CN116743961A - Visual intelligent analysis system of high altitude monitoring - Google Patents
Visual intelligent analysis system of high altitude monitoring Download PDFInfo
- Publication number
- CN116743961A CN116743961A CN202310716507.6A CN202310716507A CN116743961A CN 116743961 A CN116743961 A CN 116743961A CN 202310716507 A CN202310716507 A CN 202310716507A CN 116743961 A CN116743961 A CN 116743961A
- Authority
- CN
- China
- Prior art keywords
- data
- monitoring
- dimensional
- monitoring data
- module
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000012544 monitoring process Methods 0.000 title claims abstract description 143
- 238000004458 analytical method Methods 0.000 title claims abstract description 61
- 230000000007 visual effect Effects 0.000 title claims abstract description 18
- 238000012545 processing Methods 0.000 claims abstract description 32
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 25
- 238000011157 data evaluation Methods 0.000 claims abstract description 18
- 238000000034 method Methods 0.000 claims abstract description 16
- 238000011156 evaluation Methods 0.000 claims abstract description 15
- 238000012800 visualization Methods 0.000 claims abstract description 13
- 238000004140 cleaning Methods 0.000 claims abstract description 4
- 239000002245 particle Substances 0.000 claims description 39
- 230000006870 function Effects 0.000 claims description 30
- 238000005457 optimization Methods 0.000 claims description 8
- 238000009795 derivation Methods 0.000 claims description 4
- 238000013507 mapping Methods 0.000 claims description 4
- 239000003086 colorant Substances 0.000 claims description 3
- 238000011068 loading method Methods 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 claims 1
- 238000000605 extraction Methods 0.000 claims 1
- 238000007794 visualization technique Methods 0.000 abstract description 3
- 238000004364 calculation method Methods 0.000 abstract description 2
- 239000000284 extract Substances 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013079 data visualisation Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- 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
- G06V10/765—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects using rules for classification or partitioning the feature space
-
- 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
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/7715—Feature extraction, e.g. by transforming the feature space, e.g. multi-dimensional scaling [MDS]; Mappings, e.g. subspace methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious 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
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Multimedia (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Databases & Information Systems (AREA)
- Medical Informatics (AREA)
- Signal Processing (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the field of intelligent analysis and calculation, in particular to a visual intelligent analysis system for high altitude monitoring, which comprises the following components: and the monitoring and collecting module: the system is used for monitoring the high altitude and collecting monitoring data; and a data processing module: monitoring data collected by processing the cleaning; the dimension reduction analysis module: the method comprises the steps of performing dimension reduction processing on processed monitoring data through a dimension reduction analysis algorithm; and a data evaluation module: the method comprises the steps of performing balance degree evaluation on data after dimension reduction analysis; and a visualization module: and the visual display data evaluation module is used for visually displaying the evaluation result. The invention is based on a high-altitude monitoring system, acquires massive real-time monitoring data, realizes that video data participate in intelligent analysis and application of the data by structuring and standardizing the monitoring data, simultaneously performs dimension reduction processing on the acquired monitoring data by a dimension reduction analysis algorithm, performs balance degree evaluation, and presents the data by a visualization method to assist system operation and researchers to more intuitively analyze the hidden problem of complex data.
Description
Technical Field
The invention relates to the field of intelligent analysis and calculation, in particular to a visual intelligent analysis system for high-altitude monitoring.
Background
The monitoring is widely applied to various industries of society, particularly to the monitoring of high-altitude environment, and plays an important role in various aspects of social management, batting and preventing crimes, social security and control, emergency department outburst, public service and the like. However, structured data is required for the data intelligent analysis application, so that monitoring data is difficult to be a data operation analysis object. With the popularization of monitoring, the influence factors of the high-altitude monitoring data are gradually increased, so that the problem of uncertainty in the high-altitude monitoring data is gradually increased. The visualization process of the high-dimensional monitoring data is hindered, and therefore how to represent the high-dimensional monitoring data in a low-dimensional space becomes more and more important.
The invention is based on a high-altitude monitoring system, acquires massive real-time monitoring data, realizes that video data participate in intelligent analysis and application of the data by structuring and standardizing the monitoring data, simultaneously performs dimension reduction processing on the acquired monitoring data by a dimension reduction analysis algorithm, performs balance degree evaluation, and presents the data by a visualization method to assist system operation and researchers to more intuitively analyze the hidden problem of complex data.
Disclosure of Invention
The invention aims to solve the defects in the background technology by providing a visual intelligent analysis system for high-altitude monitoring.
The technical scheme adopted by the invention is as follows:
provided is a visual intelligent analysis system for high altitude monitoring, comprising:
and the monitoring and collecting module: the system is used for monitoring the high altitude and collecting monitoring data;
and a data processing module: the monitoring data are used for cleaning and collecting through ETL (extract-transform-load) processing;
the dimension reduction analysis module: the method comprises the steps of performing dimension reduction processing on processed monitoring data through a dimension reduction analysis algorithm;
and a data evaluation module: the method comprises the steps of performing balance degree evaluation on data after dimension reduction analysis;
and a visualization module: and the visual display data evaluation module is used for visually displaying the evaluation result.
As a preferred technical scheme of the invention: the ETL processing in the data processing module specifically comprises the steps of extracting, converting and loading the monitoring data, and converting the monitoring data collected by the monitoring collection module into structural monitoring data directly used by the dimension reduction analysis module.
As a preferred technical scheme of the invention: and the dimension reduction analysis algorithm in the dimension reduction analysis module performs dimension reduction processing on the monitoring data, and processes the low-dimension structured monitoring data by a T distribution random adjacent embedding method.
As a preferred technical scheme of the invention: the dimension reduction analysis algorithm is as follows:
setting any two structured monitoring data points X in high-dimensional monitoring data X i And x j Selecting a structured monitoring data point x j As a structured monitoring data point x i The probability of the adjacent point is p j|i And the conditional probability obeys Gaussian distribution, delta i Given an initialized low-dimensional space randomly for the standard deviation of Gaussian distribution, low-dimensional monitoring information data Y are generated in the low-dimensional space, and then the probability is expressed as p in the high-dimensional space j|i Probability in low dimensional space is expressed as q j|i : improving the conditional probability of the low-dimensional monitored information data Y and the original high-dimensional data X to a symmetric joint probability density, i.e. for any structured monitored data point X i And x j With p i|j ＝p j|i ，q i|j ＝q j|i The symmetry positions are equal, namely:
wherein x is k Representing structured monitoring data points in a high-dimensional space; y is i 、y j And y k Representing structured monitoring data points in a low-dimensional space;
gaussian distribution is still used in the high-dimensional monitored data space, but Gao Weijian is calculated by employing T-distribution random adjacency embedding in the low-dimensional monitored data spaceControlling joint probability density p between data samples ij ：
Wherein N is the number of the high-dimensional space structured monitoring data;
let the joint probability density of the low-dimensional monitoring data space be q ij ：
Wherein y is l Representing structured monitor data points in a low-dimensional space Y;
introducing a cost function W:
the new cost function is added into the W to the y i Conducting derivation to obtain a calculated gradient:
where σ represents the calculated gradient.
As a preferred technical scheme of the invention: in the dimension reduction analysis algorithm, iterative optimization is performed by improving a particle swarm algorithm, a low-dimension monitoring data space expression is used as an optimization result, a cost function W serving as an intermediate variable objective function is optimized, and low-dimension monitoring data points are continuously updated until a corresponding optimal solution is obtained:
wherein θ is learning rate, α (t) is learning momentum, t-1 and t-2 are iteration times, and operation is stopped when the preset accuracy is finally reached.
As a preferred technical scheme of the invention: the improved particle swarm algorithm comprises the following steps: let the position and velocity of the h particle at the t iteration be x respectively h，t And v h，t The particles update positions and speeds by supervising individuals and population extremum to obtain an optimal solution, and an update formula is as follows:
v h，t+1 ＝ω×v h，t +c 1 ×r 1 ×(p h -x h，t )+c 2 ×r 2 ×(g h -x h，t )
x h，t+1 ＝x h，t +λ×v h，t+1
wherein v is h，t+1 Represents the speed, x, of the h particle at the t+1st iteration h，t+1 Represents the position of the h particle at the t+1st iteration, ω represents the weight, c 1 And c 2 Represent learning factor, r 1 And r 2 Represents [0,1 ]]Random number, p h Representing the extremum of an individual, g h Represents population extremum, λ represents a velocity coefficient; omega max Represents the maximum value of the weight omega min Representing the minimum weight, f is the current objective function value of the particle,for the current average objective function value of all particles, f min For the minimum objective function value of all the current particles, T represents the maximum iteration number, and T represents the current iteration number.
As a preferred technical scheme of the invention: the data evaluation module is used for evaluating the balance degree specifically as follows:
let the high-dimensional data d have n low-dimensional values, and the probability corresponding to each n is P (d n ) The shannon entropy H (d) thereof is expressed as:
wherein P (d) e ) Value d is the e-th value of high-dimensional data d e Corresponding probability, H min (d) For minimum shannon entropy, H max (d) The maximum shannon entropy is the high-dimensional data u, and the value number is n;
the imbalance distribution UD (d) of the high-dimensional data d is:
UD(d)＝H(u)-H(d)＝logn-H(d)
the high-dimensional data d is calculated by using the standardized equalization degree: the dimensionality reduction data has n values, the discrete probability distribution is v, and u is the distribution with the same value number n, so that the standardized equilibrium degree of the distribution v is obtained:
wherein NUD (v) represents the normalized degree of equalization of v distribution, UD (v) represents the unbalanced degree distribution of v distribution, and H (v), H (u) and H (d) represent shannon entropy of v distribution, high-dimensional data u and high-dimensional data d, respectively.
As a preferred technical scheme of the invention: the visualization module is used for mapping the evaluation result of the data evaluation module based on the dynamic network layout.
As a preferred technical scheme of the invention: the dynamic network layout sets the image nodes as a spring system, and each node is subject to repulsive force of other nodes and is not overlapped.
As a preferred technical scheme of the invention: the dynamic network layout fuses hierarchical layout information between the monitoring data clusters and the monitoring node data, and displays different hierarchical layout information through different colors.
Compared with the prior art, the visual intelligent analysis system for high-altitude monitoring has the beneficial effects that:
the invention is based on a high-altitude monitoring system, acquires massive real-time monitoring data, realizes that video data participate in intelligent analysis and application of the data by structuring and standardizing the monitoring data, simultaneously performs dimension reduction processing on the acquired monitoring data by a dimension reduction analysis algorithm, performs balance degree evaluation, and presents the data by a visualization method to assist system operation and researchers to more intuitively analyze the hidden problem of complex data.
Drawings
Fig. 1 is a system block diagram of a preferred embodiment of the present invention.
The meaning of each label in the figure is: 100. a monitoring acquisition module; 200. a data processing module; 300. the dimension reduction analysis module; 400. a data evaluation module; 500. and a visualization module.
Detailed Description
It should be noted that, under the condition of no conflict, the embodiments of the present embodiments and features in the embodiments may be combined with each other, and 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 obviously, the described embodiments are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Referring to fig. 1, a preferred embodiment of the present invention provides an intelligent analysis system for monitoring and visualizing a high altitude, comprising:
monitoring and collecting module 100: the system is used for monitoring the high altitude and collecting monitoring data;
data processing module 200: the monitoring data are used for cleaning and collecting through ETL (extract-transform-load) processing;
dimension reduction analysis module 300: the method comprises the steps of performing dimension reduction processing on processed monitoring data through a dimension reduction analysis algorithm;
data evaluation module 400: the method comprises the steps of performing balance degree evaluation on data after dimension reduction analysis;
visualization module 500: for visually displaying the results of the data evaluation module 400.
The ETL processing in the data processing module 200 specifically includes extracting, converting, and loading the monitoring data, and converting the monitoring data collected by the monitoring collection module 100 into structured monitoring data for direct use by the dimension reduction analysis module 300.
The dimension reduction analysis algorithm in the dimension reduction analysis module 300 performs dimension reduction processing on the monitoring data, and processes the low-dimension structured monitoring data by a method of T distribution random adjacent embedding.
The dimension reduction analysis algorithm is as follows:
setting any two structured monitoring data points X in high-dimensional monitoring data X i And x j Selecting a structured monitoring data point x j As a structured monitoring data point x i The probability of the adjacent point is p j|i And the conditional probability obeys Gaussian distribution, delta i Given an initialized low-dimensional space randomly for the standard deviation of Gaussian distribution, low-dimensional monitoring information data Y are generated in the low-dimensional space, and then the probability is expressed as p in the high-dimensional space j|i Probability in low dimensional space is expressed as q j|i : improving the conditional probability of the low-dimensional monitored information data Y and the original high-dimensional data X to a symmetric joint probability density, i.e. for any structured monitored data point X i And x j With p i|j ＝p j|i ，q i|j ＝q j|i The symmetry positions are equal, namely:
wherein x is k Representing structured monitoring data points in a high-dimensional space; y is i 、y j And y k Representing structured monitoring data points in a low-dimensional space;
gaussian distribution is still used in the high-dimensional monitored data space, but joint probability density p between high-dimensional monitored data samples is calculated in the low-dimensional monitored data space by employing T-distribution random adjacency embedding ij ：
Wherein N is the number of the high-dimensional space structured monitoring data;
let the joint probability density of the low-dimensional monitoring data space be q ij ：
Wherein y is l Representing structured monitor data points in a low-dimensional space Y;
introducing a cost function W:
the new cost function is added into the W to the y i Conducting derivation to obtain a calculated gradient:
where σ represents the calculated gradient.
In the dimension reduction analysis algorithm, iterative optimization is performed by improving a particle swarm algorithm, a low-dimension monitoring data space expression is used as an optimization result, a cost function W serving as an intermediate variable objective function is optimized, and low-dimension monitoring data points are continuously updated until a corresponding optimal solution is obtained:
wherein θ is learning rate, α (t) is learning momentum, t-1 and t-2 are iteration times, and operation is stopped when the preset accuracy is finally reached.
The improved particle swarm algorithm comprises the following steps: let the position and velocity of the h particle at the t iteration be x respectively h，t And v h，t The particles update positions and speeds by supervising individuals and population extremum to obtain an optimal solution, and an update formula is as follows:
v h，t+1 ＝ω×v h，t +c 1 ×r 1 ×(p h -x h，t )+c 2 ×r 2 ×(g h -x h，t )
x h，t+1 ＝x h，t +λ×v h，t+1
wherein v is h，t+1 Represents the speed, x, of the h particle at the t+1st iteration h，t+1 Represents the position of the h particle at the t+1st iteration, ω represents the weight, c 1 And c 2 Represent learning factor, r 1 And r 2 Represents [0,1 ]]Random number, p h Representing the extremum of an individual, g h Represents population extremum, λ represents a velocity coefficient; omega max Represents the maximum value of the weight omega min Representing the minimum weight, f is the current objective function value of the particle,for the current average objective function value of all particles, f min For the minimum objective function value of all the current particles, T represents the maximum iteration number, and T represents the current iteration number.
The data evaluation module 400 evaluates the equalization degree as follows:
let the high-dimensional data d have n low-dimensional values, and the probability corresponding to each n is P (d n ) The shannon entropy H (d) thereof is expressed as:
wherein P (d) e ) Value d is the e-th value of high-dimensional data d e Corresponding probability, H min (d) For minimum shannon entropy, H max (d) The maximum shannon entropy is the high-dimensional data u, and the value number is n;
the imbalance distribution UD (d) of the high-dimensional data d is:
UD(d)＝H(u)-H(d)＝logn-H(d)
the high-dimensional data d is calculated by using the standardized equalization degree: the dimensionality reduction data has n values, the discrete probability distribution is v, and u is the distribution with the same value number n, so that the standardized equilibrium degree of the distribution v is obtained:
wherein NUD (v) represents the normalized degree of equalization of v distribution, UD (v) represents the unbalanced degree distribution of v distribution, and H (v), H (u) and H (d) represent shannon entropy of v distribution, high-dimensional data u and high-dimensional data d, respectively.
The visualization module 500 performs mapping processing on the evaluation result of the data evaluation module 400 based on the dynamic network layout.
The dynamic network layout sets the image nodes as a spring system, and each node is subject to repulsive force of other nodes and is not overlapped.
The dynamic network layout fuses hierarchical layout information between the monitoring data clusters and the monitoring node data, and displays different hierarchical layout information through different colors.
In this embodiment, the monitoring and collecting module 100 performs high-altitude monitoring and collects monitoring data, and then pre-stores the collected monitoring data in a time sequence database. The data processing module 200 extracts the collected monitoring data from the time series database in time sequence, and extracts, converts and loads the monitoring data through ETL processing, so as to convert the monitoring data into structured monitoring data which can be directly used by the dimension reduction analysis module 300. The dimension reduction analysis module 300 receives the structured monitoring data processed by the data processing module, and the dimension reduction analysis algorithm maintains the structure of the high-dimensional data in the low-dimensional space, so that the information of the original data can be more reserved, and the problem of data congestion and the data visualization effect are improved.
Any two monitoring data points X in the high-dimensional monitoring data X are set i And x j Selecting a monitoring data point x j As a monitoring data point x i The probability of the adjacent point is p j|i And conditional probability obeys Gaussian distribution, sigma i Given an initialized low-dimensional space randomly for the standard deviation of Gaussian distribution, low-dimensional monitoring information data Y are generated in the low-dimensional space, and then the probability is expressed as p in the high-dimensional space j|i Probability in low dimensional space is expressed as q j|i ：
Wherein x is k Representing structured monitoring data points in a high-dimensional space; y is i 、y j And y k Representing structured monitor data points, delta, in a low-dimensional space i The standard deviation of the gaussian distribution is given,
gaussian distribution is still used in the high-dimensional monitored data space, but the high-dimensional is calculated by employing T-distribution random contiguous embedding in the low-dimensional monitored data spaceMonitoring joint probability density p between data samples ij ：
Wherein N is the number of the high-dimensional space structured monitoring data;
let the joint probability density of the low-dimensional monitoring data space be q ij ：
Wherein y is l Representing structured monitor data points in a low-dimensional space Y;
if the high-dimensional monitoring data space and the low-dimensional monitoring data space are both in Gaussian distribution, the problem that data congestion is difficult to distinguish is easily caused. The method of T distribution random adjacent embedding is adopted, so that the problem of data congestion of the monitoring data points in the visualization module 500 when the high-dimensional monitoring data is visualized in a low-dimensional space is solved.
Introducing a cost function W:
the new cost function is added into the W to the y i Conducting derivation to obtain a calculated gradient:
where σ represents the calculated gradient.
Iterative optimization is performed based on an improved particle swarm algorithm:
let the position and velocity of the 32 nd particle at the 16 th iteration be x, respectively 32，16 And v 32，16 The particles update positions and speeds by supervising individuals and population extremum to obtain an optimal solution, and an update formula is as follows:
v 32，17 ＝ω×v 32，16 +c 1 ×r 1 ×(p 32 -x 32，t )+c 2 ×r 2 ×(g 32 -x 32，16 )
x 32，17 ＝x 32，16 +λ×v 32，17
wherein v is 32，17 Representing the speed of the 32 nd particle at the 17 th iteration, x 32，17 Represents the position of the 32 nd particle at the 17 th iteration, ω represents the weight, c 1 And c 2 Represent learning factor, r 1 And r 2 Represents [0,1 ]]Random number, p 32 Representing the extremum of an individual, g 32 Represents population extremum, λ represents a velocity coefficient; omega max Represents the maximum value of the weight omega min Representing the minimum weight, f is the current objective function value of the particle,for the current average objective function value of all particles, f min For the minimum objective function value of all the current particles, T represents the maximum iteration number, and T represents the current iteration number. Comparing all the particle objective function values, searching the current optimal particle objective function value and the position thereof, and sequentially updating the searching position and searching speed of the particles; and judging whether the iteration termination condition is met, if so, ending the algorithm, otherwise, recalculating the particle objective function value, and finally outputting the optimal solution.
Taking the low-dimensional monitoring data space expression as an optimization result, optimizing the intermediate variable objective function W, and continuously updating the low-dimensional monitoring data points until the corresponding optimal solution is obtained, and taking the 52 th iteration as an example to obtain the optimal solution:
wherein θ is learning rate, α (t) is learning momentum, and operation is stopped when the accuracy reaches a predetermined level.
The data evaluation module 400 performs equalization evaluation on the monitoring data obtained by the dimension reduction analysis of the dimension reduction analysis module 300:
let the high-dimensional data d have n dimension-reduction values, and the probability corresponding to each n is P (d n ) The shannon entropy H (d) thereof is expressed as:
wherein P (d) e ) Value d is the e-th value of high-dimensional data d e Corresponding probability, H min (d) For minimum shannon entropy, H max (d) The maximum shannon entropy is the high-dimensional data u, and the value number is n;
the imbalance distribution UD (d) of the high-dimensional data d is:
UD(d)＝H(u)-H(d)＝log6-H(d)
the high-dimensional data d is calculated by using the standardized equalization degree: the dimension reduction data has 6 values, the discrete probability distribution is v, u is the distribution with the same value number of 6, and the standardized equilibrium degree of the distribution v is obtained:
wherein NUD (v) represents the normalized degree of equalization of v distribution, UD (v) represents the unbalanced degree distribution of v distribution, and H (v), H (u) and H (d) represent shannon entropy of v distribution, high-dimensional data u and high-dimensional data d, respectively.
The visualization module 500 performs visualization processing based on the monitoring data evaluation result obtained by the data evaluation module 400, constructs a dynamic network layout, performs mapping processing on the evaluation result, sets image nodes in a visualized image as a spring system, each node is not overlapped due to repulsive force of other nodes, sets a monitoring data unbalance scale, performs color superposition according to the monitoring data balance, and facilitates data observation through different data layers of different color harnesses.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics thereof. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive, the scope of the invention 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. Any reference sign in a claim should not be construed as limiting the claim concerned.
Furthermore, it should be understood that although the present disclosure describes embodiments, not every embodiment is provided with a separate embodiment, and that this description is provided for clarity only, and that the disclosure is not limited to the embodiments described in detail below, and that the embodiments described in the examples may be combined as appropriate to form other embodiments that will be apparent to those skilled in the art.
Claims (10)
1. A visual intelligent analysis system for high altitude monitoring is characterized in that: comprising the following steps:
monitoring acquisition module (100): the system is used for monitoring the high altitude and collecting monitoring data;
a data processing module (200): the monitoring data are used for cleaning and collecting through ETL (extract-transform-load) processing;
dimension reduction analysis module (300): the method comprises the steps of performing dimension reduction processing on processed monitoring data through a dimension reduction analysis algorithm;
data evaluation module (400): the method comprises the steps of performing balance degree evaluation on data after dimension reduction analysis;
visualization module (500): for visually displaying the result of the evaluation by the data evaluation module (400).
2. The visual intelligent analysis system for high altitude monitoring according to claim 1, wherein: ETL processing in the data processing module (200) specifically comprises extraction, conversion and loading of monitoring data, and the monitoring data collected by the monitoring collection module (100) are converted into structural monitoring data directly used by the dimension reduction analysis module (300).
3. The visual intelligent analysis system for high altitude monitoring according to claim 1, wherein: the dimension reduction analysis algorithm in the dimension reduction analysis module (300) performs dimension reduction processing on the monitoring data, and processes the low-dimension structured monitoring data through a T distribution random adjacent embedding method.
4. The visual intelligent analysis system for high altitude monitoring according to claim 3, wherein: the dimension reduction analysis algorithm is as follows:
setting any two structured monitoring data points X in high-dimensional monitoring data X i And x j Selecting a structured monitoring data point x j As a structured monitoring data point x i The probability of the adjacent point is p j|i And the conditional probability obeys Gaussian distribution, delta i Given an initialized low-dimensional space randomly for the standard deviation of Gaussian distribution, low-dimensional monitoring information data Y are generated in the low-dimensional space, and then the probability is expressed as p in the high-dimensional space j|i Probability in low dimensional space is expressed as q j|i : improving the conditional probability of the low-dimensional monitored information data Y and the original high-dimensional data X to a symmetric joint probability density, i.e. for any structured monitored data point X i And x j With p i|j ＝p j|i ，q i|j ＝q j|i The symmetry positions are equal, namely:
wherein x is k Representing structured monitoring data points in a high-dimensional space; y is i 、y j And y k Representing structured monitoring data points in a low-dimensional space;
gaussian distribution is still used in the high-dimensional monitored data space, but joint probability density p between high-dimensional monitored data samples is calculated in the low-dimensional monitored data space by employing T-distribution random adjacency embedding ij ：
Wherein N is the number of the high-dimensional space structured monitoring data;
let the joint probability density of the low-dimensional monitoring data space be q ij ：
Wherein y is l Representing structured monitor data points in a low-dimensional space Y;
introducing a cost function W:
the new cost function is added into the W to the y i Conducting derivation to obtain a calculated gradient:
where σ represents the calculated gradient.
5. The visual intelligent analysis system for high altitude monitoring according to claim 4, wherein: in the dimension reduction analysis algorithm, iterative optimization is performed by improving a particle swarm algorithm, a low-dimension monitoring data space expression is used as an optimization result, a cost function W serving as an intermediate variable objective function is optimized, and low-dimension monitoring data points are continuously updated until a corresponding optimal solution is obtained:
wherein θ is learning rate, α (t) is learning momentum, t-1 and t-2 are iteration times, and operation is stopped when the preset accuracy is finally reached.
6. The visual intelligent analysis system for high altitude monitoring according to claim 5, wherein: the improved particle swarm algorithm comprises the following steps: let the position and velocity of the h particle at the t iteration be x respectively h，t And v h，t The particles update positions and speeds by supervising individuals and population extremum to obtain an optimal solution, and an update formula is as follows:
v h，t+1 ＝ω×v h，t +c 1 ×r 1 ×(p h -x h，t )+c 2 ×r 2 ×(g h -x h，t )
x h，t+1 ＝x h，t +λ×v h，t+1
wherein v is h，t+1 Represents the speed, x, of the h particle at the t+1st iteration h，t+1 Represents the position of the h particle at the t+1st iteration, ω represents the weight, c 1 And c 2 Represent learning factor, r 1 And r 2 Representation [0 ]，1]Random number, p h Representing the extremum of an individual, g h Represents population extremum, λ represents a velocity coefficient; omega max Represents the maximum value of the weight omega min Representing the minimum weight, f is the current objective function value of the particle,for the current average objective function value of all particles, f min For the minimum objective function value of all the current particles, T represents the maximum iteration number, and T represents the current iteration number.
7. The visual intelligent analysis system for high altitude monitoring according to claim 1, wherein: the data evaluation module (400) evaluates the equalization degree as follows:
let the high-dimensional data d have n low-dimensional values, and the probability corresponding to each n is P (d n ) The shannon entropy H (d) thereof is expressed as:
wherein P (d) e ) Value d is the e-th value of high-dimensional data d e Corresponding probability, H min (d) For minimum shannon entropy, H max (d) The maximum shannon entropy is the high-dimensional data u, and the value number is n;
the imbalance distribution UD (d) of the high-dimensional data d is:
UD(d)＝H(u)-H(d)＝logn-H(d)
the high-dimensional data d is calculated by using the standardized equalization degree: the dimensionality reduction data has n values, the discrete probability distribution is v, and u is the distribution with the same value number n, so that the standardized equilibrium degree of the distribution v is obtained:
wherein NUD (v) represents the normalized degree of equalization of v distribution, UD (v) represents the unbalanced degree distribution of v distribution, and H (v), H (u) and H (d) represent shannon entropy of v distribution, high-dimensional data u and high-dimensional data d, respectively.
8. The visual intelligent analysis system for high altitude monitoring according to claim 7, wherein: the visualization module (500) performs mapping processing on the evaluation result of the data evaluation module (400) based on the dynamic network layout.
9. The visual intelligent analysis system for high altitude monitoring according to claim 8, wherein: the dynamic network layout sets the image nodes as a spring system, and each node is subject to repulsive force of other nodes and is not overlapped.
10. The overhead monitoring visualization intelligent analysis system of claim 9, wherein: the dynamic network layout fuses hierarchical layout information between the monitoring data clusters and the monitoring node data, and displays different hierarchical layout information through different colors.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310716507.6A CN116743961A (en) | 2023-06-15 | 2023-06-15 | Visual intelligent analysis system of high altitude monitoring |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310716507.6A CN116743961A (en) | 2023-06-15 | 2023-06-15 | Visual intelligent analysis system of high altitude monitoring |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116743961A true CN116743961A (en) | 2023-09-12 |
Family
ID=87904153
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310716507.6A Pending CN116743961A (en) | 2023-06-15 | 2023-06-15 | Visual intelligent analysis system of high altitude monitoring |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116743961A (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106372756A (en) * | 2016-09-07 | 2017-02-01 | 南京工程学院 | Thermal power plant load optimization distribution method based on breeding particle swarm optimization |
| CN114492566A (en) * | 2021-12-20 | 2022-05-13 | 西南科技大学 | Weight-adjustable high-dimensional data dimension reduction method and system |
-
2023
- 2023-06-15 CN CN202310716507.6A patent/CN116743961A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106372756A (en) * | 2016-09-07 | 2017-02-01 | 南京工程学院 | Thermal power plant load optimization distribution method based on breeding particle swarm optimization |
| CN114492566A (en) * | 2021-12-20 | 2022-05-13 | 西南科技大学 | Weight-adjustable high-dimensional data dimension reduction method and system |
Non-Patent Citations (1)
| Title |
|---|
| 唐阳坤: "基于可视化技术的高性能集群监控数据分析", 中国知网硕士电子期刊, no. 11, pages 22 - 29 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109344285B (en) | Monitoring-oriented video map construction and mining method and equipment | |
| CN109214599B (en) | Method for predicting link of complex network | |
| Shen et al. | A deep multi-label learning framework for the intelligent fault diagnosis of machines | |
| CN113008559B (en) | Bearing fault diagnosis method and system based on sparse self-encoder and Softmax | |
| CN115565232A (en) | Power distribution room switch cabinet face part identification method based on improved YOLOv5 algorithm | |
| CN113642403B (en) | Crowd abnormal intelligent safety detection system based on edge calculation | |
| CN113037783A (en) | Abnormal behavior detection method and system | |
| Chen et al. | Applying Machine Learning Algorithm to Optimize Personalized Education Recommendation System | |
| CN114743089B (en) | Image recognition GIS fault diagnosis device and method based on SSA-SVM | |
| CN111353051A (en) | K-means and Apriori-based algorithm maritime big data association analysis method | |
| CN111966758B (en) | Electric power hidden trouble investigation method based on image data analysis technology | |
| CN109002746A (en) | 3D solid fire identification method and system | |
| CN116743961A (en) | Visual intelligent analysis system of high altitude monitoring | |
| CN113378962A (en) | Clothing attribute identification method and system based on graph attention network | |
| CN116304773A (en) | Community discovery method based on graph neural network | |
| CN115935285A (en) | Multi-element time series anomaly detection method and system based on mask map neural network model | |
| Singh et al. | Implication of Mathematics in Data Science Technology Disciplines | |
| CN111402223B (en) | Transformer substation defect problem detection method using transformer substation video image | |
| CN114169433A (en) | Industrial fault prediction method based on federal learning + image learning + CNN | |
| CN114140662A (en) | Insulator lightning stroke image sample amplification method based on cyclic generation countermeasure network | |
| CN111652275B (en) | Sparse astragal identification model construction method, sparse astragal identification method and sparse astragal identification system | |
| Alhaisoni et al. | SCF: smart big data classification framework | |
| CN116721302B (en) | Ice and snow crystal particle image classification method based on lightweight network | |
| CN113254742B (en) | Display device based on 5G deep learning artificial intelligence | |
| Luo | Research on Artificial Intelligence Applications Based on Data Mining Algorithms in the Era of Big Data |
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 |