CN114020934A - Method and system for integrating spatial semantic information based on knowledge graph - Google Patents
Method and system for integrating spatial semantic information based on knowledge graph Download PDFInfo
- Publication number
- CN114020934A CN114020934A CN202210005364.3A CN202210005364A CN114020934A CN 114020934 A CN114020934 A CN 114020934A CN 202210005364 A CN202210005364 A CN 202210005364A CN 114020934 A CN114020934 A CN 114020934A
- Authority
- CN
- China
- Prior art keywords
- points
- data
- knowledge
- graph
- semantic information
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/231—Hierarchical techniques, i.e. dividing or merging pattern sets so as to obtain a dendrogram
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Databases & Information Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Animal Behavior & Ethology (AREA)
- Computational Linguistics (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of robots, in particular to a method and a system for integrating spatial semantic information based on a knowledge graph; vectorizing the spatial semantic information to obtain vectorized data information; then, data information is merged into the knowledge graph; according to the invention, the semantic information is vectorized and put into the knowledge graph, so that different graph nodes can be automatically increased, increase is carried out according to actual conditions, and data information is conveniently integrated; the invention can process different spatial data information, then fuse the spatial data information into the spatial knowledge map, and can fuse the knowledge map continuously in the process to compress the data, thereby solving the problem of continuous increase of the data.
Description
Technical Field
The invention relates to the technical field of robots, in particular to a method and a system for integrating spatial semantic information based on a knowledge graph.
Background
The knowledge map is also called scientific knowledge map, is called knowledge domain visualization or knowledge domain mapping map in the book intelligence field, is a series of different graphs for displaying the relationship between the knowledge development process and the structure, describes knowledge resources and carriers thereof by using visualization technology, and excavates, analyzes, constructs, draws and displays the knowledge and the mutual relationship between the knowledge.
At present, in the robot industry, and industries such as VR, AR, MR and the like, including mixed reality and virtual reality, the traditional mode is still based on dictionary database, object characteristics, object recognition or atlas or graph theory of space objects for positioning, but the traditional mode is more troublesome and insufficient in integration when data is added.
Disclosure of Invention
The invention mainly solves the technical problem of providing a method for integrating spatial semantic information based on a knowledge graph, which can automatically increase different graph nodes by vectorizing the semantic information and putting the semantic information into the knowledge graph, follow up the actual situation for increasing and conveniently integrate data information; a method and a system for integrating the spatial semantic information based on the knowledge graph are also provided.
In order to solve the technical problems, the invention adopts a technical scheme that: the method for integrating the spatial semantic information based on the knowledge graph is provided, and comprises the following steps:
step S1, vectorizing the spatial semantic information to obtain vectorized data information;
and step S2, merging the data information into the knowledge graph.
As a modification of the present invention, step S2 includes the steps of:
step S21, defining a data set for the input data information;
s22, detecting the characteristic points in the data set and screening out the characteristic points;
step S23, clustering initialization is carried out on the feature points to obtain an initialized clustering center, and iteration processing is carried out;
s24, removing the characteristic points, then performing cluster subdivision, and keeping the characteristic points closest to the cluster center as data points;
and step S25, merging the data points into the knowledge graph.
As a further improvement of the present invention, in step S22, the curvature of the data point, the average value of the included angle between the point normal and the normal of the neighboring point, the distance from the point to the center of the neighboring area, and the average distance from the point to the neighboring point are set as parameter characteristics, and parameter characteristic thresholds are set for the parameter characteristics.
As a further improvement of the present invention, in step S22, a point at which the characteristic parameter is greater than the parameter characteristic threshold value is detected as a characteristic point.
As a further improvement of the present invention, in step S23, an adaptive octree model is established, and the initialized cluster center is obtained by segmenting through the octree model.
As a further improvement of the present invention, in step S23, an iterative update process is performed on the cluster center.
As a further improvement of the present invention, in step S24, the feature points in the data points are removed by traversing the clusters, the detailed features of the point cloud modules of the clusters are retained, so as to obtain the maximum curvature difference between the data points in each cluster, and the clusters with the maximum curvature difference larger than a set threshold are iteratively subdivided.
As a further improvement of the present invention, in step S24, after the iterative subdivision, the cluster information is updated, and the feature point closest to the cluster center is retained as the data point.
A system for integrating spatial semantic information based on knowledge graph includes:
the vector model is used for vectorizing the semantic information into data information;
and the fusion model is used for fusing the data information into the knowledge graph.
As an improvement of the invention, the fusion model comprises:
the input unit is used for defining a data set by the input data information;
the detection unit is used for detecting the characteristic points in the data set;
the initialization unit is used for carrying out clustering initialization on the feature points;
and the clustering subdivision unit is used for eliminating the characteristic points and then performing clustering subdivision.
The invention has the beneficial effects that: compared with the prior art, the semantic information is vectorized and put into the knowledge graph, so that different graph nodes can be automatically increased, increase is carried out according to actual conditions, and data information is conveniently integrated; the invention can process different spatial data information, then fuse the spatial data information into the spatial knowledge graph, and can also fuse the knowledge graph continuously in the process, merge the nodes of the knowledge graph and compress the data, thereby solving the problem of continuous increase of the data.
Drawings
FIG. 1 is a block diagram of the steps of the present invention;
fig. 2 is a block diagram of step S2 according to the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Referring to fig. 1 and fig. 2, the present invention provides a method for integrating spatial semantic information based on a knowledge graph, including the following steps:
step S1, vectorizing the spatial semantic information to obtain vectorized data information;
specifically, the method comprises the following steps: the method comprises the steps of carrying out image coding on the position, the shape, the name and the like of a spatial object according to the aspects of panorama segmentation, depth estimation, object recognition, three-dimensional object understanding and the like, then converting a content image into image vector information of k dimensions by using image convolution, wherein each dimension comprises a specific information direction, and a knowledge map is formed according to the difference of each dimension.
It can be understood that: for the growing types, whether two or more vector information can be combined into a new vector information is judged according to the distance of the information of the k dimensions of the vector. Therefore, a plurality of previous information can be compressed into one information dimension, the contained data information is increased, the storage pressure is reduced, and the same items can be combined according to continuously increased data.
And step S2, merging the data information into the knowledge graph.
According to the invention, the semantic information is vectorized and put into the knowledge graph, so that different graph nodes can be automatically increased, increase is carried out according to actual conditions, and data information is conveniently integrated; the invention can process different spatial data information, then fuse the spatial data information into the spatial knowledge map, and can fuse the knowledge map continuously in the process to compress the data, thereby solving the problem of continuous increase of the data.
Wherein, step S2 includes the following steps:
step S21, defining a data set for the input data information;
s22, detecting the characteristic points in the data set and screening out the characteristic points;
step S23, clustering initialization is carried out on the feature points to obtain an initialized clustering center, and iteration processing is carried out;
s24, removing the characteristic points, then performing cluster subdivision, and keeping the characteristic points closest to the cluster center as data points;
and step S25, merging the data points into the knowledge graph.
In the present invention, in step S22, the curvature of the data point, the average value of the included angle between the point normal and the normal of the neighborhood point, the distance from the point to the center of the neighborhood, and the average distance from the point to the neighborhood point are set as parameter characteristics, and parameter characteristic thresholds are set for the parameter characteristics; and detecting and setting the points with the characteristic parameters larger than the parameter characteristic threshold as characteristic points.
In the invention, in step S23, a self-adaptive octree model is established, and an initialized clustering center is obtained by segmenting through the octree model; and carrying out iterative updating processing on the clustering center.
In the invention, in step S24, traversing clusters to remove feature points therein, retaining detail features of point cloud modules of the clusters, thereby obtaining a maximum curvature difference between data points in each cluster, and iteratively subdividing the clusters with the maximum curvature difference larger than a set threshold; and after iterative subdivision, updating clustering information, and keeping the characteristic point closest to the clustering center as a data point.
Specifically, the input data set is defined as P = { pi (xi, yi, zi) | i =1,2, ⋯, N }, and the specific flow is as follows:
1. detecting characteristic points in the data; calculating 4 parameters of curvature Ci of a data point, an average value theta i of an included angle between a point normal and a neighborhood point normal, a distance D1i from the point to a neighborhood center and an average distance D2i from the point to the neighborhood point, and defining a characteristic threshold and a point with a characteristic discrimination parameter larger than the threshold as a characteristic point;
2. initializing self-adaptive clusters; and establishing a self-adaptive octree aiming at scattered point cloud data, setting the segmentation ending condition as the number On of data points contained in the nodes, recording the number of non-empty leaf nodes in the octree and the data points closest to the leaf node box center after the segmentation is ended, respectively taking the data points as the K value of the K-means cluster and the initialized cluster center, and then iteratively updating the cluster until the variation value of the criterion function is smaller than the tolerance epsilon.
3. Clustering and subdividing; and traversing the clusters to remove the characteristic points, updating the cluster centers, the data points and the evaluation functions after the characteristic points are removed, calculating the maximum curvature difference among the data points in each cluster in order to keep the detail characteristics of the point cloud model, performing iterative subdivision on the clusters with the maximum curvature difference larger than a set threshold, updating the cluster information again after cluster subdivision, and finally keeping the data points closest to the cluster centers.
The invention vectorizes semantic information and puts the semantic information into the graph of the knowledge graph, so that different graph nodes can be automatically increased, the increase can be carried out according to the actual situation, and the traditional dictionary library mode and the traditional graph theory mode based on semantic physical characteristics can be combined.
The invention also provides a system for integrating the spatial semantic information based on the knowledge graph, which comprises the following steps:
the vector model is used for vectorizing the semantic information into data information;
and the fusion model is used for fusing the data information into the knowledge graph.
Wherein, the fusion model includes:
the input unit is used for defining a data set by the input data information;
the detection unit is used for detecting the characteristic points in the data set;
the initialization unit is used for carrying out clustering initialization on the feature points;
and the clustering subdivision unit is used for eliminating the characteristic points and then performing clustering subdivision.
The space can be mapped, a map of the space is built, each robot and each device can carry some basic geographic information knowledge maps during initialization, then when the map is distributed to each person or each user, the data content of each robot and each device can be increased according to the actual situation, the geographic information knowledge maps of each robot and each device can be increased, and then the information is refined and summarized to learn a better geographic information knowledge map aiming at the scene of each robot and each device; the data can be collected into the whole system, information of different individuals is collected, refining is carried out, and the comprehensive positioning capacity of all robots and AR equipment is improved.
The above description is only an embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications of equivalent structures and equivalent processes performed by the present specification and drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.
Claims (10)
1. A method for integrating spatial semantic information based on knowledge graph is characterized by comprising the following steps:
step S1, vectorizing the spatial semantic information to obtain vectorized data information;
and step S2, merging the data information into the knowledge graph.
2. The method for integrating knowledge-graph-based spatial semantic information according to claim 1, wherein the step S2 comprises the following steps:
step S21, defining a data set for the input data information;
s22, detecting the characteristic points in the data set and screening out the characteristic points;
step S23, clustering initialization is carried out on the feature points to obtain an initialized clustering center, and iteration processing is carried out;
s24, removing the characteristic points, then performing cluster subdivision, and keeping the characteristic points closest to the cluster center as data points;
and step S25, merging the data points into the knowledge graph.
3. The method for integrating spatial semantic information based on a knowledge-graph according to claim 2, wherein in step S22, the curvature of the data points, the average of the angles between the normal direction of the points and the normal directions of the neighboring points, the distance from the points to the center of the neighboring region, and the average distance from the points to the neighboring points are set as parameter features, and the parameter features are all set with parameter feature thresholds.
4. The method for integrating knowledge-graph-based spatial semantic information according to claim 3, wherein in step S22, the detection of points with characteristic parameters greater than a parameter characteristic threshold is set as characteristic points.
5. The method for integrating knowledge-graph-based spatial semantic information according to claim 4, wherein in step S23, an adaptive octree model is established, and the initialized clustering center is obtained by segmentation through the octree model.
6. The method for integrating knowledge-graph-based spatial semantic information according to claim 5, wherein in step S23, the clustering center is iteratively updated.
7. The method for integrating the spatial semantic information based on the knowledge graph according to claim 6, wherein in step S24, the traversal clusters are used to remove the feature points therein, the detail features of the clustered point cloud modules are retained, so as to obtain the maximum curvature difference between the data points in each cluster, and the clusters with the maximum curvature difference larger than a set threshold are iteratively subdivided.
8. The method for integrating knowledge-graph-based spatial semantic information according to claim 7, wherein in step S24, after iterative subdivision, cluster information is updated, and feature points closest to a cluster center are retained as data points.
9. A system for integrating spatial semantic information based on knowledge graph is characterized by comprising:
the vector model is used for vectorizing the semantic information into data information;
and the fusion model is used for fusing the data information into the knowledge graph.
10. The system for integrating knowledge-graph-based spatial semantic information according to claim 9, wherein the fusion model comprises:
the input unit is used for defining a data set by the input data information;
the detection unit is used for detecting the characteristic points in the data set;
the initialization unit is used for carrying out clustering initialization on the feature points;
and the clustering subdivision unit is used for eliminating the characteristic points and then performing clustering subdivision.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210005364.3A CN114020934A (en) | 2022-01-05 | 2022-01-05 | Method and system for integrating spatial semantic information based on knowledge graph |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210005364.3A CN114020934A (en) | 2022-01-05 | 2022-01-05 | Method and system for integrating spatial semantic information based on knowledge graph |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114020934A true CN114020934A (en) | 2022-02-08 |
Family
ID=80069603
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210005364.3A Pending CN114020934A (en) | 2022-01-05 | 2022-01-05 | Method and system for integrating spatial semantic information based on knowledge graph |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114020934A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116186280A (en) * | 2022-12-27 | 2023-05-30 | 北京海致星图科技有限公司 | Knowledge-graph data compression and decompression method and system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100250479A1 (en) * | 2009-03-31 | 2010-09-30 | Novell, Inc. | Intellectual property discovery and mapping systems and methods |
| CN111125347A (en) * | 2019-12-27 | 2020-05-08 | 山东省计算中心（国家超级计算济南中心） | Knowledge graph 3D visualization method based on unity3D |
| WO2020098316A1 (en) * | 2018-11-12 | 2020-05-22 | 南京人工智能高等研究院有限公司 | Visual point cloud-based semantic vector map building method, device, and electronic apparatus |
| CN111798475A (en) * | 2020-05-29 | 2020-10-20 | 浙江工业大学 | Indoor environment 3D semantic map construction method based on point cloud deep learning |
| CN113111887A (en) * | 2021-04-26 | 2021-07-13 | 河海大学常州校区 | Semantic segmentation method and system based on information fusion of camera and laser radar |
| US20210319329A1 (en) * | 2020-03-30 | 2021-10-14 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method and apparatus for generating knowledge graph, method for relation mining |
-
2022
- 2022-01-05 CN CN202210005364.3A patent/CN114020934A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100250479A1 (en) * | 2009-03-31 | 2010-09-30 | Novell, Inc. | Intellectual property discovery and mapping systems and methods |
| WO2020098316A1 (en) * | 2018-11-12 | 2020-05-22 | 南京人工智能高等研究院有限公司 | Visual point cloud-based semantic vector map building method, device, and electronic apparatus |
| CN111125347A (en) * | 2019-12-27 | 2020-05-08 | 山东省计算中心（国家超级计算济南中心） | Knowledge graph 3D visualization method based on unity3D |
| US20210319329A1 (en) * | 2020-03-30 | 2021-10-14 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method and apparatus for generating knowledge graph, method for relation mining |
| CN111798475A (en) * | 2020-05-29 | 2020-10-20 | 浙江工业大学 | Indoor environment 3D semantic map construction method based on point cloud deep learning |
| CN113111887A (en) * | 2021-04-26 | 2021-07-13 | 河海大学常州校区 | Semantic segmentation method and system based on information fusion of camera and laser radar |
Non-Patent Citations (1)
| Title |
|---|
| 陈龙等: "自适应K-means聚类的散乱点云精简", 《中国图象图形学报》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116186280A (en) * | 2022-12-27 | 2023-05-30 | 北京海致星图科技有限公司 | Knowledge-graph data compression and decompression method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110334762B (en) | Feature matching method based on quad tree combined with ORB and SIFT | |
| CN111259936B (en) | Image semantic segmentation method and system based on single pixel annotation | |
| CN109086777B (en) | Saliency map refining method based on global pixel characteristics | |
| CN112288857A (en) | Robot semantic map object recognition method based on deep learning | |
| CN111311708B (en) | Visual SLAM method based on semantic optical flow and inverse depth filtering | |
| CN106815842A (en) | A kind of improved image significance detection method based on super-pixel | |
| CN106780639B (en) | Hash coding method based on significance characteristic sparse embedding and extreme learning machine | |
| CN114492619B (en) | Point cloud data set construction method and device based on statistics and concave-convex performance | |
| Fermin-Leon et al. | Incremental contour-based topological segmentation for robot exploration | |
| CN113592893B (en) | Image foreground segmentation method for determining combination of main body and accurate edge | |
| CN113988147B (en) | Multi-label classification method and device for remote sensing image scene based on graph network, and multi-label retrieval method and device | |
| CN114998890B (en) | Three-dimensional point cloud target detection algorithm based on graph neural network | |
| CN113392731A (en) | Modulated signal classification method and system based on graph neural network | |
| Lin et al. | Temporally coherent 3D point cloud video segmentation in generic scenes | |
| CN114020934A (en) | Method and system for integrating spatial semantic information based on knowledge graph | |
| CN112509014B (en) | Robust interpolation light stream computing method matched with pyramid shielding detection block | |
| CN112200248A (en) | Point cloud semantic segmentation method, system and storage medium under urban road environment based on DBSCAN clustering | |
| Costa et al. | Cluster analysis using self-organizing maps and image processing techniques | |
| CN112836611A (en) | Method and equipment for determining semantic graph of body part, model training and pedestrian re-identification | |
| CN117036653A (en) | Point cloud segmentation method and system based on super voxel clustering | |
| CN111860668A (en) | Point cloud identification method of deep convolution network for original 3D point cloud processing | |
| CN112580442B (en) | Behavior identification method based on multi-dimensional pyramid hierarchical model | |
| CN111986223B (en) | Method for extracting trees in outdoor point cloud scene based on energy function | |
| CN113256793A (en) | Three-dimensional data processing method and system | |
| An et al. | Automatic recognition and feature extraction of rock blocks based on 3D point cloud data analytics |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20220208 |