes:
Step S211: calculating the track length of a sub-track, wherein the sub-track is a first sub-track or a second sub-track; the track length of the sub track isDividing L into m-1 parts, wherein the length lm of each part is equal-distance sampling step length, n is the number of original track points contained from the first point to the tail point of the sub track, the number of the first point is 1, the number of the tail point is n, and p j+1 The j+1st point from the first point has coordinate information; p is p j A j-th point from the first point, having coordinate information; the sum of the distances between two adjacent points is the length L of the sub-track;
step S212: setting the total count total of sampling points to be 0 by taking the first point of the starting track segment of the sub track as a starting point;
step S213: moving the equidistant sampling step lm each time along the direction of the sub-track from the starting point until the requirement is metAnd->Wherein k is the number of original sub-track points including the first point which moves along the direction of the sub-track according to the step lm; determining coordinates of a new sampling point:
wherein xq is the abscissa of the new sampling point, yq is the ordinate of the new sampling point, and p k P is the last original track point of the new sampling point in the original sub-track k+1 For the next original track point of the new sampling point in the original sub-track, the new sampling point is located at p k And p k+1 Between xp k Is p k X p k+1 Is p k+1 Is the abscissa of yp k Is p k Yp, ordinate of (c) k+1 Is p k+1 Is the ordinate of (2); setting the total count total of sampling points as total+1;
step S214: if total is equal to m-2, ending the method; otherwise, the new sampling point is taken as a starting point, and the process proceeds to step S213.
3. The method of claim 2, wherein the distributed neighbor search strategy is to construct a neighbor matrix M PD Matrix element M PD (ii, jj) is a particle Pos ii ,Pos jj The distance between the two particles, ii, jj are respectively the particle numbers, 1 is less than or equal to ii, jj is less than or equal to NP, an update mark is set for each particle, and an initial value is 0, so that the particle does not evolve update currently; for each particle, the following operations are performed in numbered order: and acquiring a particle number, searching particles which are not evolutionarily updated around the particles corresponding to the number and are nearest to the particles corresponding to the number as nearest neighbor particles, forming a particle pair by the particles corresponding to the number and the nearest neighbor particles, completing evolutionary update of the particle pair, and assigning two particle update marks of the particle pair as 1 after the evolutionary update.
4. The method of claim 3, wherein the basis for the evolutionary update is to determine an evolutionarily updated manner for particles not evolutionarily updated in the particle pair based on the historical optimality of each particle in the particle pair itself and the currently preferred particles for both particles in the particle pair; the basis for judging the optimal history and the optimal current is the function value corresponding to the objective function.
5. A moving object track similar segment extraction device based on shape matching, the device comprising:
an initialization module: configured to obtain a first trajectory TR to be shape matched A And a second track TR B Obtain and respectively and TR A 、TR B Corresponding first track segment set TotalSeg A A second track segment set TotalSeg B ，TotalSeg A ＝{Seg Anum1 }，TotalSeg B ＝{Seg Bnum2 }, wherein Seg Anum1 For the first track segment set TotalSeg A Is not less than 1 NUM1 and not more than NUM1, NUM1 being the first track segment set TotalSeg A Number of middle track segments, seg Bnum2 For the second track segment set TotalSeg B Is not less than 1 and not more than 2 and is not more than 2, wherein NUM2 is a second track segment set TotalSeg B The number of mid-track segments; based on the first track segment set TotalSeg A The method comprises the steps of determining an original track length of a first track and a length of a first sub-track formed by a plurality of continuous track segments in the first track segment set; based on the second track segment set TotalSeg B The original track length of the second track and the length of a second sub-track formed by a plurality of continuous track segments in the second track segment set can be determined;
defining the coding mode of the particles in the particle swarm, so that the particle Pos comprises the starting track segment number id of the first sub-track AS End track segment number id AE Starting track segment number id of the second sub-track BS End track segment number id BE I.e. { id AS ,id AE ,id BS ,id BE }；
Initializing parameters of a particle swarm algorithm and randomly initializing NP particles, wherein the parameters of the particle swarm algorithm comprise the preset number NP of particles, the maximum iteration optimizing frequency N1, and corresponding first sub-tracks and second sub-tracks of the particles respectively occupy the corresponding sub-tracksProportional constraint SLP= { δ for original track length A ,δ B }, satisfyWherein LSeg A 、LSeg B LTR is the length of the first sub-track and the second sub-track corresponding to the particle A 、LTR B The original track length of the first track and the original track length of the second track are respectively;
the shape description module: a first sub-track Seg configured to correspond to each particle A Second sub-track Seg B Performing shape description based on the shape descriptor of the symbol centroid distance to obtain a shape feature sequence SDSeg respectively A And SDSeg B The method comprises the steps of carrying out a first treatment on the surface of the Based on shape feature sequence SDSeg A And SDSeg B Constructing an objective function for measuring a first sub-track Seg corresponding to the particle A And a second sub-track Seg B Similarity of (2); calculating objective function values of the particles;
and an optimization module: the method comprises the steps of performing iterative optimization on a particle swarm based on a distributed neighbor search strategy until the iterative optimization times reach N1, and performing first sub-track Seg corresponding to the particle with the largest objective function value in the particle swarm A And a second sub-track Seg B Output as a similar segment;
the shape description module includes:
a first sub-module: configured to the first sub-track Seg A Second sub-track Seg B Selecting m points respectively, sampling m points equidistantly, and using the first sub-track Seg A M points of said second sub-track Seg B Respectively form a first sub-track Segs after sampling A And a second sub-track Segs after sampling B Wherein the first sub-track Seg A M points of (1) include Seg A The first sub-track Seg is a first sub-track segment, and the second sub-track Seg is a second sub-track segment B M points of (1) include Seg B A first segment point of the start track segment and a second segment point of the end track segment; for Segs A 、Segs B Each of (3)M-sided Polygon composed of sampling points A 、Polygon B Respectively performing polygon decomposition to respectively obtain m-2 triangles taking the segment head point of the starting point track segment as the vertex; calculating m-sided Polygon A 、Polygon B Centroid coordinates Conpolygon A Conpolygon B ；
A second sub-module: configured to respectively sample the first sub-track Segs A And a second sub-track Segs after sampling B Performing normalization processing, wherein the normalization processing comprises translation transformation of the mass center of the m-sided polygon corresponding to the sub-track and area normalization of the m-sided polygon; obtaining a normalized first sub-track Segsn A And a normalized second sub-track Segsn B ；
And a third sub-module: configured to obtain a normalized first sub-track Segsn A Distance of each vertex of (2) to the corresponding centroid, segsn A The vertices of the sequence are ordered in a clockwise direction to obtain Segsn A A deflection angle of a vector formed by each vertex of the plurality of vertices and a point next to the vertex with respect to a vector formed by the origin of coordinates and the vertex; obtaining a normalized second sub-track Segsn B Distance of each vertex of (2) to the corresponding centroid, segsn B The vertices of the sequence are ordered in a clockwise direction to obtain Segsn B A deflection angle of a vector formed by each vertex of the plurality of vertices and a point next to the vertex with respect to a vector formed by the origin of coordinates and the vertex;
define centroid distance of vertex node as bd node Deflection angle of theta node Will deflect by an angle theta node Converted into positive sign or negative sign and then superimposed to the centroid distance to form the sign centroid distance shape descriptor sbd of the vertex node node I.e. the angle of rotation of the vertex satisfies θ node E [0, pi), the symbol centroid of vertex node is separated from shape descriptor sbd i ＝-bd node The rotation angle satisfies theta node E [ pi, 2 pi), the symbol centroid of vertex node is from shape descriptor sbd node ＝bd node ；
Acquiring a normalized first sub-track Segsn A The symbol centroid distance shape descriptors of each vertex of the plurality of vertices are arranged to obtain a shape feature sequence SDSeg by arranging the symbol centroid distance shape descriptions Fu Anxu corresponding to each vertex A ；
Obtaining a normalized second sub-track Segsn B The symbol centroid distance shape descriptors of each vertex of the plurality of vertices are arranged to obtain a shape feature sequence SDSeg by arranging the symbol centroid distance shape descriptions Fu Anxu corresponding to each vertex B 。
6. The system for extracting the similar segments of the moving target track based on shape matching is characterized by comprising the following components:
a processor for executing a plurality of instructions;
a memory for storing a plurality of instructions;
wherein the plurality of instructions are for storage by the memory and loading and executing by the processor the method of any of claims 1-4.
7. A computer-readable storage medium having stored therein a plurality of instructions; the plurality of instructions for loading and executing by a processor the method of any of claims 1-4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211432189.2A CN115828108B (en) | 2022-11-14 | 2022-11-14 | Shape matching-based moving object track similar segment extraction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211432189.2A CN115828108B (en) | 2022-11-14 | 2022-11-14 | Shape matching-based moving object track similar segment extraction method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115828108A CN115828108A (en) | 2023-03-21 |
| CN115828108B true CN115828108B (en) | 2023-11-28 |
Family
ID=85528384
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211432189.2A Active CN115828108B (en) | 2022-11-14 | 2022-11-14 | Shape matching-based moving object track similar segment extraction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115828108B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116602664B (en) * | 2023-07-17 | 2023-09-22 | 青岛市胶州中心医院 | Comprehensive diagnosis and treatment nursing system for neurosurgery patients |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104899263A (en) * | 2015-05-22 | 2015-09-09 | 华中师范大学 | Ship trajectory mining, analysis and monitoring method based on specific region |
| CN111397599A (en) * | 2020-02-25 | 2020-07-10 | 河海大学 | Improved ICCP (Integrated Circuit chip) underwater geomagnetic matching method based on triangular matching algorithm |
| CN113435025A (en) * | 2021-06-22 | 2021-09-24 | 同济大学 | Robot high-performance track automatic generation method combined with multistage optimization model |
| CN114117878A (en) * | 2021-11-29 | 2022-03-01 | 中国人民解放军国防科技大学 | Target motion trajectory segmented compression method based on improved particle swarm optimization |
-
2022
- 2022-11-14 CN CN202211432189.2A patent/CN115828108B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104899263A (en) * | 2015-05-22 | 2015-09-09 | 华中师范大学 | Ship trajectory mining, analysis and monitoring method based on specific region |
| CN111397599A (en) * | 2020-02-25 | 2020-07-10 | 河海大学 | Improved ICCP (Integrated Circuit chip) underwater geomagnetic matching method based on triangular matching algorithm |
| CN113435025A (en) * | 2021-06-22 | 2021-09-24 | 同济大学 | Robot high-performance track automatic generation method combined with multistage optimization model |
| CN114117878A (en) * | 2021-11-29 | 2022-03-01 | 中国人民解放军国防科技大学 | Target motion trajectory segmented compression method based on improved particle swarm optimization |
Non-Patent Citations (4)
| Title |
|---|
| A comparative analysis of trajectory similarity measures;Yaguang Tao，et al.;GIScience & Remote Sensing;全文 * |
| Review & Perspective for Distance Based Trajectory Clustering;Philippe Besse，et al.;arXiv;全文 * |
| 基于时空轨迹的热点区域提取;王侃;梅克进;朱家辉;牛新征;;电子科技大学学报(第06期);全文 * |
| 轨迹大数据:数据处理关键技术研究综述;高强;张凤荔;王瑞锦;周帆;;软件学报(第04期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115828108A (en) | 2023-03-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Nayakanti et al. | Wayformer: Motion forecasting via simple & efficient attention networks | |
| Zheng et al. | Pointcloud saliency maps | |
| Zamir et al. | Image geo-localization based on multiplenearest neighbor feature matching usinggeneralized graphs | |
| Zhao et al. | Where are you heading? dynamic trajectory prediction with expert goal examples | |
| Jia et al. | A fast ellipse detector using projective invariant pruning | |
| Gálvez-López et al. | Bags of binary words for fast place recognition in image sequences | |
| KR102096398B1 (en) | Method for recognizing position of autonomous mobile robot | |
| CN106774312B (en) | Method and device for determining moving track of mobile robot | |
| Xu et al. | A feasible density peaks clustering algorithm with a merging strategy | |
| CN115828108B (en) | Shape matching-based moving object track similar segment extraction method | |
| Saini et al. | A segmental HMM based trajectory classification using genetic algorithm | |
| KR102305230B1 (en) | Method and device for improving accuracy of boundary information from image | |
| Chen et al. | Lrgnet: Learnable region growing for class-agnostic point cloud segmentation | |
| Ren et al. | Emsin: enhanced multi-stream interaction network for vehicle trajectory prediction | |
| Habibi et al. | Human trajectory prediction using similarity-based multi-model fusion | |
| Yin et al. | Pse-match: A viewpoint-free place recognition method with parallel semantic embedding | |
| Xia et al. | A lightweight and detector-free 3d single object tracker on point clouds | |
| Kong et al. | Explicit points-of-interest driven Siamese transformer for 3D LiDAR place recognition in outdoor challenging environments | |
| Tsintotas et al. | The revisiting problem in simultaneous localization and mapping | |
| Böttger et al. | Accurate and robust tracking of rigid objects in real time | |
| Sa et al. | Depth grid-based local description for 3D point clouds | |
| CN112132184B (en) | Distribution center site selection method based on N-order neighbor analysis clustering | |
| Jaenal et al. | Unsupervised appearance map abstraction for indoor visual place recognition with mobile robots | |
| Takaishi et al. | Free-form feature classification for finite element meshing based on shape descriptors and machine learning | |
| Yang et al. | Detecting road network errors from trajectory data with partial map matching and bidirectional recurrent neural network model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |