CN116129263A - Cluster ship formation identification method based on topological structure similarity - Google Patents
Cluster ship formation identification method based on topological structure similarity Download PDFInfo
- Publication number
- CN116129263A CN116129263A CN202211725328.0A CN202211725328A CN116129263A CN 116129263 A CN116129263 A CN 116129263A CN 202211725328 A CN202211725328 A CN 202211725328A CN 116129263 A CN116129263 A CN 116129263A
- Authority
- CN
- China
- Prior art keywords
- ship
- cluster
- formation
- topological structure
- similarity
- 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
Links
- 230000015572 biosynthetic process Effects 0.000 title claims abstract description 74
- 238000000034 method Methods 0.000 title claims abstract description 40
- 230000002093 peripheral effect Effects 0.000 claims abstract description 14
- 238000001514 detection method Methods 0.000 claims abstract description 13
- 230000008569 process Effects 0.000 claims abstract description 7
- 238000004364 calculation method Methods 0.000 claims description 30
- 238000010586 diagram Methods 0.000 claims description 7
- 230000007246 mechanism Effects 0.000 claims description 6
- 238000012549 training Methods 0.000 claims description 4
- 238000013507 mapping Methods 0.000 claims description 2
- 238000012545 processing Methods 0.000 abstract description 2
- 238000005755 formation reaction Methods 0.000 description 53
- 238000004088 simulation Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000005070 sampling Methods 0.000 description 3
- 239000000463 material Substances 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/87—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using selection of the recognition techniques, e.g. of a classifier in a multiple classifier system
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention belongs to the technical field of image processing and discloses a cluster ship formation recognition method based on topological structure similarity, which comprises the following steps of firstly, constructing a single-stage anchor-frame-free ship target key point detection model, estimating cluster ship targets in any direction by using center points, and obtaining the coordinate positions of the ship center points; secondly, expressing the peripheral outline of the cluster ship target by using a Deluo inner triangle network, and drawing the topological structure of the peripheral outline of the cluster ship formation; thirdly, calculating the similarity with the public standard ship formation by utilizing topological structure information of the formation to be identified, so as to realize cluster ship formation identification; in the process, the similarity is calculated by using a topological structure, a distance relation, a distribution range, a distribution density and a cluster ship target. The invention can effectively identify the formation of the cluster ship.
Description
Technical Field
The invention belongs to the technical field of image processing, and particularly relates to a cluster ship formation identification method based on topological structure similarity.
Background
In a large-range high-resolution optical remote sensing image, the background of ship targets is complex and changeable, various ship targets are often densely arranged in any direction, and intra-class differences are small, so that a common detection and identification model has limitations, and therefore, the cluster ship formation identification of the high-resolution visible light remote sensing image becomes a great challenge.
A population is broadly an organized, interrelated collection of targets. While the target group is composed of multiple targets, the main features of the target group are the interrelationship between targets, the organizational structure between communities, and the contextual relationship between individual targets and the whole. Cluster vessels, unlike isolated vessels, often exhibit group characteristics during motion. The movement of the ship group member is constrained by the actions of other members in the group, and the movement characteristics of the ship group member are consistent with the whole group. The movement patterns of the clusters are similar, and the space positions are kept relatively fixed in a certain time. The current methods for identifying the formation of the cluster ship are relatively few and can be roughly divided into three types: the method for identifying the cluster ship formations based on the structural unit, the method for identifying the cluster ship formations based on the graph model and the method for identifying the cluster ship formations based on the shape context. According to the arrangement mode of the group ships, various basic formations can be obtained, and the whole ship group formation can be analyzed by knowing basic units in the formations, so that the formation can be intelligently perceived.
However, such methods rely on the basic units constituting the formation, and the recognition result of the early basic units directly affects the efficiency of formation recognition. In order to solve the problems that the scale of the target cluster ship targets is changed severely, the ship targets are densely arranged in any direction and the like, part of scholars adopt a graph model to identify group formations, part of key points in the graph are extracted, then corresponding points are found in images to be matched, and the corresponding points are sequentially connected to form the graph model. Meanwhile, for cluster ship formation recognition, local feature descriptors in the graph matching method cannot be directly used, ship formation targets are in close relation with each other, and a single node (ship) cannot be used for describing the relation of the whole graph (formation). Although the shape context-based target group identification method uses a histogram to count sampling points of target edges and obtain local descriptors of the shape context, the method needs a large number of sampling points, and has a limited number of ship targets for the cluster ship formation, so that the method cannot obtain an effective ship formation identification result.
Disclosure of Invention
The invention aims to provide a cluster ship formation identification method based on topological structure similarity, and aims to solve the technical problems in the prior art.
In order to achieve the above purpose, the invention provides a method for identifying the formation of a cluster ship based on the similarity of topological structures, which comprises the following steps:
step one, constructing a single-stage anchor frame-free ship target key point detection model, estimating cluster ship targets in any direction by using a center point, and acquiring the coordinate position of the ship center point;
secondly, expressing the peripheral outline of the cluster ship target by using a Deluo inner triangle network, and drawing the topological structure of the peripheral outline of the cluster ship formation;
thirdly, calculating the similarity with the public standard ship formation by utilizing topological structure information of the formation to be identified, so as to realize cluster ship formation identification; in the process, the similarity is calculated by using a topological structure, a distance relation, a distribution range, a distribution density and a cluster ship target.
In another preferred embodiment of the invention, a single-stage anchor-frame-free ship target key point detection model is constructed in the first step, ship targets with any directional arrangement and extreme aspect ratio are detected by adopting a rotating frame, and the single-stage anchor-frame-free key point ship target detection is realized by using a central network;
extracting a target feature vector by utilizing DLA backbone network to fuse semantic and spatial information; meanwhile, a CA attention mechanism is added to capture the position and channel information of the network; the CA attention mechanism includes information embedding and coordinate attention generation, output as,
in another preferred embodiment of the present invention, the acquiring the coordinate position of the ship center point in the first step is specifically that the input image is input into the backbone network to generate a thermodynamic diagram of the ship target center point
W, H, C are the width, height and category of the target to be detected respectively; and mapping the center point to +.>
The gaussian kernel function is, above all,
in another preferred embodiment of the present invention, the problem of imbalance of positive and negative samples during training is alleviated by using a focal loss in the first step, specifically expressed as,
in another preferred embodiment of the present invention, in the second step, the constraint condition of the delousing triangle is that the circumcircle of each constructed triangle does not include other vertices at the same time as the triangle vertices, and the line segments connected by the endpoints do not intersect with each other.
In another preferred embodiment of the present invention, in the third step, the topology similarity of the cluster is calculated, the topology relationship of the cluster is represented by using the neighbors of the points, the neighbors of the fixed distances are used as parameters for describing the topology information of the cluster, different fixed distances d are tried, the number of the neighbors of each point is obtained in the topology, the similarity of the topology is calculated as,
in another preferred embodiment of the present invention, the similarity calculation of the distribution range of the clusters in the third step reflects the approximate distribution range of the clusters using the area S of the topology, and the calculation formula is as follows,
in another preferred embodiment of the present invention, the similarity calculation of the cluster distribution density in the third step is performed according to the following calculation formula,
in another preferred embodiment of the present invention, the similarity calculation of the cluster distance relationship in the third step, the calculation formula is as follows,
in another preferred embodiment of the present invention, the similarity calculation of the clustered ship targets in step three, is performed according to the following calculation formula,
in summary, the invention designs a cluster ship formation recognition method based on topological structure similarity, obtains ship node information, draws a cluster topological structure, and improves description robustness. Firstly, acquiring the coordinate position of a ship center point by adopting a single-stage anchor-frame-free ship target key point detection algorithm; modeling the peripheral outline of the clustered ship target through a Delaunay triangle network, and drawing the topological structure of the clustered ship target; and finally, calculating the similarity with the standard formation, and realizing the identification of the formation of the cluster ship, thereby judging the formation and the change condition of the cluster.
Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention.
Drawings
The foregoing and/or additional aspects and advantages of the invention will become apparent and may be better understood from the following description of embodiments taken in conjunction with the accompanying drawings in which:
fig. 1 is a flow chart of an embodiment of the present application.
Fig. 2 is a schematic diagram of a ship center point location in an embodiment of the present application.
Fig. 3 is a delousing triangle network for cluster ship formation in an embodiment of the present application.
FIG. 4 is a standard aircraft carrier combat group formation in an embodiment of the present application.
Fig. 5 is a cluster ship formation pod in an embodiment of the present application.
Detailed Description
Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to like or similar elements or elements having like or similar functions throughout. The embodiments described below by referring to the drawings are illustrative only and are not to be construed as limiting the invention.
In the description of the present invention, it should be understood that the terms "longitudinal," "transverse," "vertical," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," and the like indicate or are based on the orientation or positional relationship shown in the drawings, merely to facilitate description of the invention and to simplify the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be configured and operated in a particular orientation, and thus should not be construed as limiting the invention.
In the description of the present invention, unless otherwise specified and defined, it should be noted that the terms "mounted," "connected," and "coupled" are to be construed broadly, and may be, for example, mechanical or electrical, or may be in communication with each other between two elements, directly or indirectly through intermediaries, as would be understood by those skilled in the art, in view of the specific meaning of the terms described above.
The invention provides a cluster ship formation identification method based on topological structure similarity, as shown in fig. 1, in a preferred embodiment of the invention, the method comprises the following steps:
step one, constructing a single-stage anchor frame-free ship target key point detection model, estimating cluster ship targets in any direction by using a center point, and obtaining the coordinate position of the ship center point.
Constructing a single-stage anchor frame-free ship target key point detection model:
the ship targets with any directional arrangement and extreme aspect ratio are detected by adopting the rotating frame, meanwhile, the central point of the air-to-ground target is not influenced by directional change, and the central network is used for realizing the single-stage anchor-frame-free key point ship target detection.
Firstly, fusing semantic and spatial information by using a DLA backbone network, and extracting a target feature vector; meanwhile, a CA attention mechanism is added, so that the position and channel information of the network can be captured, the attention degree of the ship target is enhanced, and the expression capability of the network for extracting the target characteristics is further improved.
The CA attention mechanism comprises information embedding and coordinate attention generation, and the output is as follows:
wherein x is c (i, j) is an input feature map,
attention weighting of the input feature for the height direction, +.>
Attention weights are given to input features for width directions.
The central point position estimation network obtains the position of a ship central point:
inputting the input image into the feature extraction backbone network proposed in the steps to generate thermodynamic diagram of the ship target center point
W, H and C are the width, height and category of the target to be detected respectively. Mapping center points to gaussian kernel functions
The gaussian kernel function is, above all,
where p is the center point position of the target, (x, y) is the target coordinate position,
is the position of the center point after the down-sampling.
Most of samples on the thermodynamic diagram are negative samples, a small number of existing center points are positive samples, the problem of unbalance of the positive and negative samples in the training process is relieved by adopting focal loss, the specific formula is that,
wherein N is the number of ship targets in the network, alpha, beta is the penalty coefficient in the training process,
predictive graph, Y, for center point xyc Is a true heat map.
And secondly, expressing the peripheral outline of the cluster ship target by using a Delong triangle network, and drawing the topological structure of the peripheral outline of the cluster ship formation. The deluxe triangle can represent the boundary of an area with any shape, and the cluster point group target added with random interference factors can be well characterized. Compared with a single target or a plurality of targets, the group behavior is an integral line, and the outer contour (convex hull) of the clustered ship targets is constructed by using the Delong triangle planing division method with simple structure and small data redundancy, so that the effective expression of the dot group shape distribution is facilitated.
Firstly, looking at a ship target central point as a node of a graph structure, setting a data set as V=P U L, and setting P as a ship central point (regarded as discrete points), wherein the points are seen from each other simultaneously; l is the endpoint set of the line segments connecting the center points, and the constraint condition of the Delong triangle is that the circumscribed circle of each constructed triangle does not contain other vertexes of the triangle at the same time, and the line segments connected by the endpoints do not cross each other.
And analyzing the ship as a discrete point, and acquiring the coordinate position of the central point of the ship target in the cluster formation by utilizing the step one, wherein the central point of the ship target is regarded as a node of a graph structure, and the node is specifically shown in fig. 2.
And constructing the peripheral outline (convex hull) of the clustered ship targets, geometrically calculating the peripheral outline (convex hull) of the clustered targets by using convex hulls of convex polygons, connecting each node among the clusters by using a Delong triangle planing method on the basis that the cluster nodes (the central points of the formed ship) are acquired, and constructing the peripheral outline (convex polygon) of the point clusters, as shown in figure 3.
And thirdly, closely relating the distribution of formation formations of the cluster ships with the convex hull, and establishing a topological structure of the cluster ship target by using the Delong triangle network in the second step. Calculating the similarity with the public standard ship formation by using the topological structure information of the formation to be identified, thereby realizing cluster ship formation identification; in the process, the similarity is calculated by using a topological structure, a distance relation, a distribution range, a distribution density and a cluster ship target.
Cluster topology similarity calculation:
using the topological relation of the clusters to utilize the neighbor representation of the points, adopting the neighbors with fixed distances as parameters for describing the topological information of the clusters, trying different fixed distances d, obtaining the neighbor number of each point in the topological structure, calculating the similarity of the topological structure as,
wherein N is the number of the neighbors of the cluster ship targets, and N is the number of the ship targets.
Similarity calculation of cluster distribution range:
the area S of the convex hull is used to reflect the approximate distribution range of the clusters, the calculation formula is as follows,
similarity calculation of cluster distribution density:
the calculation of the cluster distribution density is the ratio of the number n of the cluster ships to the area S of the convex hull, reflects the sparseness degree, has the following calculation formula,
similarity calculation of cluster distance relation:
the distance similarity of the cluster ship targets is calculated by using an external minimum matrix (the length is X and the width is Y), the shape distribution of the clusters is reflected, the calculation formula is as follows,
similarity calculation of cluster ship targets:
in order to realize more stable geometric mean value when the problem of the difference is treated, the similarity of the targets of the cluster ships is generally calculated according to the following calculation formula,
experiment: by using the technical scheme provided by the invention, simulation experiments are carried out.
1. Simulation conditions
In order to test the effectiveness of the invention, the effectiveness of cluster ship target identification based on similarity calculation is verified, and the following experiment is carried out. 200 pieces of cluster ship formation data are generated by using a simulation platform, and experiments are carried out on the basis of a Ubuntu18.04 system and a deep learning Pytorch framework.
Wherein, fig. 4 is a diagram showing the judgment of the formation and the variation degree of the clustered ship targets by calculating the similarity between the clustered ship and the disclosed clustered ship target formation; fig. 5 is a cluster ship formation and peripheral profile.
2. Simulation experiment results
(1) The statistical results of the calculation factors required by the calculation of the cluster ship formation similarity are shown in Table 1
(2) The similarity and overall similarity calculation results of the factors of the clusters are shown in Table 2
| Cluster ship | SIM Topo | SIM Area | SIM thic | SIM sp | SIM |
| Clusters 1, 2 | 0.772 | 0.271 | 0.560 | 0.864 | 0.564 |
| Clusters 1, 3 | 0.860 | 0.876 | 0.861 | 0.914 | 0.877 |
| Clusters 2, 3 | 0.862 | 0.309 | 0.604 | 0.896 | 0.616 |
In fig. 4, the cluster (b) 2 and the cluster (c) 3 are single-double aircraft carrier departure formation diagrams in fig. 3, the formation and the variation degree of the clustered ship targets are judged by calculating the similarity between the clustered ship and the disclosed clustered ship target formation in fig. 4, and the experimental results are shown in table 1 and table 2.
As can be seen from table 1, the similarity results of the clusters 1 and 3 are higher, and the formation of the clusters 1 and 3 are similar to the double-aircraft carrier departure formation by combining the single-double-aircraft carrier departure formation of the disclosed cluster ship (fig. 3, 4 (d)), and the change is not great in a certain period of time.
In summary, the invention adopts the ship target detection model based on the key points to obtain the coordinate position of the ship center point; secondly, analyzing the ship as discrete points, and drawing a topological structure (convex hull) of the peripheral outline of the cluster ship formation by using a Delong triangle planing division method according to the ship target central point as a node of the graph structure; and finally, calculating the density distribution of the cluster ship by using the area of the convex hull and the number of target points in the cluster, and further realizing comparison and identification with the public standard formation.
In the description of the present specification, reference to the terms "preferred implementation," "one embodiment," "some embodiments," "example," "a particular example" or "some examples" and the like means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiments or examples. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
While embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that: many changes, modifications, substitutions and variations may be made to the embodiments without departing from the spirit and principles of the invention, the scope of which is defined by the claims and their equivalents.
Claims (10)
1. The method for identifying the formation of the clustered ship based on the similarity of the topological structure is characterized by comprising the following steps of:
step one, constructing a single-stage anchor frame-free ship target key point detection model, estimating cluster ship targets in any direction by using a center point, and acquiring the coordinate position of the ship center point;
secondly, expressing the peripheral outline of the cluster ship target by using a Deluo inner triangle network, and drawing the topological structure of the peripheral outline of the cluster ship formation;
thirdly, calculating the similarity with the public standard ship formation by utilizing topological structure information of the formation to be identified, so as to realize cluster ship formation identification; in the process, the similarity is calculated by using a topological structure, a distance relation, a distribution range, a distribution density and a cluster ship target.
2. The method for identifying the formation of the clustered ship based on the topological structure similarity, which is characterized by comprising the following steps of: constructing a single-stage anchor-frame-free ship target key point detection model, detecting ship targets with any directional arrangement and extreme aspect ratio by adopting a rotating frame, and realizing single-stage anchor-frame-free key point ship target detection by using a central network;
extracting a target feature vector by utilizing DLA backbone network to fuse semantic and spatial information; meanwhile, a CA attention mechanism is added to capture the position and channel information of the network; the CA attention mechanism includes information embedding and coordinate attention generation, output as,
3. the method for identifying the formation of the clustered ship based on the topological structure similarity according to claim 2, wherein the method comprises the following steps: the step one of acquiring the coordinate position of the ship center point includes inputting the input image into the backbone networkGenerating thermodynamic diagram of ship target central point
W, H, C are the width, height and category of the target to be detected respectively; and mapping the center point to +.>
The gaussian kernel function is, above all,
4. the method for identifying the formation of the clustered ship based on the topological structure similarity according to claim 3, wherein the method comprises the following steps of: in the first step, focaloss is adopted to relieve the problem of unbalance of positive and negative samples in the training process, the specific formula is as follows,
5. the method for identifying the formation of the clustered ship based on the topological structure similarity, which is disclosed in claim 4, is characterized in that: in the second step, the constraint condition of the delousing triangle is that the circumscribed circle of each constructed triangle does not contain other vertexes at the same time as the vertex of the triangle, and the line segments connected by the endpoints do not intersect with each other.
6. The method for identifying the formation of the clustered ship based on the topological structure similarity, which is disclosed in claim 5, is characterized in that: calculating the similarity of the topological structure of the cluster, using the neighbor representation of the points of the topological relation of the cluster, adopting the neighbors of the fixed distances as parameters for describing the topological information of the cluster, trying different fixed distances d, obtaining the neighbor number of each point in the topological structure, calculating the similarity of the topological structure as follows,
7. the method for identifying the formation of the clustered ship based on the topological structure similarity, which is characterized by comprising the following steps of: in the third step, the similarity calculation of the cluster distribution range uses the area S of the topological structure to reflect the approximate distribution range of the clusters, the calculation formula is as follows,
8. the method for identifying the formation of the clustered ship based on the topological structure similarity, which is characterized by comprising the following steps of: in the third step, the similarity calculation of the cluster distribution density is carried out, the calculation formula is as follows,
9. the method for identifying the formation of the clustered ship based on the topological structure similarity, which is characterized by comprising the following steps of: in the third step, the similarity calculation of the cluster distance relation is carried out, the calculation formula is as follows,
10. the method for identifying the formation of the clustered ship based on the topological structure similarity, which is characterized by comprising the following steps of: in the third step, the similarity calculation of the cluster ship targets is carried out, the calculation formula is as follows,
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211725328.0A CN116129263A (en) | 2022-12-30 | 2022-12-30 | Cluster ship formation identification method based on topological structure similarity |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211725328.0A CN116129263A (en) | 2022-12-30 | 2022-12-30 | Cluster ship formation identification method based on topological structure similarity |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116129263A true CN116129263A (en) | 2023-05-16 |
Family
ID=86293922
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211725328.0A Pending CN116129263A (en) | 2022-12-30 | 2022-12-30 | Cluster ship formation identification method based on topological structure similarity |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116129263A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117372946A (en) * | 2023-09-19 | 2024-01-09 | 日照市规划设计研究院集团有限公司 | Tourist group tourist behavior identification method |
-
2022
- 2022-12-30 CN CN202211725328.0A patent/CN116129263A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117372946A (en) * | 2023-09-19 | 2024-01-09 | 日照市规划设计研究院集团有限公司 | Tourist group tourist behavior identification method |
| CN117372946B (en) * | 2023-09-19 | 2024-04-16 | 日照市规划设计研究院集团有限公司 | Tourist group tourist behavior identification method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111723675B (en) | Remote sensing image scene classification method based on multiple similarity measurement deep learning | |
| CN107092870B (en) | A kind of high resolution image Semantic features extraction method | |
| CN113658100B (en) | Three-dimensional target object detection method, three-dimensional target object detection device, electronic equipment and storage medium | |
| Maitra et al. | A hybrid cooperative–comprehensive learning based PSO algorithm for image segmentation using multilevel thresholding | |
| CN109671070B (en) | Target detection method based on feature weighting and feature correlation fusion | |
| CN112883839B (en) | Remote sensing image interpretation method based on adaptive sample set construction and deep learning | |
| CN110458172A (en) | A kind of Weakly supervised image, semantic dividing method based on region contrast detection | |
| CN107862702B (en) | Significance detection method combining boundary connectivity and local contrast | |
| CN114926469B (en) | Semantic segmentation model training method, semantic segmentation method, storage medium and terminal | |
| CN111414923B (en) | Indoor scene three-dimensional reconstruction method and system based on single RGB image | |
| Khan et al. | Color image segmentation: a novel spatial fuzzy genetic algorithm | |
| CN109034035A (en) | Pedestrian's recognition methods again based on conspicuousness detection and Fusion Features | |
| CN117746260B (en) | Remote sensing data intelligent analysis method and system | |
| CN114694038A (en) | High-resolution remote sensing image classification method and system based on deep learning | |
| CN110633734B (en) | Method for anomaly detection based on graph theory correlation theory | |
| CN111739017A (en) | Cell identification method and system of microscopic image under sample unbalance condition | |
| CN110084136A (en) | Context based on super-pixel CRF model optimizes indoor scene semanteme marking method | |
| CN116129263A (en) | Cluster ship formation identification method based on topological structure similarity | |
| CN118351320B (en) | Instance segmentation method based on three-dimensional point cloud | |
| CN109685806A (en) | Image significance detection method and device | |
| CN104021560A (en) | Image segmentation method based on multi-objective evolutionary algorithm | |
| Tran et al. | 3D point cloud registration based on the vector field representation | |
| CN109815887B (en) | Multi-agent cooperation-based face image classification method under complex illumination | |
| CN108898157B (en) | Classification method for radar chart representation of numerical data based on convolutional neural network | |
| CN116452940A (en) | Three-dimensional instance segmentation method and system based on dense and sparse convolution fusion |
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 |