CN113298195A - Method and device for generating classical trajectory of offshore target and storage medium - Google Patents

Method and device for generating classical trajectory of offshore target and storage medium Download PDF

Info

Publication number
CN113298195A
CN113298195A CN202110849499.3A CN202110849499A CN113298195A CN 113298195 A CN113298195 A CN 113298195A CN 202110849499 A CN202110849499 A CN 202110849499A CN 113298195 A CN113298195 A CN 113298195A
Authority
CN
China
Prior art keywords
track
points
point
single cluster
lneps
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.)
Pending
Application number
CN202110849499.3A
Other languages
Chinese (zh)
Inventor
李小娟
吴亚非
臧义华
马兴民
梁佳
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
CETC 15 Research Institute
Original Assignee
CETC 15 Research Institute
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by CETC 15 Research Institute filed Critical CETC 15 Research Institute
Priority to CN202110849499.3A priority Critical patent/CN113298195A/en
Publication of CN113298195A publication Critical patent/CN113298195A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/20—Analysis of motion
    • G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments

Abstract

The application discloses a method and a device for generating a classic track of an offshore object and a storage medium. The method for generating the classical trajectory of the marine target comprises the following steps: acquiring a track data set of a marine target, wherein the track data set comprises a plurality of track pointsP i The data of (a); using DBSCAN algorithm, according to tracing pointP i Eps neighborhood ofLNEps(P i ) And a track pointP i AboutLNEps(P i ) The core point of (2) clustering the track data set to obtain a plurality of clusters; and determining the direction of a single cluster in the plurality of clusters, and extracting characteristic track points along the direction of the single cluster, wherein a plurality of characteristic track points of the plurality of clusters form a classical track of the offshore target. The method and the device realize data mining of the track data of the marine target and generate the classic track of the marine target.

Description

Method and device for generating classical trajectory of offshore target and storage medium
Technical Field
The present disclosure relates to the field of trajectory analysis technologies, and in particular, to a method and an apparatus for generating a classic trajectory of an offshore object, and a storage medium.
Background
The ocean area on the earth accounts for more than 70% of the global area, and various marine targets carry out military and civil maritime activities in wide sea areas. The marine targets refer to all targets in the monitoring area which are underway, and comprise reconnaissance ships, naval fleets, aircraft carriers, fishing vessels and the like. With the increase of the offshore targets, the control of the offshore situation becomes more and more important.
Monitoring and analyzing marine targets is one of the main means to deal with marine frontier conditions. At present, the scale of marine target trajectory data is huge, how to quickly and efficiently process massive target trajectory data, effectively process noise points and extract high-value information in the target trajectory data to accurately generate a marine target classical trajectory is less in current research, and the problem is to be solved urgently.
Disclosure of Invention
The application provides a method and a device for generating a classic track of an offshore object and a storage medium, which are used for solving the technical problems in the prior art.
According to an embodiment of the first aspect of the present invention, there is provided a method for generating a classic track of an offshore object, including the following steps:
acquiring a track data set of a marine target, wherein the track data set comprises a plurality of track pointsP i The data of (a);
using a density-based noisy spatial clustering DBSCAN algorithm to calculate the noise according to the tracing pointsP i Is/are as followsEpsNeighborhood zoneLNEps(P i ) And a track pointP i AboutLNEps(P i ) The track data set is clustered to obtain a plurality of clusters;
determining the direction of a single cluster in the plurality of clusters, and extracting characteristic track points along the direction of the single cluster, wherein a plurality of characteristic track points of the plurality of clusters form a classical track of the offshore target;
wherein, the track pointsP i Eps neighborhood ofLNEps(P i ) The following formula is satisfied:
Figure 246147DEST_PATH_IMAGE001
，
wherein the content of the first and second substances,distthe distance between the points of the track is represented,P m andP n are respectivelyLNEps(P i ) The start point and the end point of (c),Epsto representLNEps(P i ) The radius of (a) is greater than (b),i、kis a variable; wherein the content of the first and second substances,Epscalculated by the following formula:
Figure 104513DEST_PATH_IMAGE002
，
wherein the content of the first and second substances,μis the arithmetic mean of the distances of adjacent track points of said plurality of track points,σis the distance variance of adjacent track points in the plurality of track points,pthe number of the anchor points in the plurality of track points is the percentage of the total number of the track points,j、wis a variable;
if it is not
Figure 109509DEST_PATH_IMAGE003
Then point of trackP i To relate toLNEps(P i ) The core point of (a), wherein,MinTimeis a preset minimum time interval and is,t n representing points of trackP n The time of (a) is,t m representing points of trackP m Time of (d).
In the method, the direction of the single cluster is calculated by the average value of the COG values of the course points of all the tracks in the single cluster.
In the above method, the method of extracting the feature trajectory point along the direction of the single cluster includes:
scanning along the direction of the single cluster, wherein a scanning line during scanning is vertical to the direction of the single cluster;
when the scanning line is coincident with the starting point or the end point of one track subsection in the single cluster, calculating the number of intersection points of the scanning line and the track subsection;
and if the number of the intersection points is not less than the preset number threshold, determining the intersection points as characteristic track points.
In the above method, the method of extracting the feature trajectory point along the direction of the single cluster includes:
scanning along the direction of the single cluster, wherein a scanning line during scanning is vertical to the direction of the single cluster;
when the scanning line is coincident with the starting point or the end point of one track subsection in the single cluster, calculating the number of intersection points of the scanning line and the track subsection;
if the number of the intersection points is not less than a preset number threshold, calculating the average coordinates of all the intersection points;
and if the distance from the average coordinate to the direction of the single cluster is not less than a preset distance threshold, determining the intersection point as a characteristic track point.
In the above method, further comprising: displaying a classic trajectory of the marine target through a fisheye view;
wherein, the formula of mapping from the normal view to the fisheye view is as follows:
Figure 271018DEST_PATH_IMAGE004
wherein, in the step (A),
Figure 686956DEST_PATH_IMAGE005
representing norm symbols, C is a coordinate vector of coordinate points in a normal view, C′Is a coordinate vector mapped into the fish-eye view;ϕis the included angle between the C axis and the + X axis,ϕ′is C′The included angle with the + X axis;f fisheye (.) representsA coordinate vector mapping function is used to map the coordinate vectors,g fisheye (.) represents an angle mapping function.
In the above-mentioned method, the first step of the method,
Figure 501460DEST_PATH_IMAGE006
。
the embodiment of another aspect of the present invention further provides a device for generating a classic track of an offshore object, including:
the acquisition unit is used for acquiring a track data set of the marine target, wherein the track data set comprises a plurality of track pointsP i The data of (a);
a processing unit for using the density-based noise-possessing spatial clustering DBSCAN algorithm to calculate the point according to the trackP i Eps neighborhood ofLNEps(P i ) And a track pointP i AboutLNEps(P i ) The track data set is clustered to obtain a plurality of clusters;
the extracting unit is used for determining the direction of a single cluster in the plurality of clusters and extracting characteristic track points along the direction of the single cluster;
the generating unit is used for forming a classical track of the offshore target according to the characteristic track points of the clusters extracted by the extracting unit;
wherein, the track pointsP i Eps neighborhood ofLNEps(P i ) The following formula is satisfied:
Figure 44568DEST_PATH_IMAGE007
，
wherein the content of the first and second substances,distthe distance between the points of the track is represented,P m andP n are respectivelyLNEps(P i ) The start point and the end point of (c),Epsto representLNEps(P i ) The radius of (a) is greater than (b),i、kis a variable; wherein the content of the first and second substances,Epscalculated by the following formula:
Figure 572501DEST_PATH_IMAGE008
，
wherein the content of the first and second substances,μis the arithmetic mean of the distances of adjacent track points of said plurality of track points,σis the distance variance of adjacent track points in the plurality of track points,pthe number of the anchor points in the plurality of track points is the percentage of the total number of the track points,j、wis a variable;
if it is not
Figure 910072DEST_PATH_IMAGE003
Then point of trackP i To relate toLNEps(P i ) The core point of (a), wherein,MinTimeis a preset minimum time interval and is,t n representing points of trackP n The time of (a) is,t m representing points of trackP m Time of (d).
In the device, the direction of the single cluster is calculated by the average value of the COG values of the course points of all the tracks in the single cluster.
In the above apparatus, the extracting unit extracts the feature trajectory points along the direction of the single cluster by:
scanning along the direction of the single cluster, wherein a scanning line during scanning is vertical to the direction of the single cluster;
when the scanning line is coincident with the starting point or the end point of one track subsection in the single cluster, calculating the number of intersection points of the scanning line and the track subsection;
and if the number of the intersection points is not less than the preset number threshold, determining the intersection points as characteristic track points.
In the above apparatus, the extracting unit extracts the feature trajectory points along the direction of the single cluster by:
scanning along the direction of the single cluster, wherein a scanning line during scanning is vertical to the direction of the single cluster;
when the scanning line is coincident with the starting point or the end point of one track subsection in the single cluster, calculating the number of intersection points of the scanning line and the track subsection;
if the number of the intersection points is not less than a preset number threshold, calculating the average coordinates of all the intersection points;
and if the distance from the average coordinate to the direction of the single cluster is not less than a preset distance threshold, determining the intersection point as a characteristic track point.
In the above apparatus, the apparatus further comprises: a display unit for displaying a classic track of the marine target through a fisheye view;
wherein, the formula of mapping from the normal view to the fisheye view is as follows:
Figure 5680DEST_PATH_IMAGE009
wherein, in the step (A),
Figure 539429DEST_PATH_IMAGE005
representing norm symbols, C is a coordinate vector of coordinate points in a normal view, C′Is a coordinate vector mapped into the fish-eye view;ϕis the included angle between the C axis and the + X axis,ϕ′is C′The included angle with the + X axis;f fisheye (.) represents a coordinate vector mapping function,g fisheye (.) represents an angle mapping function.
In the above-described apparatus, the first and second air-conditioning units,
Figure 672602DEST_PATH_IMAGE010
。
an embodiment of another aspect of the present invention further provides an apparatus of a computer, where the apparatus includes: at least one processor, a memory, an input-output unit, and a display unit; the memory is used for storing program codes, and the processor is used for calling the program codes stored in the memory to execute the method for generating the classic track of the marine target.
Embodiments of yet another aspect of the present invention also provide a computer storage medium comprising instructions which, when executed on a computer, cause the computer to perform the method for generating a classic trajectory of an offshore object as described above.
The embodiment of the application improves the traditional DBSCAN algorithm, and the Eps neighborhood and core points in the algorithm are redefined. The improved DBSCAN algorithm in the application can realize that the number of clusters to be divided does not need to be specified in advance, the shape of the cluster is not biased, and the parameters for filtering noise can be input when needed. The improved DBSCAN algorithm is adopted to perform space-time clustering analysis on massive and multi-source marine target trajectory data with space-time characteristics, so that the classic trajectory of the marine target is generated quickly and accurately.
Drawings
FIG. 1 is a schematic diagram of core points, boundary points and noise points in a DBSCAN algorithm;
FIG. 2 is a flow chart of a method for generating a classic trajectory of an offshore object according to an embodiment of the invention;
FIG. 3 is a flow chart of extracting feature trace points in one embodiment of the invention;
FIG. 4 is a flow chart of extracting feature trace points in another embodiment of the present invention;
FIG. 5 is a schematic illustration of the direction and scan lines of a single cluster of an embodiment of the present invention;
FIG. 6 is a schematic diagram of a coordinate system rotated by the direction of a cluster according to an embodiment of the present invention;
FIG. 7 is a schematic fish eye view of an embodiment of the invention;
FIG. 8 is a schematic of the checkerboard distance of a subunit from its surrounding subunits in accordance with embodiments of the present invention;
FIG. 9 is a flow chart of an embodiment of the present invention for determining the area that needs to be displayed by a fisheye view;
FIG. 10 is a schematic structural diagram of a device for generating a classic track of an offshore object according to an embodiment of the invention;
FIG. 11 is a schematic structural diagram of a device for generating a classic trajectory of an offshore object according to an embodiment of the invention;
FIG. 12 is a schematic block diagram of a computer device according to an embodiment of the invention.
Detailed Description
The following embodiments of the application provide a method, a device and a storage medium for generating a classic track of an offshore object.
The classical trajectory of a marine target refers to the frequent and regular sailing trajectory of the marine target. The embodiment of the invention carries out Clustering analysis on mass track data by an improved DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm, extracts characteristic track points by using a scanning line analysis method on the basis of the Clustering analysis, excavates the classical track of a marine target and finds the track rule.
The DBSCAN algorithm is a density-based spatial clustering algorithm, can find clusters of any shape, and has the anti-noise capability. The algorithm divides the area with sufficient density into clusters and finds arbitrarily shaped clusters in a spatial database with noise, which defines clusters as the largest set of density-connected points. The algorithm utilizes the concept of density-based clustering, i.e., requires that the number of sample points contained in a certain region in the clustering space is not less than a given threshold. The DBSCAN algorithm has the obvious advantages of high clustering speed and capability of effectively processing noise points and finding spatial clusters of any shapes.
The embodiment improves the Eps neighborhood and core point in the traditional DBSCAN algorithm, and redefines the definition.
The improved DBSCAN algorithm in this embodiment relates to the basic concept:
1) eps: the neighborhood radius at density is defined.
2) Eps neighborhood: the area within the radius Eps of a given track point is called an Eps neighborhood of the track point;
in the embodiment of the invention, the track points of the marine targetP i （iTrace point data of =1,2, 3) includes lat i ，lon i ，t i Wherein, lat i Is a track pointP i Longitude, lon of i Is a track pointP i Latitude of, t i Is a track pointP i Time of (d). Thereby, the track pointP i Eps neighborhood ofLNEps(P i ) Satisfies the following formula (1):
Figure 446654DEST_PATH_IMAGE011
wherein the content of the first and second substances,distthe distance between the points of the track is represented,P m andP n are respectivelyLNEps(P i ) The start point and the end point of (c),Epsto representLNEps(P i ) The radius of (a) is greater than (b),i、kare variables.
3) Core Point (Core Point): if it is not
Figure 32487DEST_PATH_IMAGE003
Then point of trackP i To relate toLNEps(P i ) The core point of (a), wherein,MinTimeis a preset minimum time interval and is,t n representing points of trackP n The time of (a) is,t m representing points of trackP m Time of (d).
4) Edge Point (Edge Point): the boundary point is not a core point but falls within the Eps neighborhood of a certain core point.
5) Noise Point (Outlier Point): any point that is neither a core point nor a boundary point;
6) direct density-reachable (direct density-reachable): giving a track point set D, and if p is in the Eps neighborhood of q and q is a core point, then directly enabling the density of track points p to be reachable from the track points q;
7) density-accessible (density-accessible): if there is a chain of trace points p 1., pi., pn, satisfying p1= p and pn = q, pi being reachable from pi +1 with respect to the density of Eps directly, then trace point p is reachable from trace point q with respect to the density of Eps;
8) density-connected (density-connected): if there is a set of trace points, O ∈ D, such that both trace points p and q are reachable from O with respect to the Eps density, then trace points p through q are connected with respect to the Eps density.
As shown in fig. 1, point a is a core point, points B and C are boundary points, and point N is a noise point.
As shown in fig. 2, based on the improved DBSCAN algorithm, the method for generating a classic track of an offshore target according to the embodiment of the present invention includes the following steps:
step S201, obtaining a track data set of the marine target, wherein the track data set comprises a plurality of track pointsP i The data of (a);
in the actual implementation process, the data of a plurality of track points of the marine target in a certain time period and a certain space region can be acquired to form the track data set. This embodiment is not limited to this.
Step S202, using DBSCAN algorithm, according to track pointsP i Eps neighborhood of (c), and track pointsP i AboutLNEps(P i ) The core points obtained in step S201 are clustered to obtain a plurality of clusters;
the above improved DBSCAN algorithm is used in step S202, and the trace points in step S202 are the sameP i Eps neighborhood and track pointP i AboutLNEps(P i ) The definition of the core point of (2) is as above and is not described herein again.
Step S203, determining the direction of each single cluster in the plurality of clusters obtained in step S202, and extracting characteristic track points along the direction of the single cluster; the plurality of characteristic track points extracted from the plurality of clusters form a classical track of the marine target.
Traversing the plurality of clusters obtained in step S202, and executing step S203 for each single cluster, may extract a plurality of feature trace points from the single cluster, and further extract feature trace points in all clusters. The characteristic track points extracted from all the clusters form a classical track of the marine target.
In the conventional DBSCAN algorithm, the parameter values are obtained through a plurality of experiments. The embodiment adopts the method of quantile function to calculateEpsThe calculation formula is as follows:
Figure 104348DEST_PATH_IMAGE012
wherein the content of the first and second substances,μis the arithmetic average of the distances of adjacent track points in the plurality of track points in step S201,σfor the variance of the distances of adjacent track points of the plurality of track points,pthe number of anchor points in the plurality of track points is the percentage of the total number of track points,j、ware variables.
After the clustering process for the trajectory data set is completed in step S202, the acquisition of the classical trajectory is crucial in order to quantitatively describe the overall motion law of the cluster. If the average values of the longitude, the latitude and the speed of all track points in the cluster are simply taken, the obtained classical track cannot well describe all the characteristics of the original track; the trajectory features can be described to some extent by using uniform B-spline curve fitting, but the effect of important feature segments can be ignored, and the effect of minor feature segments can be emphasized.
In order to better acquire the classical trajectory, as shown in fig. 3, the direction of a single cluster is determined in step S203, and extracting the feature trajectory point along the direction of the single cluster can be implemented by the following steps:
step S301, determining the direction of the cluster (i.e. the overall direction of the cluster);
averaging the headings of all track points in a single cluster, and defining the heading as the cluster direction, wherein the specific formula is as follows:
Figure 83238DEST_PATH_IMAGE013
wherein:
C TotalLineSegmentsindicates the direction of the cluster;
C COGi represents the first in a clusteriA Ground heading (COG) value of each track point;
Sumrepresenting the total number of trace points in the cluster.
Step S302, scanning along the direction of the cluster;
a concept of a scan Line (Sweep Line) is defined, and the scan Line is a dotted Line perpendicular to the direction of the clusters. In scanning, the scanning lines are scanned along the direction of the clusters.
Step S303, when the scanning line coincides with the starting point or the end point of a certain track subsection in the cluster, calculating the number of intersection points of the scanning line and the track subsection;
step S304, judging whether the number of the intersection points is smaller than a preset number threshold, if so, continuing to scan downwards, and returning to the step S302; otherwise, step S305 is executed;
and if the number of the intersection points of the scanning lines and the track subsections is less than a preset number threshold, discarding the intersection points and continuously scanning downwards.
And step S305, determining the intersection points as characteristic track points.
All the feature track points in a single cluster can be extracted through the steps S301 to S305.
In another embodiment, as shown in fig. 4, the direction of a single cluster is determined in step S203, and the feature track points extracted along the direction of the single cluster can be implemented by the following steps:
step S401, determining the direction of the cluster (i.e. the overall direction of the cluster);
the direction of the cluster is calculated according to equation (3).
Step S402, scanning along the direction of the cluster;
a concept of a scan Line (Sweep Line) is defined, and the scan Line is a dotted Line perpendicular to the direction of the clusters. In scanning, the scanning lines are scanned along the direction of the clusters.
Step S403, when the scanning line coincides with the starting point or the end point of a track subsection in the cluster, calculating the number of intersection points of the scanning line and the track subsection;
step S404, judging whether the number of the intersection points is smaller than a preset number threshold, if so, continuing to scan downwards, and returning to the step S402; otherwise, go to step S405;
and if the number of the intersection points of the scanning lines and the track subsections is less than a preset number threshold, discarding the intersection points and continuously scanning downwards.
Step S405, calculating the average coordinates of all the intersection points;
the average coordinate of all the intersection points is the average value of the coordinates of all the intersection points.
Step S406, judging whether the distance from the average coordinate calculated in step S405 to the cluster direction is smaller than a preset distance threshold, if so, continuing to scan downwards, and returning to step S402; otherwise, go to step S407;
and if the distance from the average coordinate of all the intersection points to the cluster direction is smaller than a preset distance threshold value, discarding the intersection points and continuing to scan downwards.
And step S407, determining the intersection points as characteristic track points.
All the feature track points in a single cluster can be extracted through the steps S401 to S407. By presetting a distance threshold value, the intersection point of which the distance from the average coordinate to the cluster direction is not less than the preset distance threshold value is selected as a characteristic track point, so that a smoother classical track can be obtained.
Fig. 5 shows a schematic diagram of the direction of the clusters (i.e., the overall direction of the clusters), the scan lines. The flow shown in fig. 4 will be described below by way of example. Assuming that the preset number threshold is 3, the number of intersections between the scanning lines 5 and 6 and the track subsections is 2, and 2 is smaller than the preset number threshold 3, so that the intersections are discarded; likewise, the intersections of the scan lines 4 with the track sub-segments are discarded because the distance between the average coordinates of these intersections to the cluster direction is less than a preset distance threshold.
In another embodiment, after the average coordinates of all the intersection points are calculated in step S405, it is determined whether the distance between the average coordinate calculated this time and the average coordinate calculated last time is smaller than a preset distance threshold, if yes, the downward scanning is continued, and the step S402 is returned to; otherwise, step S407 is executed. In this way, a smoother classical trajectory can also be obtained.
For massive trace data, due to the needThe average coordinates of all the intersection points of the scanning line and the track subsegment are calculated for multiple times, the complexity of scanning line translation is increased, and in order to simplify the calculation process, the current XY coordinate system can be rotated according to the cluster direction by the rotation angle ofθThe X-axis direction in the new XY coordinate system is parallel to the cluster direction, and therefore the calculation efficiency is improved. And (4) after the average coordinate is calculated in the new XY coordinate system, converting the calculation result into a numerical value in the original XY coordinate system.
As shown in FIG. 6, the original coordinate system is an XY coordinate system, and the XY coordinate system is rotatedθThe angle results in an X ' Y ' coordinate system, with the X ' axis in the X ' Y ' coordinate system being parallel to the direction of the cluster. The coordinate transformation formula in the XY coordinate system and the X 'Y' coordinate system is shown in formula (4):
Figure 28191DEST_PATH_IMAGE014
after the classic track of the marine target is obtained, the classic track of the marine target can be displayed through the fisheye view, and therefore visualization of the classic track is achieved.
The fish-eye lens is a lens with extremely short focal length and extremely large visual angle, compared with a common lens, the middle area of the lens is enlarged, the peripheral area of the lens is reduced, and the fish-eye view is a visual effect made by imitating the imaging effect.
The basic principle of the fisheye view is to extend the elements in the whole area to the boundary direction, and the extension degree decreases from the center to the boundary. In the fish-eye view, the central area is enlarged and the edge area is compressed, thereby realizing the enhancement of the information of the central area.
In order to realize the effect of the fisheye view, a mapping relation is required to be established to map the image from the normal view to the fisheye view.
For a more intuitive discussion, the area of the fisheye view is discussed separately from the entire picture.
Firstly, the center of the fisheye view is taken as the origin of coordinates, and the original coordinates of a point in the fisheye view under a normal view are assumed to be(x,y) In the form of vector C = (C =: (C))x,y) The coordinates under the fisheye view are (x′,y′) Recorded as vector form C′=(x′,y′)。
Thus, the formula for mapping from the normal view to the fisheye view is as follows:
Figure 350588DEST_PATH_IMAGE015
c = (C =)x,y) At an angle to the + X axis ofϕ，C′=(x′,y′) At an angle to the + X axis ofϕ′Then equation (5) can be written as:
Figure 976873DEST_PATH_IMAGE016
wherein the content of the first and second substances,
Figure 819058DEST_PATH_IMAGE005
the sign of the norm is represented,f fisheye (.) represents a coordinate vector mapping function,g fisheye (.) represents an angle mapping function.
By selecting the mapping relation in length and the mapping in angle, the mapping from the normal view to the fisheye view can be realized.
To further simplify the complexity, the fisheye view is shaped as a circle, and the effect of the fisheye view is achieved by a twist in the radial direction.
It is thus possible to obtain:
Figure 449759DEST_PATH_IMAGE017
thus equation (6) can be simplified as:
Figure 944938DEST_PATH_IMAGE018
the principle is shown in fig. 7.
To implement the fisheye view in this embodiment, only a monotonic non-linear mapping from [0,1] to [0,1] needs to be established as a warping function in the radial direction.
The fish-eye view effect is realized, and meanwhile, the amplification degree can be adjusted through the mapping relation, so that a more flexible visualization scheme can be provided.
Assuming a radius of the magnified region ofr max The distance between a certain point and the center in the region isr，r′To distance the point from the center in the fisheye view,r′andrthe mapping relation of (1) is as follows:
Figure 437231DEST_PATH_IMAGE019
wherein N is a constant. It will be appreciated that the degree of amplification can be adjusted by the magnitude of N. The greater N, the greater the degree of amplification.
Is easy to knowr/r max Has a value range of [0,1]]Take into accountr′Is after mappingrCorresponding length, so do not adoptr′/r max The radial distortion degree is characterized, and different radial distortion degrees can be obtained by comparing different parameters N. In practical implementation, an appropriate value of N can be obtained by experiment.
For distances from the center at the proximal endrLength oflOf a radial line segment of (a) having a length in a fisheye viewl′Comprises the following steps:
Figure DEST_PATH_IMAGE020
qualitatively, it can be seen that: when in userWhen the ratio of the water to the oil is small,
Figure 930660DEST_PATH_IMAGE021
，lis amplified; when in userWhen the size of the particles is larger than the required size,
Figure 217416DEST_PATH_IMAGE022
，lis compressed.
In actual implementation, the areas to be enlarged and intensified by the fish-eye view can be predetermined and then displayed by the fish-eye view. Since the classical trajectories of marine targets fall within these regions, visualization of the classical trajectories can be achieved.
And dividing the ocean area into a plurality of subunits according to a global subdivision meshing method. Firstly, determining the size of each divided subunit, and making each subunit be a square for convenient calculation, and recording the side length of each subunit asd. Presetting a track density thresholdden_thrdFor any subunit, if the track density of the subunit is greater thanden_thrdThen the subunit is considered to be a high track density subunit, and the set of all high track density subunits is recorded as-A i }. Thus, a set of high trajectory density subunits is determined among a plurality of subunits obtained by dividing the ocean area according to the global subdivision meshing methodA i ｝。
The track data is considered to have a certain continuity, that is, for a track, the data is continuously distributed and changed except for the starting point and the end point. Therefore, for an area that actually passes through a larger number of tracks, more high track density subunits may be generated after spatial meshing. Therefore, before the high-track-density subunits are amplified by using the fisheye view, the high-density subunits can be further processed in advance in a clustering mode, so that the high-track-density subunits are combined to achieve a better visualization effect.
Because the number of the high track density subunits after the final clustering is finished cannot be known in advance, the shape of the final clustering cannot be known in advance, and the purpose of clustering is to divide the high track density subunits into clusters, the DBSCAN algorithm is considered to be selected for the correspondingA i And then clustering is carried out. The specific implementation method is as follows:
step S501, selecting a great curlA i All sub-units inA core subunit.
Step S502, aiming at each core subunit, searching a non-core subunit with the direct density of the core subunit to be reached to form a cluster;
the distances between the sub-cells are defined, and considering that the sub-cells are grids that do not spatially overlap with each other, a checkerboard distance may be used to define the distances between the sub-cells. For coordinates of (x 1 ,y 1 ), (x 2 ,y 2 ) Two sub-units ofA 1 ,A 2 Chessboard distance between themD(A 1 ,A 2 ) The definition is as follows:
Figure 248826DEST_PATH_IMAGE023
the chessboard distance between a sub-cell and its surrounding sub-cells is shown in fig. 8, where a square represents a sub-cell, and the number in the sub-cell is the distance between the sub-cell filled with lines and the sub-cell where the number is located.
Selecting the radius of an Eps neighborhood as r, and aiming at a core subunitA k And a subunitA j Definition ofA j FromA k The direct density can be as follows:
Figure 282160DEST_PATH_IMAGE024
by analogy, the density can be correspondingly defined to be connected with the density:
there is a set of core subunits
Figure 98937DEST_PATH_IMAGE025
For the sub-unitA j Existence of
Figure 9124DEST_PATH_IMAGE026
So that
Figure 278563DEST_PATH_IMAGE027
Then, thenA j Slave eyeA h The density of all core subunits in the structure is accessible.
For the sub-unitA j If present, ofA p, A q FromA p, A q Has an average density ofA j Then callA p, A q The densities are connected.
Specifically, step 1: for all core subunits, all subunits whose direct density is reachable are found, forming a cluster.
Step 2: for core subunits in a cluster, repeat step 1 until the cluster is no longer populated.
In step S503, the clusters where all the core subunits connected by the non-core subunit density are located are merged. Specifically, for non-core subunits in a cluster, all core subunits connected by their density are found, and the clusters where these core subunits are located are merged. This step is repeated until all non-core subunits in the cluster have been processed.
The algorithm flow is shown in fig. 9. Finally, the area needing to be enlarged and strengthened is determined to be the combined cluster through the steps S501 to S503. Displaying these regions through a fish-eye view may show the classic trajectories of the marine targets generated in the above embodiments.
In another embodiment of the present invention, an apparatus for generating a classic trajectory of an offshore object is provided, which may apply the method of the above embodiments.
As shown in fig. 10, the device for generating the classical trajectory of the marine target includes:
an obtaining unit 601, configured to obtain a track data set of the marine target, where the track data set includes a plurality of track pointsP i The data of (a);
a processing unit 602 for using DBSCAN algorithm to determine the trace pointsP i Eps neighborhood ofLNEps(P i ) And a track pointP i AboutLNEps(P i ) The core point of (2) clustering the track data set to obtain a plurality of clusters;
an extracting unit 603, configured to determine a direction of a single cluster in the multiple clusters obtained by the processing unit 602, and extract a feature track point along the direction of the single cluster;
a generating unit 604, configured to form a classical trajectory of the marine target according to the plurality of feature track points of the plurality of clusters extracted by the extracting unit 603;
wherein, the track pointsP i Eps neighborhood ofLNEps(P i ) The following formula is satisfied:
Figure 112658DEST_PATH_IMAGE028
，
wherein the content of the first and second substances,distthe distance between the points of the track is represented,P m andP n are respectivelyLNEps(P i ) The start point and the end point of (c),Epsto representLNEps(P i ) The radius of (a) is greater than (b),i、kis a variable; wherein the content of the first and second substances,Epscalculated by the following formula:
Figure 970892DEST_PATH_IMAGE029
，
wherein the content of the first and second substances,μis the arithmetic mean of the distances of adjacent track points of said plurality of track points,σis the distance variance of adjacent track points in the plurality of track points,pthe number of the anchor points in the plurality of track points is the percentage of the total number of the track points,j、wis a variable;
if it is not
Figure 862100DEST_PATH_IMAGE030
Then point of trackP i To relate toLNEps(P i ) The core point of (a), wherein,MinTimeis a preset minimum timeThe interval between the two layers is equal to each other,t n representing points of trackP n The time of (a) is,t m representing points of trackP m Time of (d).
In the above apparatus, the direction of a single cluster is calculated by averaging the COG values of all trace points in the single cluster.
In the above apparatus, the extraction unit 603 extracts the feature trajectory points along the direction of a single cluster by:
scanning along the direction of the single cluster, wherein a scanning line during scanning is vertical to the direction of the single cluster;
when the scanning line is coincident with the starting point or the end point of one track subsection in the single cluster, calculating the number of intersection points of the scanning line and the track subsection;
and if the number of the intersection points is not less than the preset number threshold, determining the intersection points as characteristic track points.
In the above apparatus, the extraction unit 603 extracts the feature trajectory points along the direction of a single cluster by:
scanning along the direction of a single cluster, wherein a scanning line during scanning is vertical to the direction of the single cluster;
when the scanning line is coincident with the starting point or the end point of one track subsection in the single cluster, calculating the number of intersection points of the scanning line and the track subsection;
if the number of the intersection points is not less than a preset number threshold, calculating the average coordinates of all the intersection points;
and if the distance from the average coordinate to the direction of the single cluster is not less than a preset distance threshold, determining the intersection point as a characteristic track point.
As shown in fig. 11, the above apparatus further includes: a display unit 605 for displaying a classic trajectory of a marine target through a fisheye view;
wherein, the formula of mapping from the normal view to the fisheye view is as follows:
Figure 822097DEST_PATH_IMAGE031
wherein, in the step (A),
Figure 709151DEST_PATH_IMAGE005
representing norm symbols, C is a coordinate vector of coordinate points in a normal view, C′Is a coordinate vector mapped into the fish-eye view;ϕis the included angle between the C axis and the + X axis,ϕ′is C′The included angle with the + X axis;f fisheye (.) represents a coordinate vector mapping function,g fisheye (.) represents an angle mapping function.
In the above-described apparatus, the first and second air-conditioning units,
Figure 234941DEST_PATH_IMAGE032
。
in the embodiment, the traditional DBSCAN algorithm is improved, the Eps neighborhood and core points in the algorithm are redefined, the algorithm can find clustering clusters with any shapes, the number of clusters to be divided does not need to be specified in advance, the shapes of the clustering clusters are not biased, and parameters for filtering noise can be input when needed. Aiming at track data of massive and multisource offshore targets accumulated for a long time, performing space-time clustering analysis on the track data of the offshore targets with space-time characteristics by adopting an improved DBSCAN algorithm, so that space-time track points with similar behaviors can be divided into the same cluster based on time or space similarity, traversing each single cluster in a plurality of clusters obtained by clustering, determining the direction of the cluster and extracting characteristic track points along the direction of the cluster, and finally forming the classical track of the offshore targets by the characteristic track points in all the clusters.
FIG. 12 is a schematic block diagram of a computer apparatus according to another embodiment of the present application. Fig. 12 shows an apparatus of a computer, the apparatus comprising: at least one processor, a memory, an input-output unit, and a display unit;
wherein the memory is used for storing program codes, and the processor is used for calling the program codes stored in the memory to execute the method for generating the marine target classical trajectory of the embodiment.
The Processor may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic, discrete hardware components, etc. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The storage may be an internal storage unit, such as a hard disk or a memory. The memory may also be an external storage device, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card, FC), and the like. Further, the memory may also include both an internal storage unit and an external storage device. The memory is used for storing the computer program and other programs and data as required. The memory may also be used to temporarily store data that has been output or is to be output. The memory may be integrated in the processor or may be provided separately from the processor.
The input and output units may be replaced by input units and output units, which may be the same or different physical entities. When they are the same physical entity, they may be collectively referred to as an input-output unit. The input-output unit may be referred to as a transceiver.
The display unit may be a display screen.
Embodiments of a further aspect of the present invention also provide a computer storage medium comprising instructions which, when run on a computer, cause the computer to perform the method for generating a classic trajectory of an offshore object of the above embodiments.
Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
Although the embodiments of the present invention have been described above, the above description is only for the convenience of understanding the present invention, and is not intended to limit the present invention. It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.

Claims (10)

1. A method for generating a classic track of an offshore object is characterized by comprising the following steps:
acquiring a track data set of a marine target, wherein the track data set comprises a plurality of track pointsP i The data of (a);
using a density-based noisy spatial clustering DBSCAN algorithm to calculate the noise according to the tracing pointsP i Eps neighborhood ofLNEps(P i ) And a track pointP i AboutLNEps(P i ) The track data set is clustered to obtain a plurality of clusters;
determining the direction of a single cluster in the plurality of clusters, and extracting characteristic track points along the direction of the single cluster, wherein a plurality of characteristic track points of the plurality of clusters form a classical track of the offshore target;
wherein, the track pointsP i Eps neighborhood ofLNEps(P i ) The following formula is satisfied:
Figure DEST_PATH_IMAGE001
，
wherein the content of the first and second substances,distthe distance between the points of the track is represented,P m andP n are respectivelyLNEps(P i ) The start point and the end point of (c),Epsto representLNEps(P i ) The radius of (a) is greater than (b),i、kis a variable; wherein the content of the first and second substances,Epscalculated by the following formula:
Figure DEST_PATH_IMAGE002
，
wherein the content of the first and second substances,μis the arithmetic mean of the distances of adjacent track points of said plurality of track points,σis the distance variance of adjacent track points in the plurality of track points,pthe number of the anchor points in the plurality of track points is the percentage of the total number of the track points,j、wis a variable;
if it is not
Figure DEST_PATH_IMAGE003
Then point of trackP i To relate toLNEps(P i ) The core point of (a), wherein,MinTimeis a preset minimum time interval and is,t n representing points of trackP n The time of (a) is,t m representing points of trackP m Time of (d).
2. The method for generating the classic track of marine targets as claimed in claim 1, wherein the direction of the single cluster is calculated by the average value of the COG values of the course to ground of all track points in the single cluster.
3. The method for generating a classic locus of marine targets as claimed in any one of claims 1 to 2, wherein the method for extracting feature locus points along the direction of the single cluster comprises:
scanning along the direction of the single cluster, wherein a scanning line during scanning is vertical to the direction of the single cluster;
when the scanning line is coincident with the starting point or the end point of one track subsection in the single cluster, calculating the number of intersection points of the scanning line and the track subsection;
and if the number of the intersection points is not less than the preset number threshold, determining the intersection points as characteristic track points.
4. The method for generating a classic locus of marine targets as claimed in any one of claims 1 to 2, wherein the method for extracting feature locus points along the direction of the single cluster comprises:
scanning along the direction of the single cluster, wherein a scanning line during scanning is vertical to the direction of the single cluster;
when the scanning line is coincident with the starting point or the end point of one track subsection in the single cluster, calculating the number of intersection points of the scanning line and the track subsection;
if the number of the intersection points is not less than a preset number threshold, calculating the average coordinates of all the intersection points;
and if the distance from the average coordinate to the direction of the single cluster is not less than a preset distance threshold, determining the intersection point as a characteristic track point.
5. The method for generating classic trajectories of offshore objects according to claim 1, further comprising:
displaying a classic trajectory of the marine target through a fisheye view;
wherein, the formula of mapping from the normal view to the fisheye view is as follows:
Figure DEST_PATH_IMAGE004
wherein, in the step (A),
Figure DEST_PATH_IMAGE005
representing norm symbols, C is a coordinate vector of coordinate points in a normal view, C′Is a coordinate vector mapped into the fish-eye view;ϕis the included angle between the C axis and the + X axis,ϕ′is C′The included angle with the + X axis;f fisheye (.) represents a coordinate vector mapping function,g fisheye (.) represents an angle mapping function.
6. The method for generating classic loci of offshore objects according to claim 5,
Figure DEST_PATH_IMAGE006
。
7. an apparatus for generating a classical trajectory of an offshore object, comprising:
the acquisition unit is used for acquiring a track data set of the marine target, wherein the track data set comprises a plurality of track pointsP i The data of (a);
a processing unit for using the density-based noise-possessing spatial clustering DBSCAN algorithm to calculate the point according to the trackP i Eps neighborhood ofLNEps(P i ) And a track pointP i AboutLNEps(P i ) The track data set is clustered to obtain a plurality of clusters;
the extracting unit is used for determining the direction of a single cluster in the plurality of clusters and extracting characteristic track points along the direction of the single cluster;
the generating unit is used for forming a classical track of the offshore target according to the characteristic track points of the clusters extracted by the extracting unit;
wherein, the track pointsP i Eps neighborhood ofLNEps(P i ) The following formula is satisfied:
Figure DEST_PATH_IMAGE007
，
wherein the content of the first and second substances,distthe distance between the points of the track is represented,P m andP n are respectivelyLNEps(P i ) The start point and the end point of (c),Epsto representLNEps(P i ) The radius of (a) is greater than (b),i、kis a variable; wherein the content of the first and second substances,Epscalculated by the following formula:
Figure 875363DEST_PATH_IMAGE002
，
wherein the content of the first and second substances,μis the arithmetic mean of the distances of adjacent track points of said plurality of track points,σis the distance variance of adjacent track points in the plurality of track points,pthe number of the anchor points in the plurality of track points is the percentage of the total number of the track points,j、wis a variable;
if it is not
Figure DEST_PATH_IMAGE008
Then point of trackP i To relate toLNEps(P i ) The core point of (a), wherein,MinTimeis a preset minimum time interval and is,t n representing points of trackP n The time of (a) is,t m representing points of trackP m Time of (d).
8. The device for generating the classic trajectory of marine targets of claim 7, wherein the direction of the single cluster is calculated by the average of the COG values of the course to ground of all track points in the single cluster.
9. An apparatus of a computer, the apparatus comprising: at least one processor, a memory, an input-output unit, and a display unit;
wherein the memory is configured to store program code and the processor is configured to invoke the program code stored in the memory to perform the method of any of claims 1-6.
10. A computer storage medium characterized in that it comprises instructions which, when run on a computer, cause the computer to perform the method of any one of claims 1-6.
CN202110849499.3A 2021-07-27 2021-07-27 Method and device for generating classical trajectory of offshore target and storage medium Pending CN113298195A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110849499.3A CN113298195A (en) 2021-07-27 2021-07-27 Method and device for generating classical trajectory of offshore target and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110849499.3A CN113298195A (en) 2021-07-27 2021-07-27 Method and device for generating classical trajectory of offshore target and storage medium

Publications (1)

Publication Number Publication Date
CN113298195A true CN113298195A (en) 2021-08-24

Family

ID=77331153

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110849499.3A Pending CN113298195A (en) 2021-07-27 2021-07-27 Method and device for generating classical trajectory of offshore target and storage medium

Country Status (1)

Country Link
CN (1) CN113298195A (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113887590A (en) * 2021-09-22 2022-01-04 中国电子科技集团公司第二十九研究所 Target typical track and area analysis method
CN114155391A (en) * 2022-01-30 2022-03-08 浙江口碑网络技术有限公司 Trajectory processing method and device and computer equipment
CN115270920A (en) * 2022-06-21 2022-11-01 中国人民解放军91977部队 Ship target classical trajectory generation method based on density space clustering
CN115878694A (en) * 2023-01-31 2023-03-31 小米汽车科技有限公司 Track mining method and device and electronic equipment
CN117349688A (en) * 2023-12-01 2024-01-05 中南大学 Track clustering method, device, equipment and medium based on peak track

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109543715A (en) * 2018-10-23 2019-03-29 武汉理工大学 A kind of ship air route is extracted and the method for track deviation detection
CN110309383A (en) * 2019-06-17 2019-10-08 武汉科技大学 Ship trajectory clustering analysis method based on improved DBSCAN algorithm
US20200210890A1 (en) * 2016-04-05 2020-07-02 Omni Ai, Inc. Trajectory cluster model for learning trajectory patterns in video data
CN111459997A (en) * 2020-03-16 2020-07-28 中国科学院计算技术研究所 Frequent mode increment mining method of space-time trajectory data and electronic equipment
CN112733935A (en) * 2021-01-08 2021-04-30 上海大学 Fishing boat fishing mode prediction method based on Stacking algorithm
CN112749743A (en) * 2021-01-04 2021-05-04 清华大学 Track space-time clustering method, system and storage device
CN113032502A (en) * 2021-02-09 2021-06-25 北京工业大学 Ship anomaly detection method based on improved track segment DBSCAN clustering
CN113032378A (en) * 2021-03-05 2021-06-25 北京工业大学 Ship behavior pattern mining method based on clustering algorithm and pattern mining

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20200210890A1 (en) * 2016-04-05 2020-07-02 Omni Ai, Inc. Trajectory cluster model for learning trajectory patterns in video data
CN109543715A (en) * 2018-10-23 2019-03-29 武汉理工大学 A kind of ship air route is extracted and the method for track deviation detection
CN110309383A (en) * 2019-06-17 2019-10-08 武汉科技大学 Ship trajectory clustering analysis method based on improved DBSCAN algorithm
CN111459997A (en) * 2020-03-16 2020-07-28 中国科学院计算技术研究所 Frequent mode increment mining method of space-time trajectory data and electronic equipment
CN112749743A (en) * 2021-01-04 2021-05-04 清华大学 Track space-time clustering method, system and storage device
CN112733935A (en) * 2021-01-08 2021-04-30 上海大学 Fishing boat fishing mode prediction method based on Stacking algorithm
CN113032502A (en) * 2021-02-09 2021-06-25 北京工业大学 Ship anomaly detection method based on improved track segment DBSCAN clustering
CN113032378A (en) * 2021-03-05 2021-06-25 北京工业大学 Ship behavior pattern mining method based on clustering algorithm and pattern mining

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
孙书盛: "基于轨迹大数据的内河集装箱船舶航行效率研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑（月刊）》 *

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113887590A (en) * 2021-09-22 2022-01-04 中国电子科技集团公司第二十九研究所 Target typical track and area analysis method
CN114155391A (en) * 2022-01-30 2022-03-08 浙江口碑网络技术有限公司 Trajectory processing method and device and computer equipment
CN115270920A (en) * 2022-06-21 2022-11-01 中国人民解放军91977部队 Ship target classical trajectory generation method based on density space clustering
CN115878694A (en) * 2023-01-31 2023-03-31 小米汽车科技有限公司 Track mining method and device and electronic equipment
CN117349688A (en) * 2023-12-01 2024-01-05 中南大学 Track clustering method, device, equipment and medium based on peak track
CN117349688B (en) * 2023-12-01 2024-03-19 中南大学 Track clustering method, device, equipment and medium based on peak track

Similar Documents

Publication Publication Date Title
CN113298195A (en) Method and device for generating classical trajectory of offshore target and storage medium
US10699134B2 (en) Method, apparatus, storage medium and device for modeling lane line identification, and method, apparatus, storage medium and device for identifying lane line
WO2019096902A1 (en) System and method for real-time large image homography processing
US11482009B2 (en) Method and system for generating depth information of street view image using 2D map
CN104217459B (en) A kind of spheroid character extracting method
CN111028327A (en) Three-dimensional point cloud processing method, device and equipment
US20210201050A1 (en) Generating training data from overhead view images
US11948331B2 (en) Guided batching
CN111192313A (en) Method for robot to construct map, robot and storage medium
CN115601511B (en) Three-dimensional reconstruction method and device, computer equipment and computer readable storage medium
US20160180586A1 (en) System and method for data compression and grid regeneration
US20230177822A1 (en) Large scene neural view synthesis
CN108256417A (en) Architecture against regulations recognition methods based on outdoor scene Point Cloud Processing
CN111931790A (en) Laser point cloud extraction method and device
Feng et al. PRIF: Primary ray-based implicit function
Zhu et al. 3D reconstruction of plant leaves for high-throughput phenotyping
CN114648640B (en) Target object monomer method, device, equipment and storage medium
JP6765353B2 (en) Structure inspection system, structure inspection equipment and structure inspection method
CN110335322B (en) Road recognition method and road recognition device based on image
US20140347373A1 (en) Method of generating terrain model and device using the same
WO2021056501A1 (en) Feature point extraction method, movable platform and storage medium
CN113034556B (en) Frequency domain correlation semi-dense remote sensing image matching method
Zhou et al. Underwater sonar target detection based on YOLOv5
JP6945772B1 (en) Learning device, object detection device and learning method
WO2020189703A1 (en) Structure detection device, structure detection method and structure detection processing program

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
WD01 Invention patent application deemed withdrawn after publication
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20210824