CN116539167B - High-voltage power supply working temperature distribution data analysis method - Google Patents
High-voltage power supply working temperature distribution data analysis method Download PDFInfo
- Publication number
- CN116539167B CN116539167B CN202310806240.XA CN202310806240A CN116539167B CN 116539167 B CN116539167 B CN 116539167B CN 202310806240 A CN202310806240 A CN 202310806240A CN 116539167 B CN116539167 B CN 116539167B
- Authority
- CN
- China
- Prior art keywords
- node
- embedded vector
- categories
- power supply
- voltage power
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 50
- 238000007405 data analysis Methods 0.000 title claims abstract description 13
- 239000013598 vector Substances 0.000 claims abstract description 235
- 238000010586 diagram Methods 0.000 claims abstract description 77
- 230000008859 change Effects 0.000 claims abstract description 34
- 238000007621 cluster analysis Methods 0.000 claims abstract description 14
- 238000004364 calculation method Methods 0.000 claims description 20
- 238000004422 calculation algorithm Methods 0.000 claims description 14
- 238000003064 k means clustering Methods 0.000 claims description 7
- 230000014759 maintenance of location Effects 0.000 claims description 5
- 230000004069 differentiation Effects 0.000 description 6
- 230000002159 abnormal effect Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 230000011218 segmentation Effects 0.000 description 4
- 230000009467 reduction Effects 0.000 description 3
- 230000002776 aggregation Effects 0.000 description 2
- 238000004220 aggregation Methods 0.000 description 2
- 235000002020 sage Nutrition 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 230000004931 aggregating effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000017525 heat dissipation Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
- G01J5/00—Radiation pyrometry, e.g. infrared or optical thermometry
- G01J5/48—Thermography; Techniques using wholly visual means
- G01J5/485—Temperature profile
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
- G06T17/20—Finite element generation, e.g. wire-frame surface description, tesselation
-
- 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
- G06V10/762—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using clustering, e.g. of similar faces in social networks
-
- 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
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- 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
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Computer Graphics (AREA)
- Geometry (AREA)
- Image Analysis (AREA)
Abstract
The application relates to the technical field of image processing, in particular to a high-voltage power supply working temperature distribution data analysis method, which comprises the following steps: at least two embedded vector diagram structures are constructed based on pixel point change sequences corresponding to power infrared thermal image pixels, differences of every two node categories in the same embedded vector diagram structure and consistency of every two node categories in the embedded vector diagram structure are calculated respectively according to at least two node categories in the embedded vector diagram structures, distinguishing degrees corresponding to the embedded vector diagram structures are calculated based on the differences and the consistency of every two node categories, and finally the number of node categories corresponding to the embedded vector diagram structure with the highest distinguishing degree is used as classification parameters of cluster analysis so as to classify the working temperature of a high-voltage power supply, so that the accuracy of temperature category division of the working temperature of the high-voltage power supply can be improved, and the working cost is further reduced.
Description
Technical Field
The application relates to the technical field of image processing, in particular to a high-voltage power supply working temperature distribution data analysis method.
Background
When the high-voltage power supply works, temperature change usually occurs, because energy conversion occurs, a part of energy in other forms is converted into heat energy, and the heat energy is not emitted, namely, a temporary large amount of accumulation phenomenon occurs. And the identification of the working temperature of the high-voltage power supply is particularly important.
In the existing method, when abnormal data in the working temperature of the high-voltage power supply is identified, the temperature category in the normal state is usually obtained through prior knowledge calculation, and then the temperature category is compared with the actual temperature category, wherein when the actual temperature category is calculated, the pixel points with similar temperature are often divided into one category through a clustering method, and the number of categories obtained by dividing the pixel points with similar temperature into one category through the clustering method is easy to generate errors, so that the abnormal region cannot be accurately identified. That is, the conventional clustering method is not accurate enough for classifying the temperature types of the working temperatures of the high-voltage power supply, so that the abnormal data identification of the working temperatures of the high-voltage power supply is caused to have errors, and the working cost is high.
Disclosure of Invention
In view of the foregoing, it is necessary to provide a method for analyzing data of operating temperature distribution of a high-voltage power supply, which can improve accuracy of temperature classification of operating temperature of the high-voltage power supply, further improve accuracy of data analysis, and reduce operating cost.
The first aspect of the application provides a high-voltage power supply working temperature distribution data analysis method, which is applied to the field of high-voltage power supply working temperature data processing, and comprises the following steps: constructing at least two embedded vector graph structures based on a pixel point change sequence corresponding to the power infrared thermal image pixels, wherein two adjacent embedded vector graph structures correspond to each other, and each embedded vector graph structure comprises at least two node categories; according to at least two node categories in the embedded vector diagram structure, respectively calculating the difference of every two node categories in the same embedded vector diagram structure and the consistency in the mutually corresponding embedded vector diagram structures; based on the difference and consistency of the two-node categories, calculating the distinction degree corresponding to the embedded vector diagram structure; and taking the node category number corresponding to the embedded vector diagram structure with the highest distinction as the classification parameter of cluster analysis so as to classify the working temperature of the high-voltage power supply.
In one embodiment, the constructing at least two embedded vector graph structures based on the pixel point change sequence corresponding to the power infrared thermal image pixels, where two adjacent embedded vector graph structures correspond to each other, and each embedded vector graph structure includes at least two node categories, specifically includes: step 1, constructing a triangle mesh map structure based on a pixel point change sequence corresponding to a power infrared thermal image pixel; step 2, calculating an embedded vector corresponding to each node based on the triangle mesh structure and a preset vector calculation method; step 3, taking the embedded vector corresponding to each node as a node value, taking cosine similarity of the embedded vectors corresponding to every two nodes as a side value, and taking each pixel point as a node to construct an embedded vector diagram structure; and 4, carrying out iterative computation for preset times on the embedded vector diagram structure instead of the triangular mesh diagram structure in the step 2, and obtaining at least two embedded vector diagram structures.
In one embodiment, the constructing a triangle mesh structure based on the pixel point change sequence corresponding to the power infrared thermal image pixel specifically includes: and acquiring sequences corresponding to pixel values of each pixel point of the power infrared thermal image at different moments, confirming a pixel point change sequence corresponding to each pixel point, taking each pixel point as a node, taking the pixel point change sequence of each pixel point as a node value, and taking cosine similarity of the pixel point change sequences corresponding to every two nodes as a side value, so as to construct the triangular network graph structure.
In one embodiment, the calculating the difference between every two node categories in the same embedded vector graph structure according to at least two node categories in the embedded vector graph structure specifically includes: based on the class center embedded vector corresponding to each node class, cosine similarity between each pixel point of each two node classes and the corresponding class center embedded vector is calculated respectively; determining the pixel points with the cosine similarity larger than a preset threshold value as the corresponding retention points of each node category; matching the reserved points corresponding to the two node categories according to a preset rule, and confirming the matched pair; based on the matching pair, confirming the corresponding difference weight of the two node categories and the similarity parameter of the matching pair; and calculating the difference of the two-node categories in the same embedded vector diagram structure based on the difference weight corresponding to the two-node categories and the matching pair similarity parameter.
In one embodiment, the calculating the consistency of every two node categories in the embedded vector graph structure according to at least two node categories in the embedded vector graph structure specifically includes: calculating the ratio of each pixel point in each two node categories in the corresponding node category of the adjacent embedded vector graph structure so as to divide the corresponding node category of the adjacent embedded vector graph structure into a complete sub-category and a scattered sub-category; and calculating the consistency of the two-node categories in the mutually corresponding embedded vector graph structure based on the corresponding ratio of the scattered sub-categories of the two-node categories and the difference between the two-node categories and the corresponding node categories of the adjacent embedded vector graph structure.
In one embodiment, the calculating the degree of distinction corresponding to the embedded vector graph structure based on the difference and the consistency of the two-node categories specifically includes: inputting the difference and the consistency of the two-node categories into a difference value calculation formula, and calculating a difference value corresponding to the two-node categories; and taking the average value corresponding to the distinguishing value corresponding to all the node categories in the embedded vector diagram structure as the distinguishing degree corresponding to the embedded vector diagram structure.
In one embodiment, the calculating the difference value corresponding to the two-node class by inputting the difference and the consistency of the two-node class into a difference value calculation formula specifically includes:
wherein ,refers to the differential weight corresponding to every two node categories, ++>Means matching pair similarity parameters corresponding to every two node categories, < ->Refers toQuantity of combination->Refers to the ratio of scattered sub-categories of one node category in the corresponding node category, +.>Refers to the +.>Difference between two categories in the two combinations, +.>Refers to the ratio of scattered sub-categories of another node category in the corresponding node category, +.>Refers to the +.f in the corresponding node class of another node class>Differences in two categories in the individual combinations.
In one embodiment, the classifying parameters of the cluster analysis, which are the number of node categories corresponding to the highest-distinction embedded vector graph structure, are used for classifying the working temperature of the high-voltage power supply, and specifically include: counting the corresponding distinction between at least two embedded vector graph structures, and confirming the node category number corresponding to the embedded vector graph structure with the highest distinction; and taking the K value as a parameter of a K-means clustering algorithm to divide the infrared thermal image into K temperature categories.
According to the embodiment of the application, at least two embedded vector graph structures are constructed based on a pixel point change sequence corresponding to a power infrared thermal image pixel, wherein two adjacent embedded vector graph structures are mutually corresponding, each embedded vector graph structure comprises at least two node categories, then the difference of every two node categories in the same embedded vector graph structure and the consistency of every two node categories in the mutually corresponding embedded vector graph structure are respectively calculated according to the at least two node categories in the embedded vector graph structure, the differentiation degree corresponding to the embedded vector graph structure is calculated based on the difference and the consistency of every two node categories, and finally the node category number corresponding to the embedded vector graph structure with the highest differentiation degree is used as a classification parameter of cluster analysis so as to classify the working temperature of the high-voltage power supply. The node class number corresponding to the embedded vector diagram structure with the highest degree of distinction is selected as the classification parameter of the cluster analysis, and then the classification of the working temperature of the high-voltage power supply is carried out, so that the accuracy of the temperature class division of the working temperature of the high-voltage power supply can be improved, and the working cost is further reduced.
Drawings
Fig. 1 is a flow chart of a method for analyzing data of high-voltage power supply operating temperature distribution according to an embodiment of the present application.
Fig. 2 is a schematic diagram of a first sub-flow of a high-voltage power supply operating temperature distribution data analysis method according to an embodiment of the present application.
Fig. 3 is a schematic diagram of a second sub-flow of the high-voltage power supply operating temperature distribution data analysis method according to an embodiment of the present application.
Fig. 4 is a schematic diagram of a third sub-flow of the high-voltage power supply operating temperature distribution data analysis method according to an embodiment of the present application.
Fig. 5 is a schematic diagram of a fourth sub-flowchart of a method for analyzing high-voltage power supply operating temperature distribution data according to an embodiment of the present application.
Fig. 6 is a schematic diagram of a fifth sub-flowchart of a method for analyzing high-voltage power supply operating temperature distribution data according to an embodiment of the present application.
Fig. 7 is a schematic diagram of a sixth sub-flowchart of a method for analyzing high-voltage power supply operating temperature distribution data according to an embodiment of the present application.
Detailed Description
In describing embodiments of the present application, words such as "exemplary," "or," "such as," and the like are used to mean serving as an example, instance, or illustration. Any embodiment or design described herein as "exemplary" or "for example" should not be construed as preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary," "or," "such as," and the like are intended to present related concepts in a concrete fashion.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the description of the application herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. It is to be understood that, unless otherwise indicated, a "/" means or. For example, A/B may represent A or B. The "and/or" in the present application is merely one association relationship describing the association object, indicating that three relationships may exist. For example, a and/or B may represent: a exists alone, A and B exist simultaneously, and B exists alone. "at least one" means one or more. "plurality" means two or more than two. For example, at least one of a, b or c may represent: seven cases of a, b, c, a and b, a and c, b and c, a, b and c.
It should be further noted that the terms "first" and "second" in the description and claims of the present application and the accompanying drawings are used for respectively similar objects, and are not used for describing a specific order or sequence. The method disclosed in the embodiments of the present application or the method shown in the flowchart, including one or more steps for implementing the method, may be performed in an order that the steps may be interchanged with one another, and some steps may be deleted without departing from the scope of the claims.
The embodiment of the application firstly provides a high-voltage power supply working temperature distribution data analysis method which is applied to the field of high-voltage power supply working temperature data processing. Referring to fig. 1, a method for analyzing data of operating temperature distribution of a high-voltage power supply includes:
s101, constructing at least two embedded vector graph structures based on a pixel point change sequence corresponding to a high-voltage power supply infrared thermal image pixel, wherein two adjacent embedded vector graph structures correspond to each other, and each embedded vector graph structure comprises at least two node categories.
The high-voltage power supply infrared thermal image is obtained by shooting the working state of the high-voltage power supply through the infrared thermal image acquisition equipment. It should be noted that, the high-voltage power supply also emits heat during normal operation, and when the voltage is converted, not only energy is lost, but also heat energy is generated, and these all cause the power supply to generate heat. When the high-voltage power supply has abnormal conditions, the heat emitted can be increased along with different abnormal conditions, such as overload, poor heat dissipation and the like, so that the temperature of the high-voltage power supply can be increased, and the heat emitted by the high-voltage power supply can be represented by the high-voltage power supply infrared thermal image so as to be used for subsequent data analysis and processing.
The pixel point change sequence refers to a pixel value sequence formed by pixel values of pixels corresponding to infrared thermal image pixels of the high-voltage power supply at different moments according to a time sequence, and the pixel point change sequence can further reflect temperature change of the high-voltage power supply. The embedded vector diagram structure refers to a diagram structure derived from embedded vectors corresponding to pixel point change sequences corresponding to high-voltage power supply infrared thermal image pixels, and can be generated based on a triangulation method.
It should be noted that, the mutual correspondence of two adjacent embedded vector graph structures means that there is a correspondence relationship between the embedded vector graph structures, and may be that the following embedded vector graph structure is iteratively generated according to the preceding embedded vector graph structure, for example, the second embedded vector graph structure is further generated by the first embedded vector graph structure, the third embedded vector graph structure is further generated by the second embedded vector graph structure, and so on to iteratively generate at least two embedded vector graph structures.
Specifically, the node categories refer to classification categories of pixel points, each embedded vector graph structure includes at least two node categories, namely at least two node categories exist through each embedded vector graph structure, and the at least two node categories can be node categories corresponding to the embedded vector graph structure obtained by performing preset dimension reduction and clustering on the embedded vector graph structure.
S102, according to at least two node categories in the embedded vector diagram structure, respectively calculating the difference of every two node categories in the same embedded vector diagram structure and the consistency in the mutually corresponding embedded vector diagram structures.
The difference of the two node categories in the same embedded vector diagram structure refers to the degree of difference between two target node categories selected in the same embedded vector diagram structure, and the greater the difference between the node categories is, the more obvious the classification effect is proved, and the higher the typing accuracy is. The consistency of the two-node type in the mutually corresponding embedded vector graph structure refers to the coincidence ratio of the pixel points of the two-node type in the mutually corresponding embedded vector graph structure, for example, the coincidence ratio of the pixel points of the A-node type and the pixel points of the A-node type in the mutually corresponding embedded vector graph structure, namely the consistency in the mutually corresponding embedded vector graph structure is obtained by iterative calculation, the mutually corresponding embedded vector graph structure with the A-node type can be the last embedded vector graph structure, the corresponding embedded vector graph structure with the current A-node type is the A-node type, and the coincidence ratio of the pixel points of the A-node type and the pixel points of the A-node type is the consistency in the mutually corresponding embedded vector graph structure.
It should be noted that, the differences of the node classes in the same embedded vector graph structure and the consistency of the node classes in the embedded vector graph structures corresponding to each other are respectively calculated, and the purpose is to prepare for subsequent determination of the embedded vector graph structure with good classification effect by taking the differences and the consistency as calculation parameters so as to determine the optimal classification quantity.
And S103, calculating the distinction degree corresponding to the embedded vector diagram structure based on the difference and the consistency of the two-node categories.
The degree of distinction corresponding to the embedded vector graph structure refers to a parameter obtained by calculation according to the difference and consistency between node categories in the embedded vector graph structure and representing the classification effect in the embedded vector graph structure. In other words, when the degree of distinction corresponding to the embedded vector diagram structure is higher, the classification effect corresponding to the current embedded vector diagram structure is better, that is, the number of classification categories is more appropriate.
And S104, taking the node category number corresponding to the highest-distinction embedded vector diagram structure as a classification parameter of cluster analysis so as to classify the working temperature of the high-voltage power supply.
After the discrimination degrees corresponding to all the embedded vector graph structures are obtained, sorting the discrimination degrees corresponding to all the embedded vector graph structures to confirm the embedded vector graph structure with the highest discrimination degree, and further confirm the node category number corresponding to the embedded vector graph structure with the highest discrimination degree, so that the node category number corresponding to the embedded vector graph structure with the highest discrimination degree is used as the classification number of the subsequent cluster analysis aiming at the working temperature of the high-voltage power supply.
Specifically, referring to fig. 2, the number of node categories corresponding to the highest-distinction embedded vector graph structure is used as a classification parameter of cluster analysis to classify the working temperature of the high-voltage power supply, and specifically includes:
s201, counting the corresponding distinction between at least two embedded vector graph structures, and confirming the node category number corresponding to the embedded vector graph structure with the highest distinction;
s202, the K value of the number of node categories corresponding to the highest-distinction embedded vector diagram structure is used as a parameter of a K-means clustering algorithm, so that the infrared thermal image is divided into K temperature categories.
After counting the corresponding discrimination degrees between at least two embedded vector graph structures, sorting the discrimination degrees corresponding to all the embedded vector graph structures to confirm the node category number corresponding to the embedded vector graph structure with the highest discrimination degree, and further confirm the node category number corresponding to the embedded vector graph structure with the highest discrimination degree, wherein the node category number corresponding to the embedded vector graph structure with the highest discrimination degree is used as the K value of a subsequent K-means clustering algorithm, and the K value is the number of the subsequent classification categories.
It should be noted that the K-means clustering algorithm (K-means clustering algorithm) is an iterative solution clustering analysis algorithm, and includes the steps of selecting K objects randomly as initial cluster centers if data is to be classified into K groups, then calculating the distance between each object and each seed cluster center, and assigning each object to the cluster center closest to the object. The cluster centers and the objects assigned to them represent a cluster. For each sample assigned, the cluster center of the cluster is recalculated based on the existing objects in the cluster. This process will repeat until a certain termination condition is met. The termination condition may be that no (or a minimum number of) objects are reassigned to different clusters, no (or a minimum number of) cluster centers are changed again, and the sum of squares of errors is locally minimum. With specific reference to the prior art, the present disclosure is not further limited.
According to the embodiment of the application, at least two embedded vector graph structures are constructed based on a pixel point change sequence corresponding to a power infrared thermal image pixel, wherein two adjacent embedded vector graph structures are mutually corresponding, each embedded vector graph structure comprises at least two node categories, then the difference of every two node categories in the same embedded vector graph structure and the consistency of every two node categories in the mutually corresponding embedded vector graph structure are respectively calculated according to the at least two node categories in the embedded vector graph structure, the differentiation degree corresponding to the embedded vector graph structure is calculated based on the difference and the consistency of every two node categories, and finally the node category number corresponding to the embedded vector graph structure with the highest differentiation degree is used as a classification parameter of cluster analysis so as to classify the working temperature of the high-voltage power supply. The node class number corresponding to the embedded vector diagram structure with the highest degree of distinction is selected as the classification parameter of the cluster analysis, and then the classification of the working temperature of the high-voltage power supply is carried out, so that the accuracy of the temperature class division of the working temperature of the high-voltage power supply can be improved, and the working cost is further reduced.
In one embodiment of the present application, and referring to fig. 3, the step S101: the method comprises the steps of constructing at least two embedded vector diagram structures based on a pixel point change sequence corresponding to a high-voltage power supply infrared thermal image pixel, wherein two adjacent embedded vector diagram structures correspond to each other, each embedded vector diagram structure comprises at least two node categories, and the method specifically comprises the following steps:
s301, step 1: and constructing a triangular mesh map structure based on a pixel point change sequence corresponding to the high-voltage power supply infrared thermal image pixels.
The triangulation network graph structure is a graph structure obtained through a triangulation method, and the triangulation method is preferably a Delaunay triangulation method. It should be noted that triangulation: let V be a finite set of points in the two-dimensional real number domain, edge E be a closed line segment composed of points in the set of points as endpoints, and E be a set of E. Then a triangulation t= (V, E) of the point set V is a plan G that satisfies the condition: 1. edges in the plan view do not contain any points in the point set other than the end points. 2. There are no intersecting edges. 3. All the faces in the plan view are triangular faces, and the aggregate of all the triangular faces is the convex hull of the scattered point set V. The most widely used triangulation in practice is Delaunay triangulation, which is a special triangulation. Starting from the Delaunay side: delaunay edges of the Delaunay triangulation algorithm: assuming one edge E in E (two endpoints are a, b), E is called Delaunay edge if the following condition is satisfied: there is a circle passing through the points a and b, and the circle (note that the circle is the circle, and the circle is the same with the three points at most) does not contain any other points in the point set V, which is also called an empty circle characteristic. If one triangulation T of the point set V contains only Delaunay edges, then the triangulation is called Delaunay triangulation.
Specifically, referring to fig. 4, the constructing a triangle mesh structure based on the pixel point change sequence corresponding to the high-voltage power supply infrared thermal image pixel specifically includes:
s401, acquiring sequences corresponding to pixel values of each pixel point of the power infrared thermal image at different moments, and confirming a pixel point change sequence corresponding to each pixel point;
s402, taking each pixel point as a node, taking a pixel point change sequence of each pixel point as a node value, and taking cosine similarity of pixel point change sequences corresponding to every two nodes as a side value, so as to construct the triangle network diagram structure.
The infrared thermal image at each moment can be obtained, the pixel values at different moments of the same pixel point are formed into a sequence according to a time sequence, which is called a pixel point change sequence, and then a triangular network graph structure is obtained for all the pixel points on the image through a Delaunay triangulation method, namely, nodes are each pixel point, node values are the change sequence of each pixel point, and side values are cosine similarity of the change sequences of two corresponding nodes.
S302, step 2: and calculating an embedded vector corresponding to each node based on the triangle mesh structure and a preset vector calculation method.
After the triangle graph structure is obtained, the embedded vector corresponding to each node can be correspondingly calculated according to a preset vector calculation method, in this embodiment, the preset vector calculation method may be a graph SAGE algorithm, the graph SAGE algorithm (Graph SAmple and aaareGatE) belongs to an Instructivesning algorithm, an aggregation function is learned, and the ebedding expression of the target node is learned by aggregating the characteristic information of the neighbors of the nodes. From its name, it can be seen that the core steps of the algorithm are neighbor sampling and feature aggregation, respectively, and graphSage is a machine learning task in our general sense and has generalization capability for unknown nodes.
It should be noted that, by the triangle network graph structure and the preset vector calculation method, an embedded vector corresponding to each node is calculated, where the embedded vector corresponding to each node includes information of the node and a neighborhood node.
S303, step 3: and taking the embedded vector corresponding to each node as a node value, taking cosine similarity of the embedded vectors corresponding to every two nodes as a side value, and taking each pixel point as a node to construct an embedded vector diagram structure.
When the embedded vector corresponding to each node is obtained, the embedded vector corresponding to each node is used as a node value, the pixel point is maintained as the node, the connection relation of each side is unchanged, and meanwhile, the cosine similarity of the embedded vectors corresponding to every two nodes is used as a side value, so that the embedded vector graph structure is constructed.
It should be noted that, after the embedded vector graph structure is obtained, the corresponding node category may be obtained through the embedded vector graph structure, which specifically may be: and firstly converting each node value into a scalar by using an LDA dimension reduction method for the nodes of the embedded vector graph structure, and then clustering the nodes by using an otsu multi-threshold segmentation method to obtain the node category corresponding to the current embedded vector graph structure. The LDA dimension reduction refers to projecting the nodes of the embedded vector graph structure into a space with lower dimension, so that the projected points can form the situation of distinguishing one cluster by one cluster according to the category, the points of the same category are closer to the space after projection, and the distances of the points of different categories are farther. The otsu multi-threshold segmentation method is a maximum inter-class variance method (oxford algorithm), which is an algorithm for determining a threshold value, and is called as the maximum inter-class variance method, because the image fixed threshold value binarization is performed by using the threshold value, and the inter-class variance is maximum, which is to divide the image into a background part and a foreground part according to the gray characteristic of the image, so that the segmentation with the maximum inter-class variance means that the error segmentation probability is minimum.
S304, step 4: and (3) carrying out iterative computation for preset times on the embedded vector diagram structure instead of the triangular mesh diagram structure in the step (2) to obtain at least two embedded vector diagram structures.
It should be noted that, the number of the embedded vector graph structures may be set according to actual needs, that is, the iterative computation of the preset times is set, after the embedded vector graph structures are obtained, the embedded vector graph structures are used as input, the step 2 is returned, the embedded vector graph structures replace the triangular network graph structures in the step 2, the second embedded vector computation is performed on each node to obtain new embedded vectors, for example, the first embedded vector graph structures are subjected to the computation of the embedded vector of each node by the graph method to obtain the second embedded vector corresponding to each node, at this time, the second embedded vector of each node contains node information in a larger range than the first embedded vector, then the step 3 is used for building the embedded vector graph structures to obtain second embedded vector graph structures, then the second embedded vector graph structures are used as input to replace the first embedded vector graph structures in the step 2, and then the third embedded vector graph structures are obtained, and the iterative computation of the preset times is performed according to the analogy. In addition, a plurality of embedded vector graph structures are obtained through calculation, the number of nodes in each embedded vector graph structure is the same, but the range of neighborhood information synthesized by each node is different, and the node category in each embedded vector graph structure is also different.
In this embodiment, a plurality of embedded vector graph structures are further constructed through the pixel point change sequence of each pixel point, so that subsequent classification number calculation is performed according to the plurality of embedded vector graph structures, accuracy of temperature class division of the working temperature of the high-voltage power supply is improved, accuracy of data analysis is further improved, and working cost is reduced.
In one embodiment of the present application, and referring to fig. 5, S102: according to at least two node categories in the embedded vector graph structure, the consistency of every two node categories in the embedded vector graph structure corresponding to each other is calculated, and the method specifically comprises the following steps:
s501, based on the class center embedded vector corresponding to each node class, the cosine similarity between each pixel point of each node class and the corresponding class center embedded vector is calculated.
The class center embedded vector is an embedded vector corresponding to a class center, the class center is a cosine similarity sum of the embedded vector of each node in the class and the embedded vectors of all other points in the class, and a point corresponding to the maximum sum value is taken as the class center. After the class center embedded vectors of the two-node classes are respectively determined, the cosine similarity between each pixel point of the two-node classes and the corresponding class center embedded vector is calculated.
S502, determining the pixel points with cosine similarity larger than a preset threshold as the corresponding retention points of each node category.
After obtaining cosine similarity of each pixel point of every two node categories and the corresponding category center embedded vector, comparing each cosine similarity with a preset threshold value one by one, and determining the pixel point corresponding to the cosine similarity larger than the preset threshold value as a retention point corresponding to each node category. The preferred preset threshold is 0.5, i.e., a node with a similarity greater than 0.5 is reserved as a reserved point in the class.
And S503, matching the reserved points corresponding to the two-node categories according to a preset rule, and confirming the matched pair.
The number of the reserved points corresponding to the two node categories may be different, and the matching pairs are matched in a one-to-one correspondence, so that when the number of the reserved points corresponding to the two node categories is different, the reserved points are screened out in the node categories with a large number, the reserved points with the same number as those in the node categories with a small number are screened out, and the screening rule may be to screen out the reserved points with the front cosine similarity in the node categories with a large number until the reserved points with the same number as those in the node categories with a small number are screened out. For example, there are 30 points in class a, 25 points in class B, and finally 25 points in KM matching, 25 points in both classes, and the 25 retention points with the greatest weight in class a are selected to participate in matching.
And matching the reserved points corresponding to the two node categories according to a preset rule, and confirming a matching pair, wherein the specific method can be that after calculating reserved points in the two target categories, KM matching of the reserved points in the two target categories is calculated, the reserved point of one category is used as a left node, the reserved point of the other category is used as a right node, each node on the left side is connected with all nodes on the right side, the edge value is the cosine similarity of embedded vectors of the two corresponding nodes, and a matching result is obtained through a maximum matching principle, namely one-to-one matching of the left node and the right node is obtained, so that the matching pair is obtained.
S504, confirming the difference weight corresponding to the two-node category and the similarity parameter of the matching pair based on the matching pair.
The difference weight corresponding to the two-node class may be a difference weight average value of all matching pairs of the two-node class, and the matching pair similarity parameter may be a cosine similarity average value of the matching pairs.
S505, calculating the difference of the two-node categories in the same embedded vector diagram structure based on the difference weight corresponding to the two-node categories and the matching pair similarity parameter.
After the differential weight and the matching pair similarity parameter corresponding to the two-node category are obtained, the product calculation is carried out on the differential weight and the matching pair similarity parameter corresponding to the two-node category, and the difference of the two-node category in the same embedded vector diagram structure is calculated.
In one embodiment of the present application, referring to fig. 6, the step S102 is: according to at least two node categories in the embedded vector diagram structure, the consistency of every two node categories in the embedded vector diagram structure corresponding to each other is calculated, and the method specifically comprises the following steps:
s601, calculating the ratio of each pixel point in every two node categories in the corresponding node category of the adjacent embedded vector diagram structure so as to divide the corresponding node category of the adjacent embedded vector diagram structure into a complete sub-category and a scattered sub-category.
The adjacent embedded vector graph structure can be the last embedded vector graph structure of the current embedded vector graph structure, the corresponding node class of the node in the two-by-two node class in the last embedded vector graph structure is called a corresponding node class, the ratio of the node belonging to the corresponding node class in each node class to the total number of nodes in the corresponding node class is calculated, the corresponding node class is divided according to the ratio, the corresponding node class with the ratio larger than 0.7 is called a complete sub-class of the current node class, and the corresponding node class with the ratio smaller than 0.7 is called a scattered sub-class of the current node class.
The larger the ratio of the full sub-category corresponding to the current node category to the corresponding node category of the adjacent embedded vector graph structure is, the larger the internal consistency of the current node category is, and the higher the purity of the point is. For example, for example: in contrast to a class containing all A, B, C and a part a, a part B, and a part C, the purity of the nodes in the previous class is higher, the previous class indicates A, B, C that the nodes are similar but different, the nodes are classified into one class in the next embedded vector graph structure, the classes A, B, C are similar, the next class indicates that some nodes in the class a are more similar than some nodes in the class B, C, rather than the nodes in the same class, i.e., the classification result of the previous class is inaccurate, and the current classification result is obtained by expanding the range on the basis of the previous one, i.e., the difference of the internal nodes of the current target class is larger.
S602, calculating the consistency of the two-by-two node categories in the mutually corresponding embedded vector diagram structure based on the corresponding ratio of the scattered sub-categories of the two-by-two node categories and the difference of the two-by-two node categories and the corresponding node categories of the adjacent embedded vector diagram structure.
After the corresponding node categories of the adjacent embedded vector graph structures are divided into the complete sub-category and the scattered sub-category, the ratio of the scattered sub-category of every two node categories in the corresponding node category and the difference between every two node categories and the corresponding node category of the adjacent embedded vector graph structures (obtained by referring to the calculation mode of the difference) are calculated, and the consistency of every two node categories in the embedded vector graph structures corresponding to each other is calculated.
In one embodiment of the present application, and referring to fig. 7, the step S103: based on the difference and consistency of the two-node categories, calculating the degree of distinction corresponding to the embedded vector diagram structure specifically comprises the following steps:
s701, inputting the difference and the consistency of the two-by-two node categories into a difference value calculation formula, and calculating a difference value corresponding to the two-by-two node categories.
S702, taking the average value corresponding to the distinguishing value corresponding to all the node categories in the embedded vector diagram structure as the distinguishing degree corresponding to the embedded vector diagram structure.
Specifically, the step of inputting the difference and the consistency of the two-by-two node categories into a difference value calculation formula to calculate the difference value corresponding to the two-by-two node categories specifically includes:
wherein ,refers to the differential weight corresponding to every two node categories, ++>Means matching pair similarity parameters corresponding to every two node categories, < ->Means the number of combinations, & gt>Refers to the ratio of scattered sub-categories of one node category in the corresponding node category, +.>Refers to the +.>Difference between two categories in the two combinations, +.>Refers to the ratio of scattered sub-categories of another node category in the corresponding node category, +.>Refers to the +.f in the corresponding node class of another node class>Differences in two categories in the individual combinations.
It should be noted that, the larger the difference value corresponding to every two node categories in the current embedded vector diagram structure is, the larger the degree of distinction between the categories is, and the larger the degree of distinction corresponding to the current embedded vector diagram structure is, the more accurate the number of the node categories corresponding to the current embedded vector diagram structure is. And then the node category number determined in the mode is used as the temperature category number, namely the K value of the K-means clustering method, the subsequent temperature classification is carried out, and the obtained classification result is more accurate.
According to the embodiment of the application, at least two embedded vector graph structures are constructed based on a pixel point change sequence corresponding to a power infrared thermal image pixel, wherein two adjacent embedded vector graph structures are mutually corresponding, each embedded vector graph structure comprises at least two node categories, then the difference of every two node categories in the same embedded vector graph structure and the consistency of every two node categories in the mutually corresponding embedded vector graph structure are respectively calculated according to the at least two node categories in the embedded vector graph structure, the differentiation degree corresponding to the embedded vector graph structure is calculated based on the difference and the consistency of every two node categories, and finally the node category number corresponding to the embedded vector graph structure with the highest differentiation degree is used as a classification parameter of cluster analysis so as to classify the working temperature of the high-voltage power supply. The node class number corresponding to the embedded vector diagram structure with the highest degree of distinction is selected as the classification parameter of the cluster analysis, and then the classification of the working temperature of the high-voltage power supply is carried out, so that the accuracy of the temperature class division of the working temperature of the high-voltage power supply can be improved, and the working cost is further reduced.
It will be evident to those skilled in the art that the application is not limited to the details of the foregoing illustrative embodiments, and that the present application may be embodied in other specific forms without departing from the spirit or essential characteristics thereof. The above-described embodiments of the application are, therefore, to be considered in all respects as illustrative and not restrictive, the scope of the application being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.
Claims (8)
1. The high-voltage power supply working temperature distribution data analysis method is applied to the field of high-voltage power supply working temperature data processing, and is characterized by comprising the following steps of:
constructing at least two embedded vector diagram structures based on a pixel point change sequence corresponding to a high-voltage power supply infrared thermal image pixel, wherein two adjacent embedded vector diagram structures correspond to each other, and each embedded vector diagram structure comprises at least two node categories;
according to at least two node categories in the embedded vector diagram structure, respectively calculating the difference of every two node categories in the same embedded vector diagram structure and the consistency in the mutually corresponding embedded vector diagram structures;
based on the difference and consistency of the two-node categories, calculating the distinction degree corresponding to the embedded vector diagram structure;
and taking the node category number corresponding to the embedded vector diagram structure with the highest distinction as the classification parameter of cluster analysis so as to classify the working temperature of the high-voltage power supply.
2. The method for analyzing working temperature distribution data of high-voltage power supply according to claim 1, wherein at least two embedded vector graph structures are constructed based on a pixel point change sequence corresponding to the pixels of the infrared thermal image of the high-voltage power supply, wherein two adjacent embedded vector graph structures correspond to each other, each embedded vector graph structure comprises at least two node categories, and specifically comprises:
step 1, constructing a triangular mesh map structure based on a pixel point change sequence corresponding to a high-voltage power supply infrared thermal image pixel;
step 2, calculating an embedded vector corresponding to each node based on the triangle mesh structure and a preset vector calculation method;
step 3, taking the embedded vector corresponding to each node as a node value, taking cosine similarity of the embedded vectors corresponding to every two nodes as a side value, and taking each pixel point as a node to construct an embedded vector diagram structure;
and 4, carrying out iterative computation for preset times on the embedded vector diagram structure instead of the triangular mesh diagram structure in the step 2, and obtaining at least two embedded vector diagram structures.
3. The method for analyzing high-voltage power supply operating temperature distribution data according to claim 2, wherein the constructing a triangle mesh structure based on the pixel point change sequence corresponding to the high-voltage power supply infrared thermal image pixels specifically comprises:
acquiring sequences corresponding to pixel values of each pixel point of the power infrared thermal image at different moments, and confirming a pixel point change sequence corresponding to each pixel point;
and taking each pixel point as a node, taking a pixel point change sequence of each pixel point as a node value, and taking cosine similarity of the pixel point change sequences corresponding to every two nodes as a side value to construct the triangular mesh map structure.
4. The method for analyzing high-voltage power supply operating temperature distribution data according to claim 3, wherein the method for acquiring the variability specifically comprises:
based on the class center embedded vector corresponding to each node class, cosine similarity between each pixel point of each two node classes and the corresponding class center embedded vector is calculated respectively;
determining the pixel points with the cosine similarity larger than a preset threshold value as the corresponding retention points of each node category;
matching the reserved points corresponding to the two node categories according to a preset rule, and confirming the matched pair;
based on the matching pair, confirming the corresponding difference weight of the two node categories and the similarity parameter of the matching pair;
and calculating the difference of the two-node categories in the same embedded vector diagram structure based on the difference weight corresponding to the two-node categories and the matching pair similarity parameter.
5. The method for analyzing high-voltage power supply operating temperature distribution data according to claim 4, wherein the method for acquiring consistency comprises the following steps:
calculating the ratio of each pixel point in each two node categories in the corresponding node category of the adjacent embedded vector graph structure so as to divide the corresponding node category of the adjacent embedded vector graph structure into a complete sub-category and a scattered sub-category;
and calculating the consistency of the two-node categories in the mutually corresponding embedded vector graph structure based on the corresponding ratio of the scattered sub-categories of the two-node categories and the difference between the two-node categories and the corresponding node categories of the adjacent embedded vector graph structure.
6. The method for analyzing high-voltage power supply operating temperature distribution data according to claim 5, wherein the calculating the degree of distinction corresponding to the embedded vector diagram structure based on the difference and the consistency of the two-node categories specifically comprises:
inputting the difference and the consistency of the two-node categories into a difference value calculation formula, and calculating a difference value corresponding to the two-node categories;
and taking the average value corresponding to the distinguishing value corresponding to all the node categories in the embedded vector diagram structure as the distinguishing degree corresponding to the embedded vector diagram structure.
7. The method for analyzing high-voltage power supply operating temperature distribution data according to claim 6, wherein the step of inputting the difference and the consistency of the two-node categories into a difference value calculation formula to calculate the difference value corresponding to the two-node categories comprises the following steps:
wherein ,refers to the differential weight corresponding to every two node categories, ++>Means matching pair similarity parameters corresponding to every two node categories, < ->Means the number of combinations, & gt>Refers to the ratio of scattered sub-categories of one node category in the corresponding node category, +.>Refers to the +.>Difference between two categories in the two combinations, +.>Refers to the ratio of scattered sub-categories of another node category in the corresponding node category, +.>Refers to the +.f in the corresponding node class of another node class>Differences in two categories in the individual combinations.
8. The method for analyzing high-voltage power supply operating temperature distribution data according to any one of claims 1 to 7, wherein the classifying parameters of the cluster analysis are the number of node categories corresponding to the highest-distinction embedded vector graph structure, and specifically include:
counting the corresponding distinction between at least two embedded vector graph structures, and confirming the node category number corresponding to the embedded vector graph structure with the highest distinction;
and taking the K value as a parameter of a K-means clustering algorithm to divide the infrared thermal image into K temperature categories.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310806240.XA CN116539167B (en) | 2023-07-04 | 2023-07-04 | High-voltage power supply working temperature distribution data analysis method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310806240.XA CN116539167B (en) | 2023-07-04 | 2023-07-04 | High-voltage power supply working temperature distribution data analysis method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116539167A CN116539167A (en) | 2023-08-04 |
| CN116539167B true CN116539167B (en) | 2023-09-08 |
Family
ID=87443945
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310806240.XA Active CN116539167B (en) | 2023-07-04 | 2023-07-04 | High-voltage power supply working temperature distribution data analysis method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116539167B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116705180B (en) * | 2023-08-08 | 2023-10-31 | 山东北国发展集团有限公司 | N2O catalytic decomposition monitoring method and system based on multidimensional data analysis |
| CN116821666B (en) * | 2023-08-31 | 2023-11-03 | 陕西威思曼高压电源股份有限公司 | Real-time monitoring method for power data of high-energy ion beam high-voltage amplifier |
| CN117130409A (en) * | 2023-10-26 | 2023-11-28 | 沈阳仕腾科技有限公司 | Communication machine room temperature control method and system |
| CN117746260B (en) * | 2024-02-20 | 2024-05-31 | 山东省国土测绘院 | Remote sensing data intelligent analysis method and system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101159015A (en) * | 2007-11-08 | 2008-04-09 | 清华大学 | Two-dimension human face image recognizing method |
| CN112560740A (en) * | 2020-12-23 | 2021-03-26 | 中国水利水电科学研究院 | PCA-Kmeans-based visible light remote sensing image change detection method |
| CN113191419A (en) * | 2021-04-27 | 2021-07-30 | 河海大学 | Sag homologous event detection and type identification method based on track key point matching and region division |
| WO2023087184A1 (en) * | 2021-11-17 | 2023-05-25 | 京东方科技集团股份有限公司 | Image processing method and apparatus, computing device, and medium |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112101381B (en) * | 2020-08-30 | 2022-10-28 | 西南电子技术研究所（中国电子科技集团公司第十研究所） | Tensor collaborative drawing discriminant analysis remote sensing image feature extraction method |
-
2023
- 2023-07-04 CN CN202310806240.XA patent/CN116539167B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101159015A (en) * | 2007-11-08 | 2008-04-09 | 清华大学 | Two-dimension human face image recognizing method |
| CN112560740A (en) * | 2020-12-23 | 2021-03-26 | 中国水利水电科学研究院 | PCA-Kmeans-based visible light remote sensing image change detection method |
| CN113191419A (en) * | 2021-04-27 | 2021-07-30 | 河海大学 | Sag homologous event detection and type identification method based on track key point matching and region division |
| WO2023087184A1 (en) * | 2021-11-17 | 2023-05-25 | 京东方科技集团股份有限公司 | Image processing method and apparatus, computing device, and medium |
Non-Patent Citations (1)
| Title |
|---|
| 车联网中基于Ｋ均值聚类的拥塞控制方法;李国钦等;《计算机工程与设计》;第41卷(第2期);第334-338页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116539167A (en) | 2023-08-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116539167B (en) | High-voltage power supply working temperature distribution data analysis method | |
| Lodha et al. | Aerial lidar data classification using adaboost | |
| Everson et al. | Multi-class ROC analysis from a multi-objective optimisation perspective | |
| Antunes et al. | Knee/elbow estimation based on first derivative threshold | |
| CN112800231B (en) | Power data verification method and device, computer equipment and storage medium | |
| Meng'Ao et al. | Research and improvement of DBSCAN cluster algorithm | |
| CN106326923A (en) | Sign-in position data clustering method in consideration of position repetition and density peak point | |
| Chebbout et al. | Comparative study of clustering based colour image segmentation techniques | |
| CN115169501A (en) | Community detection method based on close similarity of common neighbor node clustering entropy | |
| Wang et al. | Automatic cluster number selection by finding density peaks | |
| CN113436223B (en) | Point cloud data segmentation method and device, computer equipment and storage medium | |
| JP2005513632A (en) | Split an image using the water source method | |
| Wang et al. | A three-way adaptive density peak clustering (3W-ADPC) method | |
| CN111724298B (en) | Dictionary optimization and mapping method for digital rock core super-dimensional reconstruction | |
| CN115270904B (en) | Method and system for spatialization of proper-age permanent population in compulsory education stage | |
| CN113065604B (en) | Air target grouping method based on DTW-DBSCAN algorithm | |
| CN105828434B (en) | Subnet division type DV-hop wireless sensor network positioning method | |
| Liu | Min-Max-Jump distance and its applications | |
| Chen et al. | A New Method of Selecting K-means Initial Cluster Centers Based on Hotspot Analysis | |
| Chang et al. | Single cell tracking based on Voronoi partition via stable matching | |
| Małyszko et al. | Adaptive rough entropy clustering algorithms in image segmentation | |
| CN112733624A (en) | People stream density detection method, system storage medium and terminal for indoor dense scene | |
| CN112765414A (en) | Graph embedding vector generation method and graph embedding-based community discovery method | |
| Li et al. | Simultaneous clustering and model selection for tensor affinities | |
| Pan et al. | An evolutionary approach based on symmetric nonnegative matrix factorization for community detection in dynamic networks |
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 |