CN117786438B - Meta-universe digital twin method and system - Google Patents
Meta-universe digital twin method and system Download PDFInfo
- Publication number
- CN117786438B CN117786438B CN202410205215.0A CN202410205215A CN117786438B CN 117786438 B CN117786438 B CN 117786438B CN 202410205215 A CN202410205215 A CN 202410205215A CN 117786438 B CN117786438 B CN 117786438B
- Authority
- CN
- China
- Prior art keywords
- data
- environment
- sensor
- abnormal
- environmental
- 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 43
- 230000002159 abnormal effect Effects 0.000 claims abstract description 106
- 230000007613 environmental effect Effects 0.000 claims abstract description 83
- 230000008859 change Effects 0.000 claims abstract description 67
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 26
- 238000001514 detection method Methods 0.000 claims abstract description 21
- 238000005070 sampling Methods 0.000 claims description 32
- 238000004590 computer program Methods 0.000 claims description 6
- 238000005286 illumination Methods 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 abstract description 10
- 230000005856 abnormality Effects 0.000 abstract description 8
- 238000004458 analytical method Methods 0.000 abstract description 7
- 238000005065 mining Methods 0.000 abstract description 6
- 230000003203 everyday effect Effects 0.000 description 9
- 230000002354 daily effect Effects 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 6
- 230000006870 function Effects 0.000 description 6
- 238000010606 normalization Methods 0.000 description 6
- 230000003993 interaction Effects 0.000 description 5
- 238000007405 data analysis Methods 0.000 description 4
- 238000001914 filtration Methods 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000009499 grossing Methods 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 230000002547 anomalous effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
Landscapes
- Testing Or Calibration Of Command Recording Devices (AREA)
Abstract
The invention relates to the technical field of metauniverse digital twin, in particular to a metauniverse digital twin method and a system, comprising the following steps: collecting data of various environmental sensors in a real scene; obtaining the relative difference between any two moments according to the environmental distribution difference change of any two moments; according to the relative difference change conditions of various environmental data in the real scene between different variables of adjacent dates and the same time, abnormal data are obtained, noise is removed from the abnormal data, and the accuracy of data prediction of meta-universe digital twin is achieved. According to the invention, the data abnormal condition is effectively identified by mining the data characteristics in the environment change process, the high-precision prediction of the data is achieved by optimizing the abnormality detection algorithm, the more real and richer virtual experience is realized, and the accuracy of virtual object analysis and prediction in the meta-space environment is improved.
Description
Technical Field
The application relates to the technical field of metauniverse digital twin, in particular to a metauniverse digital twin method and a system.
Background
The meta universe and the digital twin can map information of physical objects, devices, environments and the like in the real world into the virtual world, so that data interaction and sharing between the virtual world and the real world are realized. The digital twin technology provides rich digital twin models for various virtual objects in the metauniverse, and enables the virtual objects in the metauniverse environment to mirror, analyze and predict the behaviors of the digital twin objects through real-time data collected from sensors and other connecting devices and associated with the digital twin physical objects in the real world, so that the digital twin technology is greatly enriched in the complexity of application scenes from an internet of things platform to the metauniverse environment and expansion from a system level to a system level. Meanwhile, the meta universe and the digital twin can fuse and share data from different data sources, such as sensor data, geographic information data, service data and the like, so as to realize a more comprehensive and accurate digital twin model.
When the digital twin technology collects real-time data through the sensor and other connecting devices, noise may exist in the collected data due to factors such as environment and the like, and further the reliability of analysis and prediction results of virtual objects in the meta-universe environment may be affected by the digital twin technology. Therefore, the obtained data needs to be subjected to preprocessing operations such as data denoising.
When the existing filtering algorithm is used for denoising data, partial normal data is also changed due to the global property and the like of the filtering algorithm, in order to avoid abnormal conditions and error processing mechanisms in the data interaction process, the obtained abnormal data is processed through the abnormal analysis of the obtained data, so that the purpose of denoising the data is achieved.
Disclosure of Invention
In order to solve the technical problems, the invention provides a metauniverse digital twin method and a system, and the adopted technical scheme is as follows:
in a first aspect, an embodiment of the present invention provides a metauniverse digital twin method, including the steps of:
collecting data of various environmental sensors in a real scene, wherein the data are environmental data collected at each collecting date and each collecting time in a sampling period;
Obtaining the relative difference between any two moments according to the environmental distribution difference change of any two moments; obtaining the probability that the current time of the current date contains abnormal data according to the relative difference between the same time of adjacent dates and the relative difference between the current date and each date at the same time; obtaining the possibility that the current time of the current date contains abnormal environment data according to the relative difference between the same dates of adjacent times and the relative difference between the current time and the same date of each time;
Obtaining the environment value correlation between any two environment sensors according to the similarity condition between the environment data of any two environment sensors in the sampling period; for each environment sensor, replacing each environment sensor with the greatest correlation with the environment value of each environment sensor to obtain a replaced environment sensor layout; obtaining the possibility that each environment sensor is an abnormal sensor according to the probability of abnormal data contained in the environment sensor layout before and after the replacement of each environment sensor and the difference change condition of the possibility of abnormal environment data;
And obtaining abnormal data according to the possibility that each environment sensor is an abnormal sensor and the suspicious threshold value, denoising the abnormal data, and improving the accuracy of the meta-universe digital twin data.
Preferably, the various environmental sensors include: temperature sensor, humidity sensor, barometric sensor, illumination sensor.
Preferably, the obtaining the relative difference between any two moments according to the environmental distribution difference change between any two moments includes:
Acquiring a real scene plane graph, marking the positions of all the environmental sensors in the image according to the same proportion, and obtaining the image at each sampling moment; clustering Euclidean distance between environment sensors in the image and environment data by adopting a clustering algorithm to obtain clusters; matching Euclidean distances between cluster centers in images at any two moments by adopting a Hungary algorithm to obtain each matching pair;
calculating the absolute value of the difference between the cluster numbers in the images at any two moments; calculating Euclidean distance sum values between all matched pairs in the images at any two moments; and taking the product of the absolute value of the difference and the sum as the relative difference between any two moments.
Preferably, the obtaining the probability that the current time of the current date contains abnormal data according to the relative difference between the same time of adjacent dates and the relative difference between the current date and each date at the same time includes:
calculating a normalized value of a sum value of relative differences between all adjacent dates and the same time in a sampling period, and taking a difference result obtained by subtracting the normalized value from 1 as a distribution change degree;
And calculating the sum value of the relative differences between the current date and all the acquisition dates at the same time, and taking the normalized value of the product of the sum value and the distribution change degree as the probability that the current date and the current time contain abnormal data.
Preferably, the obtaining the possibility that the current date and the current time contain abnormal environmental data according to the relative difference between the same dates of adjacent times and the relative difference between the current time and each time on the same date includes:
obtaining cluster center change stability according to the relative difference between the same dates at all adjacent moments in the sampling period;
And calculating the sum value of the relative differences between the current time and all the time on the same date, and taking the normalized value of the product of the sum value and the cluster center change stability as the possibility that the current time of the current date contains abnormal environment data.
Preferably, the obtaining cluster center change stability according to the relative difference between the same dates at all adjacent moments in the sampling period includes:
For each date in the sampling period, taking the sum of the relative differences between all adjacent moments under the same date as the offset change degree of the date;
And calculating the normalized value of the absolute value of the difference between the offset change degrees of all adjacent dates, and taking the difference result obtained by subtracting the normalized value from 1 as cluster center change stability.
Preferably, the obtaining the correlation of the environmental values between any two environmental sensors according to the similarity between the environmental data of any two environmental sensors in the sampling period includes:
For the data of each acquisition time of each acquisition date in the sampling period, calculating the absolute value of the difference value of the data of any two environmental sensors, and taking the sum value of the absolute value of the difference value of all the acquisition times of all the acquisition dates as the environmental value correlation between any two environmental sensors.
Preferably, the obtaining the possibility that each environmental sensor is an abnormal sensor according to the probability of the abnormal data contained in the layout of the environmental sensors before and after the replacement of each environmental sensor and the difference change of the possibility of the abnormal environmental data comprises:
for the environment sensor a, acquiring the probability of containing abnormal data before the replacement of the environment sensor a and the probability of containing abnormal data after the replacement;
Calculating a ratio and a difference absolute value between the probability of containing abnormal data before replacement and the probability of containing abnormal data after replacement respectively, and calculating a product of the ratio and the difference absolute value as a first product;
And aiming at the possibility that the environment sensor a contains abnormal environment data before and after replacement, acquiring a second product by adopting a calculation method of the first product, and taking a normalized value of a product result of the first product and the second product as the possibility that the environment sensor a is the abnormal environment sensor.
Preferably, the obtaining the anomaly data according to the likelihood that each environmental sensor is an anomaly sensor and the suspicious threshold value includes:
Performing anomaly detection on the environmental data of all the environmental sensors at the current time of the current date by adopting an LOF anomaly detection algorithm to obtain outlier factors of all the environmental sensors; taking the mean value of the outlier factors of the environment sensors and the possibility of the abnormal sensors as the suspicious property of the environment sensors; and marking the environmental data of the environmental sensor with the suspicious property larger than a preset suspicious threshold value as abnormal data.
In a second aspect, an embodiment of the present invention further provides a metauniverse digital twin system, including a memory, a processor, and a computer program stored in the memory and running on the processor, where the processor implements the steps of any one of the methods described above when the processor executes the computer program.
The invention has at least the following beneficial effects:
According to the invention, by constructing the scene plan of each environment sensor in the real scene, the obtained data is conveniently analyzed by combining the real space position of the physical world, so that the data analysis has more practical significance; clustering all the environmental sensors in the scene plan according to the environmental data values and the difference between the positions of the environmental sensors to obtain the overall environmental data distribution condition of the real scene in the acquisition region at different moments; the method constructs relative differences through the overall environment distribution situation between adjacent days at the same moment, is used for representing the characteristics of daily change and invariance of the environment data in the real scene, and digs the differential characteristics in the real environment data; then, by combining the difference characteristics of all adjacent days in the sampling period, constructing a distribution change degree, analyzing the difference from the whole sampling period, and analyzing the stability of the corresponding environmental temperature of the environmental data every day; meanwhile, discarding accidental events with larger correlation between adjacent days, analyzing the inevitable events among different days, and constructing the probability of containing abnormal data in the current time of the current date by combining the distribution change degree, thereby improving the accuracy of abnormal data analysis;
Meanwhile, according to the environmental distribution difference change condition among all adjacent acquisition moments every day, the probability of containing abnormal environmental data in the current moment of the current date is built through the tiny deviation change of the environmental data in the real scene and the regularity characteristic of the environmental data, the change difference among the acquisition moments of the environmental data acquired in the current day is analyzed, the characteristic of the tiny change of the environmental data is analyzed from a local angle, the less obvious abnormal data in the mining environmental data are improved, and the abnormal data identification is more accurate; according to the invention, the environment value correlation is constructed by analyzing the environment data difference between any two environment sensors at all acquisition moments in the sampling period, and the data abnormality is analyzed from the space angle; meanwhile, the environment sensor with the highest environment value correlation with any environment sensor is used for replacing the original environment sensor, probability difference calculated between the replaced environment sensor and the replaced environment sensor is used for constructing the possibility that each environment sensor is an abnormal sensor, so that outlier factors of each environment sensor obtained by an LOF abnormal detection algorithm are optimized, the suspicious of each environment sensor is constructed, denoising optimization is carried out on abnormal data, the accuracy of an abnormal result is greatly improved, and the reliability of real-time data collected by a digital twin technology through the sensors and other connecting devices is improved; according to the invention, the data abnormal condition is effectively identified by mining the data characteristics in the environment change process, and the high-precision prediction of the data is achieved by optimizing the abnormal detection algorithm, so that more real and richer virtual experience is realized.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a metauniverse digital twin method provided by the invention;
FIG. 2 is a flowchart of screening abnormal data.
Detailed Description
In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following detailed description is given below of a meta-universe digital twin method and system according to the invention, and the detailed implementation, structure, characteristics and effects thereof are described in detail below with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
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 invention belongs.
The invention provides a metauniverse digital twin method and a system specific scheme by combining the drawings.
The invention provides a metauniverse digital twin method and a system.
Specifically, the following metauniverse digital twin method is provided, please refer to fig. 1, the method includes the following steps:
step S001, collecting various environmental data in the real scene.
The meta universe and the digital twin can map information of physical objects, devices, environments and the like in the real world into the virtual world, so that data interaction and sharing between the virtual world and the real world are realized. The data interaction process comprises the steps of acquiring various data sources used in the data interaction process from a real scene, such as sensor data, geographic information data, service data and the like, so as to realize a comprehensive and accurate digital twin model.
When an application program is used for making decisions and controlling the data of the digital twin model, accurate and real-time data are needed to support analysis, decision and control of the model and the algorithm, if the data acquired from a real scene are abnormal, the model analysis is inaccurate, and therefore the final control is abnormal. The sensor data is characterized by large data volume, real time and the like compared with geographic information data, service data and the like aiming at the data in change, namely the data obtained from the sensor, and the characteristics can cause abnormal data to easily appear in the process of collecting the sensor data. Therefore, the embodiment collects the environmental sensor data in the real scene, effectively identifies the abnormal condition of the data by mining the data characteristics in the environmental change process, and achieves high-precision prediction of the data by optimizing the abnormal detection algorithm, thereby realizing more real and richer virtual experience.
Wherein, obtain the data in the real scene through using various environmental sensor, the kind of environmental sensor includes: temperature sensor, humidity sensor, air pressure sensor, illumination sensor etc. this embodiment takes temperature sensor, humidity sensor, air pressure sensor, illumination sensor as the example and analyzes, will be in real scene according to gathering regional evenly distributed a plurality of sensors, sets up the quantity of various environmental sensor in this embodiment and is m=4, and the implementer can set up corresponding quantity of environmental sensor according to actual place needs.
And meanwhile, numbering the obtained environment sensor, and recording the data value obtained by the environment sensor according to the acquisition time. The environment sensor uploads the acquired data once per hour.
Thus, the collection of various environmental data is completed.
Step S002, carrying out anomaly detection on various collected environmental data, analyzing the anomaly condition existing in the environmental data, optimizing an LOF anomaly detection algorithm, and achieving high-accuracy prediction on the environmental data.
The obtained sensor data is detected using an existing anomaly detection algorithm, and the anomaly value of the obtained data is noted as LOF. However, the LOF algorithm only analyzes according to the historical data of the sensor, and does not consider the correlation with the data values obtained by other sensors, and meanwhile, the corresponding parameter K value in the LOF algorithm is artificially given and is not necessarily suitable for the current scene, so that the obtained data abnormal value may be inaccurate. Therefore, it is necessary to correct the obtained LOF value, thereby increasing the accuracy of the abnormality detection result.
In order to facilitate the targeted analysis of the environmental data, the embodiment takes a temperature sensor as an example to analyze the temperature data collected from the real scene.
In this embodiment, the time period is divided into 24 hours, that is, each temperature sensor has a sensor temperature-time curve every day, where the current date is denoted by D, the current time is denoted by t, and there are D acquisition dates in the acquisition period and 24 acquisition times in each acquisition date, where the empirical value of D is set to 30.
In order to distinguish the difference of temperature distribution changes in real scenes at different moments, the embodiment obtains the temperature distribution situation of the scene corresponding to each moment, and judges whether the temperature distribution situation corresponding to other moments is abnormal or not according to the change of the obtained temperature distribution situation.
Firstly, a scene plane graph is acquired, each position of a temperature sensor is marked in the image, and meanwhile, one temperature sensor is selected as an origin, so that a plane rectangular coordinate system is constructed, and the obtained data are conveniently analyzed.
Clustering is performed according to the distribution condition of the temperature values of the temperature sensors in the image, wherein a DBSCAN clustering algorithm is used as a clustering algorithm, and is not described in detail in the embodiment, so that each cluster and the cluster center of the corresponding cluster are obtained. Analyzing the change condition of the temperature distribution difference between the images at any two moments to obtain the relative difference at any two moments, and taking the relative difference between the (v) th hour on the (u) th day and the (v) th hour on the (u+1) th day as an example for analysis:
Wherein, Represents the relative difference between the v hour on the u day and the v hour on the u+1 day,Representing the absolute value of the difference between the number of clustered clusters in the images at the (u+1) th and (v) th daysRepresenting the minimum cluster number in two images of the (u+1) th and (v) th days, v/hRepresenting the Euclidean distance between the cluster centers of the kth matching pair in the two images of the (u+1) th and (v) th hours, wherein the matching process of the cluster centers is as follows: inputting the cluster center of each cluster obtained in the (v) th hour on the (u) th day and the cluster center of each cluster obtained in the (v) th hour on the (u+1) th day into a Hungary algorithm to obtain output matching pairs, wherein the number of the matching pairs is/>The hungarian algorithm is a well-known technique, and this embodiment will not be described in detail.
In the embodiment, the distribution change degree of the temperature between adjacent dates in the sampling period is obtained by analyzing the change of the centers of the corresponding clusters between all the moments in the sampling period:
Wherein, For the degree of change in the distribution of temperature between adjacent dates in a sampling period,/>Representing a normalization function,/>Total number of days indicating date of data acquisition,/>The relative difference between the v-th hour on the u-th day and the v-th hour on the u+1-th day is shown.
The smaller the distance between the cluster center obtained after clustering corresponding to the obtained temperature sensor temperature value every day and the cluster center obtained at the same time of the next day is, and the smaller the difference of the number of the obtained clusters is, the smaller the degree of change of the corresponding temperature distribution every day is, and the more stable the corresponding temperature distribution every day is.
According to the reliability degree of the condition that the temperature values are similar at the same acquisition time when different acquisition dates appear in the real scene in the sampling period, if the temperature abnormal condition appears, the temperature data difference between adjacent days at the same acquisition time can be represented.
Therefore, by analyzing the relative difference of the temperature between the current time t of the current date D and the same time t thereof and different dates, the probability that the current time t of the current date D contains abnormal data is obtained:
Wherein, For the probability of containing abnormal data in the current time t of the current date D,/>As a normalization function,/>For the degree of change in the distribution of temperature between adjacent dates in a sampling period,/>The total number of days on the date of data acquisition is indicated,The relative difference between the t-th hour on the u-th day and the t-th hour on the D-th day is shown.
The smaller the degree of change of the temperature value distribution between the same time points of adjacent date of the data is, the stronger the stability of the temperature distribution is, and the more similar the temperature distribution situation is obtained when the data are collected at the same time point under each date is, namely the smaller the probability that the temperature data obtained under all the collection dates at the current time point t of the current date D contains abnormal temperature data is.
According to the method, the temperature distribution change condition of the temperature values obtained by the temperature sensor at the adjacent daily acquisition time is analyzed, and the offset change degree of the center of the daily cluster is obtained, and the offset change degree of the day e is taken as an example:
Wherein, Represents the degree of shift change on day e,/>The relative difference between s hours on day e and s+1 hours on day e is shown.
That is, the more similar the sensor temperature distribution is to the sensor temperature distribution at the next time of the day e, the smaller the sensor temperature change degree is, namely the offset change degreeThe smaller.
The method for calculating the stability of the corresponding daily cluster center change degree is as follows:
Wherein, Representing cluster center change stability between all dates within a sampling period,/>As a normalization function,/>Total number of days indicating date of acquisition,/>Absolute value of difference indicating the degree of change in offset between day e and day e+1.
That is, the greater the stability of the degree of change in the obtained cluster according to the temperature distribution of the temperature values of the daily temperature sensor, the greater the stability of the temperature change obtained by the daily temperature sensor.
The possibility that the group of temperature data obtained at the current moment contains abnormal temperature data can be analyzed by analyzing the data obtained at the current moment, the cluster center change stability among all acquisition dates in the sampling period and the offset degree of the cluster center obtained at the moment adjacent to the current moment according to the cluster center obtained by the method:
Wherein, Representing the probability of containing abnormal temperature data in the current time t of the current date D,/>Representing a normalization function,/>The cluster center change stability among all dates in the sampling period is represented, t represents the current time of the current date and the corresponding current day acquisition time,/>, andThe relative difference between the p-th hour on day D and the t-th hour on day D is shown.
That is, the stronger the stability of the daily change of the temperature cluster center is, the smaller the difference of the number of cluster centers between all the moments in the current date D and the current moment t is, the smaller the possibility that the group of temperature data obtained at the current moment of the current date D contains abnormal temperature data is.
In this embodiment, the correlation of the temperature data of the temperature sensors at different positions is analyzed, and whether the obtained data is abnormal data or not is analyzed according to the correlation of the temperature sensors at different positions, so that filtering optimization is performed on the obtained data. The data correlation acquiring method between the temperature sensors corresponding to the different positions is as follows:
Analyzing the temperature data obtained by each temperature sensor on a temperature-time curve, and obtaining the temperature value correlation between any two temperature sensors by analyzing the similarity of the data curve and the data curves of other temperature sensors:
Wherein, Total number of days indicating date of acquisition,/>The absolute value of the difference between the temperature values obtained by the a-th temperature sensor and the b-th temperature sensor obtained in the j-th hour on the i-th day is shown. That is, the closer the temperature values obtained at each of the two temperature sensors are each day, the stronger the correlation between the temperature values of the two temperature sensors is.
Collecting a group of temperature sensor data at the current collection time, selecting the correlation with the data needing to be subjected to abnormality detectionThe strongest temperature value of the temperature sensor is replaced, so that the temperature distribution condition of the current time of the current date is calculated, the possibility that the current temperature sensor is an abnormal sensor is judged according to the difference of the temperature distribution conditions obtained before and after replacement, and the temperature sensor a is taken as an example for calculation:
Wherein, Indicating the possibility that the temperature sensor a is an abnormal sensor,/>The normalization function is represented as a function of the normalization,Representing the probability of containing abnormal temperature data in the current time t of the current date D,/>Representing the probability of abnormal temperature data contained in the current time t of the current date D after the temperature sensor a is replaced,/>For the probability of containing abnormal data in the current time t of the current date D,/>The probability that the current date D and the current time t contain abnormal data after the temperature sensor a is replaced is indicated.
I.e. the greater the difference between the probability and the likelihood of the group of data containing abnormal data according to the above-mentioned results, i.e. the calculated, when the calculated temperature sensor a is replaced、/>The larger and the greater the likelihood that it is anomalous data before replacement, i.e., the sought/>、/>The larger the probability and probability that it is then the outlier data, the smaller the probability that it is、/>The smaller the temperature sensor a is, the greater the possibility of being an abnormal sensor is explained.
The temperature data of all the temperature sensors at the current date D and the current time t are subjected to anomaly detection according to an LOF anomaly detection algorithm, which is a known technique, and is not described in detail in this embodiment, wherein the outlier factor obtained for the temperature sensor a is. Binding outlier factor/>Possibility of being abnormal with temperature sensor a/>Together, the suspicious nature of the temperature sensor a is obtained:
Wherein, Representing the suspicious nature of the temperature sensor a,/>Indicating the possibility that the temperature sensor a is an abnormal sensor,/>An outlier factor representing the temperature sensor a.
The larger the indication is, the higher the suspicious of the temperature sensor a is. The suspicious threshold eta is set, the data with the obtained temperature anomaly value larger than the threshold eta is marked as anomaly data, wherein the experience value of the suspicious threshold is set to be 0.7, and an implementer can set the suspicious threshold according to actual conditions. The screening flow chart of the abnormal data is shown in fig. 2.
Thus, abnormality detection of the temperature data of the obtained temperature sensor is completed. The data abnormality detection method of other various environment sensors is the same as the temperature data abnormality detection method of the temperature sensor, so that abnormal constants in the collected data of various environment sensors can be obtained.
And step S003, denoising the abnormal data according to the abnormal data in the environment data obtained after the abnormal detection, so as to realize the data reliability of the meta-universe digital twin.
And according to the method, the abnormal data in all the obtained environmental sensors are subjected to optimized denoising by using a local smoothing algorithm. The local smoothing algorithm is a known technique, and this embodiment is not described in detail.
And the digital twinning is carried out according to the obtained processed environment data, so that the environment data in the corresponding scene in the meta-universe environment can be conveniently predicted or other digital twinning behaviors can be conveniently realized. The digital twin technique is a known technique, and the embodiment is not described in detail.
Based on the same inventive concept as the above method, the embodiment of the present invention further provides a metauniverse digital twin system, which includes a memory, a processor, and a computer program stored in the memory and running on the processor, where the processor implements the steps of any one of the above metauniverse digital twin methods when executing the computer program.
In summary, according to the embodiment of the invention, by constructing the scene plan of each environmental sensor in the real scene, the obtained data is conveniently analyzed by combining the real space position of the physical world, so that the data analysis has more practical significance; clustering all the environmental sensors in the scene plan according to the environmental data values and the difference between the positions of the environmental sensors to obtain the overall environmental data distribution condition of the real scene in the acquisition region at different moments; according to the embodiment of the invention, the relative difference is constructed through the overall environment distribution situation between adjacent days at the same moment, so that the characteristic that the environment data changes every day in a real scene is represented, and the differential characteristics in the real environment data are mined; then, by combining the difference characteristics of all adjacent days in the sampling period, constructing a distribution change degree, analyzing the difference from the whole sampling period, and analyzing the stability of the corresponding environmental temperature of the environmental data every day; meanwhile, discarding accidental events with larger correlation between adjacent days, analyzing the inevitable events among different days, and constructing the probability of containing abnormal data in the current time of the current date by combining the distribution change degree, thereby improving the accuracy of abnormal data analysis;
Meanwhile, according to the environment distribution difference change condition among all adjacent acquisition moments every day, the embodiment of the invention constructs the probability of containing abnormal environment data in the current moment of the current date through the tiny deviation change of the environment data in the real scene and the regularity characteristic thereof, analyzes the change difference among the acquisition moments of the environment data acquired in the current day, analyzes the tiny change characteristic of the environment data from a local angle, improves the less obvious abnormal data in the mining environment data, and ensures that the abnormal data identification is more accurate; according to the embodiment of the invention, the environment value correlation is constructed by analyzing the environment data difference between any two environment sensors at all acquisition moments in the sampling period, and the data abnormality is analyzed from the space angle; meanwhile, the environment sensor with the highest environment value correlation with any environment sensor is used for replacing the original environment sensor, probability difference calculated between the replaced environment sensor and the replaced environment sensor is used for constructing the possibility that each environment sensor is an abnormal sensor, so that outlier factors of each environment sensor obtained by an LOF abnormal detection algorithm are optimized, the suspicious of each environment sensor is constructed, denoising optimization is carried out on abnormal data, the accuracy of an abnormal result is greatly improved, and the reliability of real-time data collected by a digital twin technology through the sensors and other connecting devices is improved; according to the embodiment of the invention, the data abnormal condition is effectively identified by mining the data characteristics in the environment change process, and the high-precision prediction of the data is achieved by optimizing the abnormal detection algorithm, so that more real and richer virtual experience is realized.
It should be noted that: the sequence of the embodiments of the present invention is only for description, and does not represent the advantages and disadvantages of the embodiments. And the foregoing description has been directed to specific embodiments of this specification. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
In this specification, each embodiment is described in a progressive manner, and the same or similar parts of each embodiment are referred to each other, and each embodiment mainly describes differences from other embodiments.
The above embodiments are only for illustrating the technical solution of the present application, and not for limiting the same; the technical solutions described in the foregoing embodiments are modified or some of the technical features are replaced equivalently, so that the essence of the corresponding technical solutions does not deviate from the scope of the technical solutions of the embodiments of the present application, and all the technical solutions are included in the protection scope of the present application.
Claims (5)
1. A metauniverse digital twin method, characterized in that it comprises the steps of:
collecting data of various environmental sensors in a real scene, wherein the data are environmental data collected at each collecting date and each collecting time in a sampling period;
Obtaining the relative difference between any two moments according to the environmental distribution difference change of any two moments; obtaining the probability that the current time of the current date contains abnormal data according to the relative difference between the same time of adjacent dates and the relative difference between the current date and each date at the same time; obtaining the possibility that the current time of the current date contains abnormal environment data according to the relative difference between the same dates of adjacent times and the relative difference between the current time and the same date of each time;
Obtaining the environment value correlation between any two environment sensors according to the similarity condition between the environment data of any two environment sensors in the sampling period; for each environment sensor, replacing each environment sensor with the greatest correlation with the environment value of each environment sensor to obtain a replaced environment sensor layout; obtaining the possibility that each environment sensor is an abnormal sensor according to the probability of abnormal data contained in the environment sensor layout before and after the replacement of each environment sensor and the difference change condition of the possibility of abnormal environment data;
Obtaining abnormal data according to the possibility that each environment sensor is an abnormal sensor and the suspicious threshold value, and denoising the abnormal data;
The method for obtaining the relative difference between any two moments according to the environmental distribution difference change of any two moments comprises the following steps:
Acquiring a real scene plane graph, marking the positions of all the environmental sensors in the image according to the same proportion, and obtaining the image at each sampling moment; clustering Euclidean distance between environment sensors in the image and environment data by adopting a clustering algorithm to obtain clusters; matching Euclidean distances between cluster centers in images at any two moments by adopting a Hungary algorithm to obtain each matching pair;
Calculating the absolute value of the difference between the cluster numbers in the images at any two moments; calculating Euclidean distance sum values between all matched pairs in the images at any two moments; taking the product of the absolute value of the difference and the sum as the relative difference between any two moments;
the obtaining the probability that the current time of the current date contains abnormal data according to the relative difference between the same time of adjacent dates and the relative difference between the current date and each date at the same time comprises the following steps:
calculating a normalized value of a sum value of relative differences between all adjacent dates and the same time in a sampling period, and taking a difference result obtained by subtracting the normalized value from 1 as a distribution change degree;
Calculating the sum value of the relative differences between the current date and all the acquisition dates at the same time, and taking the normalized value of the product of the sum value and the distribution change degree as the probability that the current date and the current time contain abnormal data;
The obtaining the possibility that the current date and the current time contain abnormal environment data according to the relative difference between the same dates of adjacent time and the relative difference between the current time and each time on the same date comprises the following steps:
obtaining cluster center change stability according to the relative difference between the same dates at all adjacent moments in the sampling period;
Calculating the sum value of the relative differences between the current time and all the time on the same date, and taking the normalized value of the product of the sum value and the cluster center change stability as the possibility that the current time of the current date contains abnormal environment data;
Obtaining cluster center change stability according to the relative difference between the same dates at all adjacent moments in the sampling period, wherein the cluster center change stability comprises the following steps:
For each date in the sampling period, taking the sum of the relative differences between all adjacent moments under the same date as the offset change degree of the date;
calculating the normalized value of the absolute value of the difference between the offset change degrees of all adjacent dates, and taking the difference result of subtracting the normalized value from 1 as cluster center change stability;
The method for obtaining the possibility that each environmental sensor is an abnormal sensor according to the probability of abnormal data contained in the layout of the environmental sensors before and after replacement of each environmental sensor and the difference change condition of the possibility of abnormal environmental data comprises the following steps:
for the environment sensor a, acquiring the probability of containing abnormal data before the replacement of the environment sensor a and the probability of containing abnormal data after the replacement;
Calculating a ratio and a difference absolute value between the probability of containing abnormal data before replacement and the probability of containing abnormal data after replacement respectively, and calculating a product of the ratio and the difference absolute value as a first product;
And aiming at the possibility that the environment sensor a contains abnormal environment data before and after replacement, acquiring a second product by adopting a calculation method of the first product, and taking a normalized value of a product result of the first product and the second product as the possibility that the environment sensor a is the abnormal environment sensor.
2. The metauniverse digital twinning method of claim 1, wherein the various environmental sensors comprise: temperature sensor, humidity sensor, barometric sensor, illumination sensor.
3. The metauniverse digital twin method of claim 1 wherein the deriving the environmental value correlation between any two environmental sensors based on similarities between the environmental data of any two environmental sensors over a sampling period comprises:
For the data of each acquisition time of each acquisition date in the sampling period, calculating the absolute value of the difference value of the data of any two environmental sensors, and taking the sum value of the absolute value of the difference value of all the acquisition times of all the acquisition dates as the environmental value correlation between any two environmental sensors.
4. The metauniverse digital twin method of claim 1 wherein obtaining anomaly data based on a likelihood that each environmental sensor is an anomaly sensor and a suspicion threshold comprises:
Performing anomaly detection on the environmental data of all the environmental sensors at the current time of the current date by adopting an LOF anomaly detection algorithm to obtain outlier factors of all the environmental sensors; taking the mean value of the outlier factors of the environment sensors and the possibility of the abnormal sensors as the suspicious property of the environment sensors; and marking the environmental data of the environmental sensor with the suspicious property larger than a preset suspicious threshold value as abnormal data.
5. A metauniverse digital twin system comprising a memory, a processor and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of the method of any one of claims 1-4 when the computer program is executed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410205215.0A CN117786438B (en) | 2024-02-26 | 2024-02-26 | Meta-universe digital twin method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410205215.0A CN117786438B (en) | 2024-02-26 | 2024-02-26 | Meta-universe digital twin method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117786438A CN117786438A (en) | 2024-03-29 |
| CN117786438B true CN117786438B (en) | 2024-05-10 |
Family
ID=90389466
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410205215.0A Active CN117786438B (en) | 2024-02-26 | 2024-02-26 | Meta-universe digital twin method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117786438B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115169763A (en) * | 2022-09-08 | 2022-10-11 | 广东广宇科技发展有限公司 | Waterlogging prediction method, digital twin system, electronic device and storage medium |
| CN115297184A (en) * | 2022-09-30 | 2022-11-04 | 网思科技股份有限公司 | Data acquisition system for digital twins |
| CN115471554A (en) * | 2022-09-15 | 2022-12-13 | 北京河图联合创新科技有限公司 | Method, device and storage medium for acquiring user behavior data of metauniverse space |
| CN115983901A (en) * | 2023-03-17 | 2023-04-18 | 山东捷瑞数字科技股份有限公司 | Equipment accessory demand prediction method, system and medium based on digital twin |
| CN116108506A (en) * | 2023-04-12 | 2023-05-12 | 广东奥飞数据科技股份有限公司 | Meta-universe digital asset security management system |
| CN116108645A (en) * | 2022-12-30 | 2023-05-12 | 友谊国际工程咨询股份有限公司 | Engineering digital twin method based on meta universe |
| CN117251695A (en) * | 2023-10-07 | 2023-12-19 | 腾讯科技（深圳）有限公司 | Training method, device, equipment and storage medium of data prediction model |
| CN117519486A (en) * | 2024-01-02 | 2024-02-06 | 虚拟现实(深圳)智能科技有限公司 | Metaverse digital human interaction methods, devices, equipment and storage media |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP6368227B2 (en) * | 2014-11-25 | 2018-08-01 | アズビル株式会社 | Abnormality judgment system and environmental sensor assembly |
| EP4338022A2 (en) * | 2021-05-06 | 2024-03-20 | Strong Force IoT Portfolio 2016, LLC | Quantum, biological, computer vision, and neural network systems for industrial internet of things |
-
2024
- 2024-02-26 CN CN202410205215.0A patent/CN117786438B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115169763A (en) * | 2022-09-08 | 2022-10-11 | 广东广宇科技发展有限公司 | Waterlogging prediction method, digital twin system, electronic device and storage medium |
| CN115471554A (en) * | 2022-09-15 | 2022-12-13 | 北京河图联合创新科技有限公司 | Method, device and storage medium for acquiring user behavior data of metauniverse space |
| CN115297184A (en) * | 2022-09-30 | 2022-11-04 | 网思科技股份有限公司 | Data acquisition system for digital twins |
| CN116108645A (en) * | 2022-12-30 | 2023-05-12 | 友谊国际工程咨询股份有限公司 | Engineering digital twin method based on meta universe |
| CN115983901A (en) * | 2023-03-17 | 2023-04-18 | 山东捷瑞数字科技股份有限公司 | Equipment accessory demand prediction method, system and medium based on digital twin |
| CN116108506A (en) * | 2023-04-12 | 2023-05-12 | 广东奥飞数据科技股份有限公司 | Meta-universe digital asset security management system |
| CN117251695A (en) * | 2023-10-07 | 2023-12-19 | 腾讯科技（深圳）有限公司 | Training method, device, equipment and storage medium of data prediction model |
| CN117519486A (en) * | 2024-01-02 | 2024-02-06 | 虚拟现实(深圳)智能科技有限公司 | Metaverse digital human interaction methods, devices, equipment and storage media |
Non-Patent Citations (2)
| Title |
|---|
| 数字孪生技术在社区老年人安全健康监测领域的应用探究;张捷;钱虹;周宏远;;中国医疗器械杂志;20191130(第06期);全文 * |
| 环境偶双极的数字孪生及其应用;程祖国;罗敏;;科技创新与应用;20200225(第06期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117786438A (en) | 2024-03-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115223063B (en) | Deep learning-based unmanned aerial vehicle remote sensing wheat new variety lodging area extraction method and system | |
| CN113761259B (en) | Image processing method, device and computer equipment | |
| CN114359787A (en) | Target attribute identification method and device, computer equipment and storage medium | |
| CN113408566A (en) | Target detection method and related equipment | |
| CN112183212A (en) | Weed identification method and device, terminal equipment and readable storage medium | |
| CN111831704A (en) | Abnormal data determination method and device, storage medium and electronic equipment | |
| CN113268641B (en) | User data processing method based on big data and big data server | |
| CN114912720A (en) | Memory network-based power load prediction method, device, terminal and storage medium | |
| CN111352926B (en) | Method, device, equipment and readable storage medium for data processing | |
| Ding et al. | Mit-avt clustered driving scene dataset: Evaluating perception systems in real-world naturalistic driving scenarios | |
| CN117786438B (en) | Meta-universe digital twin method and system | |
| CN114626886A (en) | Questionnaire data analysis method and system | |
| CN111507495B (en) | Method and device for predicting missing anemometry data | |
| CN117913808A (en) | A distributed photovoltaic power generation prediction method and device | |
| CN111144424A (en) | Personnel feature detection and analysis method based on clustering algorithm | |
| CN111325255A (en) | Specific crowd delineating method and device, electronic equipment and storage medium | |
| CN117576823A (en) | Queuing and calling system terminal | |
| CN111797290B (en) | Data processing method, device, storage medium and electronic device | |
| CN116484193A (en) | Crop yield prediction method, system, equipment and medium | |
| Dominguez et al. | A classification and data visualization tool applied to human migration analysis | |
| CN113868597A (en) | A regression fairness measure for age estimation | |
| CN117745472B (en) | River management method and system based on lightweight sensing model | |
| CN118098443B (en) | Online upgrading system and method for infrared gas sensor | |
| CN114896479B (en) | Online learning method, system and computer readable storage medium | |
| CN111079526A (en) | Carrier pigeon genetic relationship analysis method, device and storage medium |
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 |