CN114817014A - Method for avoiding graph nodes in three-dimensional scene - Google Patents
Method for avoiding graph nodes in three-dimensional scene Download PDFInfo
- Publication number
- CN114817014A CN114817014A CN202210391939.XA CN202210391939A CN114817014A CN 114817014 A CN114817014 A CN 114817014A CN 202210391939 A CN202210391939 A CN 202210391939A CN 114817014 A CN114817014 A CN 114817014A
- Authority
- CN
- China
- Prior art keywords
- dimensional
- information
- node
- graph
- model
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/36—Prevention of errors by analysis, debugging or testing of software
- G06F11/3668—Testing of software
- G06F11/3672—Test management
- G06F11/3688—Test management for test execution, e.g. scheduling of test suites
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/36—Prevention of errors by analysis, debugging or testing of software
- G06F11/3668—Testing of software
- G06F11/3672—Test management
- G06F11/3692—Test management for test results analysis
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T19/00—Manipulating 3D models or images for computer graphics
- G06T19/20—Editing of 3D images, e.g. changing shapes or colours, aligning objects or positioning parts
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2200/00—Indexing scheme for image data processing or generation, in general
- G06T2200/04—Indexing scheme for image data processing or generation, in general involving 3D image data
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Hardware Design (AREA)
- Biophysics (AREA)
- Software Systems (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Quality & Reliability (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Graphics (AREA)
- Genetics & Genomics (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Physiology (AREA)
- Molecular Biology (AREA)
- Architecture (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Geometry (AREA)
- Computational Linguistics (AREA)
- Processing Or Creating Images (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a method for avoiding graph nodes in a three-dimensional scene, which comprises the following steps: collecting region information and constructing a three-dimensional model; constructing a three-dimensional graph and collecting node information of the three-dimensional graph; screening nodes of the three-dimensional graph; running a simulation graph to perform avoidance test; optimizing avoidance information of the three-dimensional graphs; and optimizing the related path information, and greatly improving the feasibility of the subsequent path.
Description
Technical Field
The invention relates to the technical field of three-dimensional testing, in particular to a method for avoiding graph nodes in a three-dimensional scene.
Background
Three-dimensional models are polygonal representations of objects, typically displayed by a computer or other video device, and the displayed objects may be real-world entities or fictional objects. Anything that exists in physical nature can be represented by a three-dimensional model, which is often generated using specialized software such as a three-dimensional modeling tool, but can be generated in other ways. The three-dimensional model may be generated manually or according to a certain algorithm as data of points and other information sets. Although usually present in a virtual manner in a computer or computer file, similar models described on paper can also be considered as three-dimensional models. Three-dimensional models are used broadly wherever three-dimensional graphics are used. In fact, their application has been earlier than the popularity of three-dimensional graphics on personal computers. Many computer games use pre-rendered three-dimensional model images as sprites for real-time computer rendering, and three-dimensional model tests are becoming the main test mode for people to debug various devices as science and technology continuously progresses;
the accuracy of each group of node information collected by the existing method for mutually avoiding the graph nodes in the three-dimensional scene is low, so that the measurement accuracy is low easily, and the working quality is reduced; in addition, the existing method for mutually avoiding the graph nodes in the three-dimensional scene needs manual recording by workers, so that the working efficiency of the workers is reduced, the related path information cannot be optimized, and the feasibility of a subsequent path is reduced; therefore, a method for avoiding graph nodes from each other in a three-dimensional scene is provided.
Disclosure of Invention
The invention aims to solve the defects in the prior art, and provides a method for avoiding graph nodes in a three-dimensional scene.
In order to achieve the purpose, the invention adopts the following technical scheme:
a method for avoiding graph nodes from each other in a three-dimensional scene comprises the following steps:
collecting region information and constructing a three-dimensional model;
constructing a three-dimensional graph and collecting node information of the three-dimensional graph;
screening nodes of the three-dimensional graph;
running a simulation graph to perform avoidance test;
and optimizing avoidance information of the three-dimensional graphs.
Preferably, the collecting region information and constructing a corresponding three-dimensional model specifically include:
the server receives image data sent by the remote sensing satellite, captures environment information from the image data according to the area name, extracts obstacle feature data from the environment information, and obtains the actual size of an obstacle according to the obstacle feature data;
constructing a corresponding obstacle model according to the obstacle feature data, and reducing the model according to the scaling;
constructing a space rectangular coordinate system, performing data fusion on the space rectangular coordinate system and the image data, collecting coordinate point information from the obstacle characteristic data, and splicing an obstacle model to the space rectangular coordinate system according to the coordinate point information to produce a three-dimensional model corresponding to the area name;
collecting all the three-dimensional models, constructing an area storage library to store the three-dimensional models, and constructing a data connection between the area storage library and a server;
the environment information is the building, the terrain and the vegetation coverage information of the selected area of the staff, and the information is presented in a digital form; including the coordinates of the position of the building and the digital representation of the vegetation coverage area.
Preferably, the data connection is constructed between the local storage library and the server, and the method comprises the following steps: creating a database table in the regional storage base, and recording the three-dimensional model into the database table;
and creating a class model, and establishing connection between the server and the regional storage library in the initialization process of the class model.
Preferably, a three-dimensional graph is constructed, and node information of the three-dimensional graph is collected, specifically:
the server collects image data of each obstacle model under different visual angles, Gaussian smoothing processing is carried out on the image data to filter noise interference, and then filtering processing is carried out on the image data after the noise interference is removed;
comparing the image data of the same barrier model under different viewing angles to find out corner points, and constructing a corner point coordinate set of the barrier model according to the found corner points; the set of corner coordinates is calculated by the following equation (1):
Tn＝((x 1 ，y 1 )、(x 2， y 2 )，(x 3 ，y 3 )，…，(x n ，y n ) }; formula (1)
Wherein x is n The abscissa, y, representing the corner point n A vertical coordinate representing a corner point;
when the observation visual angle of the obstacle model is changed, calculating the gray level change of pixel points in the window before and after sliding, and referring to formulas (2), (3) and (4):
diagonalizing the covariance matrix M of the formula (3), collecting two groups of characteristic values after diagonalization, and judging the two groups of characteristic values;
measuring the response of each angular point to obtain a measurement value, comparing the measurement value with a threshold value, and marking the measurement value larger than the threshold value, wherein a pixel point corresponding to the measurement value is the angular point position coordinate in the image; the specific metric is as follows (5):
R＝detM-k(traceM) 2 formula (5)
Wherein k represents a constant and generally takes a value of 0.04-0.06;
the threshold is a server default threshold or a threshold set by a worker;
carrying out point cloud clustering processing on the three-dimensional graph, detecting image gradient information through 3DHarris, and calculating coordinate information of nodes of the three-dimensional graph to obtain a calculation result;
comparing the calculation result with a set threshold value; and if the calculation result is greater than the set threshold, the node is a local maximum value point, the node is judged to be a key node, and otherwise, the node is not the key node.
Preferably, the judging of the two groups of feature values specifically includes:
when the two groups of characteristic values are large, namely the window contains angular points;
when one of the two groups of characteristic values is larger and the other is smaller, the window contains edges;
when the characteristic values are all small, the window is in a flat area.
Preferably, the screening of the nodes of the three-dimensional graph specifically comprises:
training and optimizing a learning network model, and importing the collected node coordinate information of the three-dimensional graph into the learning network model;
the learning network model normalizes the pixel point images corresponding to the node coordinate information of the three-dimensional graph, and measures the probability distribution of the gray level of each interval of the normalized pixel point images;
and collecting measurement results, screening out nodes lower than the preset node coordinate standard, comparing all groups of node information in the same three-dimensional graph, deleting repeated and redundant node information, and constructing a node record table to record all groups of node information.
Preferably, the training and optimizing of the learning network model specifically includes:
the learning network model is in communication connection with a node standard library of a server, and node information stored in the node standard library is extracted;
wherein, the node information is the collected coordinate point information;
constructing a node information set according to the collected node information, selecting one node information as verification data, and repeatedly using the verification data to verify the precision of the test model;
selecting any subset as a test set for each group of node information, then taking the rest subsets as a training set, carrying out primary prediction on each group of data, and outputting the data with the highest prediction result accuracy as an optimal parameter;
and carrying out standardization processing on the training data set according to the optimal parameters, finally conveying the training samples to the learning network model, and updating the learning network model in real time through iterative training.
Preferably, the avoidance measurement of the plurality of three-dimensional graphs specifically includes:
moving the preset path parameters of each group of three-dimensional graphs, and recording the coordinate information of each node of each group of three-dimensional graphs in real time;
and calculating the distance of the x axis, the y axis and the z axis of each group of coordinates of different three-dimensional graphs, judging that the two groups of three-dimensional graphs collide when the distance of the node coordinates of the two groups of three-dimensional graphs is less than 0, and recording the information of the two groups of node coordinates which collide.
Preferably, the updating the avoidance information of the plurality of three-dimensional graphs according to the analysis result specifically includes:
the server receives the coordinate information of the collided nodes;
and constructing a prediction model, collecting coordinate information paths of all groups of collided nodes, capturing a genetic algorithm from the Internet by a server, and updating the path information of all groups.
Compared with the prior art, the invention has the beneficial effects that:
1. compared with the conventional avoidance method, the method for avoiding the graph nodes from each other in the three-dimensional scene comprises the steps of collecting image data of barrier models in three-dimensional models of each region under different visual angles through a server, finding out corresponding corner points for matching, constructing a corner point coordinate set of each barrier according to each set of found corner points, detecting used image gradient information through 3DHarris, calculating the coordinate information of each set of key nodes of each set of three-dimensional models, introducing the collected coordinate information of each set of nodes into a learning network model, normalizing pixel point images corresponding to the coordinate information of each set of nodes through the learning network model, measuring the probability distribution of gray levels of the normalized pixel point images from interval to interval, screening out nodes lower than the standard of the node coordinates set by a worker, and deleting repeated and redundant node information, the node record table is built to record each group of node information, and the node information is screened out by building a learning network model, so that the accuracy of each group of node information can be improved, the accuracy of a test result is improved, and the working quality is improved;
2. the method for avoiding the graph nodes from each other in the three-dimensional scene includes that a server moves each group of three-dimensional graphs according to route information set by the server, coordinate information of each node of each group of three-dimensional graphs is recorded in real time, distance calculation is carried out on an x axis, a y axis and a z axis of each group of coordinates of different three-dimensional graphs, when the distance between the node coordinates of the two groups of three-dimensional graphs is smaller than 0, collision of the two groups of three-dimensional graphs is judged, coordinate information of the two groups of nodes which collide is recorded at the same time, the server receives the coordinate information of the nodes which collide, a prediction model is constructed, paths of the node coordinate information which collide are collected, the server automatically captures related genetic algorithms from the internet, updates each group of path information, synchronizes the updated path information to the corresponding three-dimensional graphs, and can collect and record the collision node information, the manual recording of workers is not needed, the working efficiency of the workers is improved, meanwhile, the related path information is optimized, and the feasibility of the follow-up path is greatly improved.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention.
Fig. 1 is a flow chart of a method for avoiding graph nodes from each other in a three-dimensional scene according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments.
Example 1
Referring to fig. 1, a method for avoiding graph nodes from each other in a three-dimensional scene according to the present application includes the steps of:
collecting region information and constructing a three-dimensional model;
the server receives image data sent by the remote sensing satellite, captures environment information from the image data according to the area name, extracts obstacle feature data from the environment information, and obtains the actual size of an obstacle according to the obstacle feature data.
And constructing a corresponding obstacle model according to the obstacle characteristic data, and reducing the model according to the scaling.
Wherein the scaling is set by a system default or manually according to needs;
and constructing a space rectangular coordinate system, performing data fusion on the space rectangular coordinate system and the image data, collecting coordinate point information from the barrier characteristic data, and splicing the barrier model to the space rectangular coordinate system according to the coordinate point information to produce the three-dimensional model corresponding to the area name.
Collecting all the three-dimensional models, constructing an area storage library to store the three-dimensional models, and constructing a data connection between the area storage library and a server;
wherein, constructing a data connection between the local repository and the server comprises the steps of:
and creating a database table in the regional storage base, and recording the three-dimensional model into the database table.
And creating a class model, and establishing connection between the server and the regional storage library in the initialization process of the class model.
The environment information is the building, the terrain and the vegetation coverage information of the selected area of the staff, and the information is presented in a digital form; including the coordinates of the position of the building and the digital representation of the vegetation coverage area.
And constructing a three-dimensional graph and collecting node information of the three-dimensional graph.
And the server collects the image data of each obstacle model under different viewing angles, performs Gaussian smoothing processing on the image data to filter noise interference, and then performs filtering processing on the image data after the noise interference is removed.
Comparing the image data of the same barrier model under different viewing angles to find out corner points, and constructing a corner point coordinate set of the barrier model according to the found corner points; the set of corner coordinates is calculated by the following equation (1):
T n ＝((x 1 ，y 1 )，(x 2 ，y 2 )，(x 3 ，y 3 )，…，(x n ，y n ) }; formula (1)
Wherein x is n The abscissa, y, representing the corner point n Represents the ordinate of the corner point;
when the observation visual angle of the obstacle model is changed, calculating the gray level change of pixel points in the window before and after sliding, and referring to formulas (2), (3) and (4):
and (3) carrying out diagonalization on the covariance matrix M of the formula (3), collecting two groups of characteristic values after the diagonalization, and judging the two groups of characteristic values.
The judgment of the two groups of characteristic values specifically comprises the following steps:
when both sets of feature values are large, the window contains corner points.
When the two sets of feature values are one larger and one smaller, the window contains edges.
When the characteristic values are all small, the window is in a flat area.
Measuring the response of each angular point to obtain a measurement value, comparing the measurement value with a threshold value, and marking the measurement value larger than the threshold value, wherein a pixel point corresponding to the measurement value is the angular point position coordinate in the image; the specific metric is as follows (5):
R＝detM-k(traceM) 2 formula (5)
Wherein k represents a constant and generally takes a value of 0.04-0.06;
wherein the threshold is a server default threshold or a threshold set by a worker.
And carrying out point cloud clustering processing on the three-dimensional graph, detecting image gradient information through 3DHarris, and calculating coordinate information of nodes of the three-dimensional graph to obtain a calculation result.
Comparing the calculation result with a set threshold value; and if the calculation result is greater than the set threshold, the node is a local maximum value point, the node is judged to be a key node, and otherwise, the node is not the key node.
Screening the nodes of the three-dimensional graph specifically comprises the following steps:
and training and optimizing the learning network model, and importing the collected node coordinate information of the three-dimensional graph into the learning network model.
The training optimization of the learning network model specifically comprises the following steps:
and the learning network model is in communication connection with the node standard library of the server, and node information stored in the node standard library is extracted.
Wherein, the node information is the collected coordinate point information.
And constructing a node information set according to the collected node information, selecting one node information as verification data, and repeatedly using the verification data to verify the precision of the test model.
And selecting any subset as a test set for each group of node information, then taking the rest subsets as a training set, carrying out primary prediction on each group of data, and outputting the data with the highest prediction result accuracy as the optimal parameters.
And carrying out standardization processing on the training data set according to the optimal parameters, finally conveying the training samples to the learning network model, and updating the learning network model in real time through iterative training.
And the learning network model performs normalization processing on the pixel point images corresponding to the node coordinate information of the three-dimensional graph, and simultaneously performs interval-by-interval gray level probability distribution measurement on the normalized pixel point images.
And collecting measurement results, screening out nodes lower than the preset node coordinate standard, comparing all groups of node information in the same three-dimensional graph, deleting repeated and redundant node information, and constructing a node record table to record all groups of node information.
Operating a simulation graph to perform avoidance testing, comprising:
and moving the preset path parameters of each group of three-dimensional graphs, and recording the coordinate information of each node of each group of three-dimensional graphs in real time.
And calculating the distance of the x axis, the y axis and the z axis of each group of coordinates of different three-dimensional graphs, judging that the two groups of three-dimensional graphs collide when the distance of the node coordinates of the two groups of three-dimensional graphs is less than 0, and recording the information of the two groups of node coordinates which collide.
Optimizing and updating avoidance information of a plurality of three-dimensional graphs, comprising the following steps:
and the server receives the coordinate information of the node with the collision.
And constructing a prediction model, collecting all groups of collided node coordinate information paths, capturing a genetic algorithm from the Internet by the server, updating all groups of path information, and synchronizing the updated path information into the corresponding three-dimensional graph.
In one embodiment, compared with the conventional avoidance method, the method for avoiding the graph nodes from each other in the three-dimensional scene collects image data of the barrier models in the three-dimensional models of each region under different viewing angles through the server, finds out corresponding corner points for matching, constructs a corner point coordinate set of each barrier according to each set of found corner points, detects the used image gradient information through 3DHarris, calculates each set of key node coordinate information of each set of three-dimensional graphs, simultaneously introduces each set of collected node coordinate information into the learning network model, normalizes the pixel point images corresponding to each set of node coordinate information through the learning network model, measures the probability distribution of the gray level of each region from one region to another region of the normalized pixel point images, and screens out the nodes lower than the set node coordinate standard of a worker, meanwhile, the repeated and redundant node information is deleted, a node record table is built to record each group of node information, and the node information is screened out by building a learning network model, so that the accuracy of each group of node information can be improved, the accuracy of a test result is improved, and the working quality is improved;
the method for avoiding the graph nodes from each other in the three-dimensional scene includes that a server moves each group of three-dimensional graphs according to route information set by the server, coordinate information of each node of each group of three-dimensional graphs is recorded in real time, distance calculation is carried out on an x axis, a y axis and a z axis of each group of coordinates of different three-dimensional graphs, when the distance between the node coordinates of the two groups of three-dimensional graphs is smaller than 0, collision of the two groups of three-dimensional graphs is judged, coordinate information of the two groups of nodes which collide is recorded at the same time, the server receives the coordinate information of the nodes which collide, a prediction model is constructed, paths of the node coordinate information which collide are collected, the server automatically captures related genetic algorithms from the internet, updates each group of path information, synchronizes the updated path information to the corresponding three-dimensional graphs, and can collect and record the collision node information, the manual recording of workers is not needed, the working efficiency of the workers is improved, meanwhile, the related path information is optimized, and the feasibility of the follow-up path is greatly improved.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be considered to be within the technical scope of the present invention, and the technical solutions and the inventive concepts thereof according to the present invention should be equivalent or changed within the scope of the present invention.
Claims (9)
1. A method for avoiding graph nodes from each other in a three-dimensional scene is characterized by comprising the following steps:
collecting region information and constructing a three-dimensional model;
constructing a three-dimensional graph and collecting node information of the three-dimensional graph;
screening nodes of the three-dimensional graph;
running a simulation graph to perform avoidance test;
and optimizing avoidance information of the three-dimensional graphs.
2. The method for avoiding each other of graph nodes in a three-dimensional scene according to claim 1, wherein the collecting region information and constructing the corresponding three-dimensional model specifically comprises:
the server receives image data sent by the remote sensing satellite, captures environment information from the image data according to the area name, extracts obstacle feature data from the environment information, and obtains the actual size of an obstacle according to the obstacle feature data;
constructing a corresponding obstacle model according to the obstacle feature data, and reducing the model according to the scaling;
constructing a space rectangular coordinate system, performing data fusion on the space rectangular coordinate system and the image data, collecting coordinate point information from the obstacle characteristic data, and splicing an obstacle model to the space rectangular coordinate system according to the coordinate point information to produce a three-dimensional model corresponding to the area name;
collecting all the three-dimensional models, constructing a regional storage library to store the three-dimensional models, and constructing a data connection between the regional storage library and a server.
3. The method of claim 2, wherein the step of building a data connection between the local repository and the server comprises the steps of:
creating a database table in the regional storage base, and recording the three-dimensional model into the database table;
and creating a class model, and establishing connection between the server and the regional storage library in the initialization process of the class model.
4. The method for avoiding the graph nodes from each other in the three-dimensional scene according to claim 1, wherein a three-dimensional graph is constructed, and node information of the three-dimensional graph is collected, specifically:
the server collects image data of each obstacle model under different visual angles, Gaussian smoothing processing is carried out on the image data to filter noise interference, and then filtering processing is carried out on the image data after the noise interference is removed;
comparing the image data of the same barrier model under different visual angles to find out angular points, and constructing an angular point coordinate set of the barrier model according to the found angular points; the set of corner coordinates is calculated by the following equation (1):
T n ＝{(x 1 ，y 1 )，(x 2 ，y 2 )，(x 3 ，y 3 )，…，(x n ，y n ) }; formula (1)
Wherein x is n The abscissa, y, representing the corner point n Represents the ordinate of the corner point;
s2.3, when the observation visual angle of the obstacle model is changed, calculating the gray level change of pixel points in the window before and after sliding, and referring to formulas (2), (3) and (4):
E(u,v)＝∑w(x n ,y n )[I(x n +u,x n +v)-I(x n ,x n )] 2 formula (2)
Diagonalizing the covariance matrix M of the formula (3), collecting two groups of characteristic values after diagonalization, and judging the two groups of characteristic values;
measuring the response of each angular point to obtain a measurement value, comparing the measurement value with a threshold value, and marking the measurement value larger than the threshold value, wherein a pixel point corresponding to the measurement value is the angular point position coordinate in the image; the specific metric is as follows (5):
R＝detM-k(traceM) 2 formula (5)
Wherein k represents a constant and generally takes a value of 0.04-0.06;
carrying out point cloud clustering processing on the three-dimensional graph, detecting image gradient information through 3DHarris, and calculating coordinate information of nodes of the three-dimensional graph to obtain a calculation result;
comparing the calculation result with a set threshold value; and if the calculation result is greater than the set threshold, the node is a local maximum value point, the node is judged to be a key node, and otherwise, the node is not the key node.
5. The method for avoiding each other of the graph nodes in the three-dimensional scene according to claim 4, wherein the judging of the two groups of feature values specifically comprises:
when the two groups of characteristic values are both large, namely the window contains angular points;
when one of the two groups of characteristic values is larger and the other is smaller, the window contains edges;
when the characteristic values are all small, the window is in a flat area.
6. The method for avoiding each other of the graph nodes in the three-dimensional scene according to claim 1, wherein the screening of the nodes of the three-dimensional graph specifically comprises:
training and optimizing a learning network model, and importing the collected node coordinate information of the three-dimensional graph into the learning network model;
the learning network model normalizes the pixel point images corresponding to the node coordinate information of the three-dimensional graph, and measures the probability distribution of the gray level of each interval of the normalized pixel point images;
and collecting measurement results, screening out nodes lower than the preset node coordinate standard, comparing all groups of node information in the same three-dimensional graph, deleting repeated and redundant node information, and constructing a node record table to record all groups of node information.
7. The method for avoiding the graph nodes from each other in the three-dimensional scene according to claim 5, wherein the training optimization is performed on the learning network model, specifically:
the learning network model is in communication connection with a node standard library of a server, and node information stored in the node standard library is extracted;
wherein, the node information is the collected coordinate point information;
constructing a node information set according to the collected node information, selecting one node information as verification data, and repeatedly using the verification data to verify the precision of the test model;
selecting any subset as a test set for each group of node information, then taking the rest subsets as a training set, carrying out primary prediction on each group of data, and outputting the data with the highest prediction result accuracy as an optimal parameter;
and carrying out standardization processing on the training data set according to the optimal parameters, finally conveying the training samples to the learning network model, and updating the learning network model in real time through iterative training.
8. The method for avoiding the graph nodes from each other in the three-dimensional scene according to claim 1, wherein the avoiding measurement of the plurality of three-dimensional graphs specifically comprises:
moving the preset path parameters of each group of three-dimensional graphs, and recording the coordinate information of each node of each group of three-dimensional graphs in real time;
and calculating the distance of the x axis, the y axis and the z axis of each group of coordinates of different three-dimensional graphs, judging that the two groups of three-dimensional graphs collide when the distance of the node coordinates of the two groups of three-dimensional graphs is less than 0, and recording the information of the two groups of node coordinates which collide.
9. The method for avoiding the graph nodes from each other in the three-dimensional scene according to claim 1, wherein the updating of the avoidance information of the plurality of three-dimensional graphs according to the analysis result specifically includes:
the server receives the coordinate information of the collided nodes;
and constructing a prediction model, collecting all groups of collided node coordinate information paths, capturing a genetic algorithm from the Internet by the server, updating all groups of path information, and synchronizing the updated path information into the corresponding three-dimensional graph.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210391939.XA CN114817014B (en) | 2022-04-14 | 2022-04-14 | Method for avoiding graph nodes from each other in three-dimensional scene |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210391939.XA CN114817014B (en) | 2022-04-14 | 2022-04-14 | Method for avoiding graph nodes from each other in three-dimensional scene |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114817014A true CN114817014A (en) | 2022-07-29 |
| CN114817014B CN114817014B (en) | 2024-10-15 |
Family
ID=82535695
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210391939.XA Active CN114817014B (en) | 2022-04-14 | 2022-04-14 | Method for avoiding graph nodes from each other in three-dimensional scene |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114817014B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20160071870A (en) * | 2014-12-12 | 2016-06-22 | 삼성전자주식회사 | Apparatus and method for providing 3d graphic contents |
| US20180308281A1 (en) * | 2016-04-01 | 2018-10-25 | draw, Inc. | 3-d graphic generation, artificial intelligence verification and learning system, program, and method |
| CN113031597A (en) * | 2021-03-02 | 2021-06-25 | 南京理工大学 | Autonomous obstacle avoidance method based on deep learning and stereoscopic vision |
| CN113777622A (en) * | 2021-08-31 | 2021-12-10 | 通号城市轨道交通技术有限公司 | Method and device for identifying rail obstacle |
| CN113962019A (en) * | 2021-09-23 | 2022-01-21 | 海南大学 | Intelligent driving automobile safety protection system based on virtual reality technology |
-
2022
- 2022-04-14 CN CN202210391939.XA patent/CN114817014B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20160071870A (en) * | 2014-12-12 | 2016-06-22 | 삼성전자주식회사 | Apparatus and method for providing 3d graphic contents |
| US20180308281A1 (en) * | 2016-04-01 | 2018-10-25 | draw, Inc. | 3-d graphic generation, artificial intelligence verification and learning system, program, and method |
| CN113031597A (en) * | 2021-03-02 | 2021-06-25 | 南京理工大学 | Autonomous obstacle avoidance method based on deep learning and stereoscopic vision |
| CN113777622A (en) * | 2021-08-31 | 2021-12-10 | 通号城市轨道交通技术有限公司 | Method and device for identifying rail obstacle |
| CN113962019A (en) * | 2021-09-23 | 2022-01-21 | 海南大学 | Intelligent driving automobile safety protection system based on virtual reality technology |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114817014B (en) | 2024-10-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110246112B (en) | Laser scanning SLAM indoor three-dimensional point cloud quality evaluation method based on deep learning | |
| CN111105495A (en) | Laser radar mapping method and system fusing visual semantic information | |
| CN110135282B (en) | Examinee return plagiarism cheating detection method based on deep convolutional neural network model | |
| KR20060097074A (en) | Apparatus and method for generating shape model of object and automatic search for feature point of object using same | |
| CN115760922B (en) | Target tracking method, device, system, terminal equipment and storage medium | |
| CN116363319B (en) | A modeling method, modeling device, equipment and medium of a building roof | |
| CN116152863B (en) | Person information identification method, device, electronic equipment and storage medium | |
| CN112232311B (en) | Face tracking method and device and electronic equipment | |
| CN118675106A (en) | Real-time monitoring method, system, device and storage medium for falling rocks based on machine vision | |
| CN115937680A (en) | A geospatial positioning method for Wailongwu buildings based on deep neural network | |
| CN113516771B (en) | A method for extracting building change features based on real-life 3D models | |
| US20050185834A1 (en) | Method and apparatus for scene learning and three-dimensional tracking using stereo video cameras | |
| CN111860097B (en) | Abnormal behavior detection method based on fuzzy theory | |
| CN116958477A (en) | A three-dimensional modeling method and system for building structures based on construction drawings | |
| CN108992033B (en) | Grading device, equipment and storage medium for vision test | |
| CN114817014B (en) | Method for avoiding graph nodes from each other in three-dimensional scene | |
| CN118644667B (en) | Method and device for detecting surface defects of hub based on deep learning | |
| CN117197730B (en) | Repair evaluation method for urban space distortion image | |
| CN115578455B (en) | Positioning method for reserved hole in concrete structure room | |
| CN119600471A (en) | A target detection method and system integrating DOM and DSM data | |
| CN119863504B (en) | A method and system for calculating the effective area of building windows based on BIM model | |
| CN114092762A (en) | Method, device and electronic device for generating training samples for remote sensing image classification | |
| CN120612644B (en) | Financial deposit investigation method, device, equipment and medium based on computer vision | |
| CN117827929B (en) | Tissue member mining method, medium and device based on tissue characteristic attribute self-adaptive update | |
| CN119850801B (en) | A virtual character construction system and method based on multimodal information 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |