CN115941807B - Efficient data compression method for park security system - Google Patents
Efficient data compression method for park security system Download PDFInfo
- Publication number
- CN115941807B CN115941807B CN202211657753.0A CN202211657753A CN115941807B CN 115941807 B CN115941807 B CN 115941807B CN 202211657753 A CN202211657753 A CN 202211657753A CN 115941807 B CN115941807 B CN 115941807B
- Authority
- CN
- China
- Prior art keywords
- data
- data point
- sensor
- coordinate system
- degree
- 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 49
- 238000013144 data compression Methods 0.000 title claims abstract description 17
- 230000005856 abnormality Effects 0.000 claims abstract description 95
- 230000002159 abnormal effect Effects 0.000 claims abstract description 76
- 230000002776 aggregation Effects 0.000 claims abstract description 34
- 238000004220 aggregation Methods 0.000 claims abstract description 34
- 238000012937 correction Methods 0.000 claims description 12
- 238000013528 artificial neural network Methods 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims 1
- 230000011218 segmentation Effects 0.000 description 9
- 238000004458 analytical method Methods 0.000 description 8
- 238000012216 screening Methods 0.000 description 7
- 238000007906 compression Methods 0.000 description 5
- 230000006835 compression Effects 0.000 description 5
- 238000007726 management method Methods 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000009434 installation Methods 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000002349 favourable effect Effects 0.000 description 2
- 238000012502 risk assessment Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 239000000779 smoke Substances 0.000 description 1
Classifications
-
- 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
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Abstract
The invention relates to the technical field of data compression, in particular to a high-efficiency data compression method for a park security system. The method comprises the following steps: constructing a three-dimensional coordinate system corresponding to each type of sensor based on data in a park acquired by the sensors of different types, and acquiring the category of the area where each sensor is located; obtaining the distribution aggregation degree of each data point based on the category, and further determining the corresponding neighborhood distance; obtaining the abnormality degree of each data point based on the data value of each data point and the data value of the data point in the neighborhood distance of each data point; and marking the pixel points with the abnormality degree larger than the abnormality degree threshold value, obtaining data points to be corrected, correcting the abnormality degree of each data point to be corrected, obtaining normal data points and abnormal data points, and compressing and storing corresponding data based on the normal data points and the abnormal data points. The invention improves the efficiency of data compression on the basis of ensuring that important data of the park security system is not lost.
Description
Technical Field
The invention relates to the technical field of data compression, in particular to a high-efficiency data compression method for a park security system.
Background
An important task for intelligent park management is to manage security operations for the park. The intelligent security system of the park is through application of the internet of things technology, so that the sub security systems are interconnected and intercommunicated, data of the park can be extracted and analyzed in real time, emergency treatment is conducted on the park in time, and the function of real-time prevention is achieved. In the intelligent security system operation management of garden, arrange various sensor equipment in the garden, through the internet of things technology, with the data transmission that the sensor gathered to intelligent garden security system in, carry out scientific analysis and decision.
However, since the number of sensors arranged in the campus is large, and the data of the sensors are various and are continuously updated in time sequence, the data amount collected by the sensors is large, which results in a large storage amount of the data. Because of limited network transmission bandwidth and storage space, the cost of the device needs to be increased if all the acquired data is stored. The existing method generally adopts a lossy compression mode to compress data acquired by a sensor, and the method can cause partial information loss, can not ensure that risks existing in a park are accurately analyzed, and further has larger potential safety hazards. Therefore, how to efficiently compress data of the park security system and ensure that important data is not lost is a problem to be solved.
Disclosure of Invention
In order to solve the problem that the data of the park security system cannot be compressed efficiently on the basis of ensuring that important data are not lost in the prior art, the invention aims to provide the data efficient compression method of the park security system, and the adopted technical scheme is as follows:
the invention provides a high-efficiency compression method for data of a park security system, which comprises the following steps:
acquiring data in a park in real time by utilizing different types of sensors;
constructing a three-dimensional coordinate system corresponding to each type of sensor based on data in the park acquired by each type of sensor, wherein the three-dimensional coordinate system comprises at least two data points; obtaining the category corresponding to the area where each sensor is located based on the park plan and the trained neural network; obtaining the distribution aggregation degree corresponding to each data point according to the category corresponding to the area where each sensor is located and the position of each data point in the three-dimensional coordinate system;
determining a neighborhood distance corresponding to each data point in the three-dimensional coordinate system based on the distribution aggregation degree and the total number of the data points corresponding to the regions belonging to the same category in the region corresponding to each data point in the three-dimensional coordinate system; obtaining the abnormality degree of each data point based on the data value of each data point and the data value of the data point in the neighborhood distance of each data point; marking the pixel points with the abnormality degree larger than the abnormality degree threshold, acquiring data points to be corrected according to marked data points and non-marked data points, and correcting the abnormality degree of each data point to be corrected to obtain corrected abnormality degree;
And acquiring normal data points and abnormal data points based on the abnormal degree of each data point and the abnormal degree of each data point to be corrected, and compressing and storing corresponding data based on the normal data points and the abnormal data points.
Preferably, the acquiring the data point to be corrected according to the marked data point and the non-marked data point includes:
for any type of sensor: calculating the ratio of the number of marked data points in the three-dimensional coordinate system corresponding to the sensor of the type to the total number of data points in the three-dimensional coordinate system corresponding to the sensor of the type, and taking the ratio as the duty ratio of abnormal data points in the three-dimensional coordinate system corresponding to the sensor of the type;
the sensor with the largest duty ratio of abnormal data points in the three-dimensional coordinate system is marked as a reference sensor; the three-dimensional coordinate system corresponding to the other types of sensors except the reference type of sensor is recorded as a coordinate system to be analyzed;
for the q-th coordinate system to be analyzed:
and marking the area corresponding to the nth non-marked data point in the coordinate system to be analyzed as a target area, judging whether the area corresponding to the marked data point exists in the three-dimensional coordinate system corresponding to the sensor of the reference type or not as the target area, and if so, marking the nth non-marked data point in the coordinate system to be analyzed as a data point to be corrected.
Preferably, correcting the degree of abnormality of each data point to be corrected to obtain the corrected degree of abnormality includes:
marking the marked data points in which the corresponding area in the three-dimensional coordinate system corresponding to the sensor of the reference type is the target area as the reference data points; and calculating the degree of abnormality of the data point to be corrected according to the Euclidean distance between the sensor for collecting the data value of the data point to be corrected and the sensor for collecting the data value of each reference data point and the data value of the data point to be corrected.
Preferably, the degree of abnormality after correction of the nth data point to be corrected in the qth coordinate system to be analyzed is calculated by using the following formula:
wherein beta is ′ qn For the degree of abnormality after correction of the nth data point to be corrected in the qth coordinate system to be analyzed, W is the number of reference data points, d n,w For Euclidean distance between the sensor collecting the data value of the nth data point to be corrected and the sensor collecting the data value of the w-th reference data point in the q-th coordinate system to be analyzed, exp () is an exponential function based on natural constant, beta qn The degree of abnormality of the nth data point to be corrected in the qth coordinate system to be analyzed.
Preferably, the obtaining the distribution aggregation degree corresponding to each data point according to the category corresponding to the area where each sensor is located and the position of each data point in the three-dimensional coordinate system includes:
For the ith data point in the three-dimensional coordinate system corresponding to any one type of sensor:
calculating the mean value of Euclidean distances between the ith data point and all data points corresponding to the region belonging to the same category in the region corresponding to the ith data point in the three-dimensional coordinate system corresponding to the sensor of the type, and marking the mean value as a first mean value; the region corresponding to the ith data point is the region where the sensor is located when the data value of the ith data point is acquired;
taking a natural constant as a base, taking the value of an exponential function taking the negative first mean value as an index as the distribution aggregation degree corresponding to the ith data point.
Preferably, the determining the neighborhood distance corresponding to each data point in the three-dimensional coordinate system based on the distribution aggregation degree and the total number of data points corresponding to the regions belonging to the same category in the region corresponding to each data point in the three-dimensional coordinate system includes:
for the ith data point in the three-dimensional coordinate system corresponding to any one type of sensor: calculating the product of the total number of data points corresponding to the region belonging to the same category and the distribution aggregation degree corresponding to the ith data point in the three-dimensional coordinate system corresponding to the sensor of the type, and marking the product as a first product; rounding down the first product, and marking the result of rounding down as a first distance; and taking the sum of the first distance and the initial neighborhood distance as the neighborhood distance corresponding to the ith data point.
Preferably, the obtaining the abnormality degree of each data point based on the data value of each data point and the data value of the data point in the neighborhood distance of each data point includes:
for the ith data point in the three-dimensional coordinate system corresponding to any one type of sensor:
obtaining the local distribution degree of the ith data point according to the difference between the data value of the ith data point and the data value of each data point in the neighborhood distance of the ith data point and the data value of the ith data point;
taking the ratio of the average value of the local distribution degree of all data points in the neighborhood distance of the ith data point and the local distribution degree of the ith data point as the abnormality degree of the ith data point.
Preferably, the acquiring normal data point and abnormal data point based on the abnormality degree of each data point and the abnormality degree after correction of each data point to be corrected, and compressing and storing the corresponding data based on the normal data point and the abnormal data point, includes:
taking corrected data points with the abnormality degree larger than the abnormality degree threshold value as abnormal data points, taking all data points except the data points to be corrected as abnormal data points, and taking all data points except the abnormal data points in all three-dimensional coordinate systems as normal data points;
And respectively encoding the data value of the normal data point and the data value of the abnormal data point by adopting run-length encoding, and storing the encoded data.
The invention has at least the following beneficial effects:
1. according to the method, a three-dimensional coordinate system corresponding to each type of sensor is built based on data in each type of sensor in a park, and in consideration of certain relevance among data acquired by the sensors of the same type, which are arranged in similar areas in the park, the method utilizes a neural network to acquire the category corresponding to each area in a park plan, calculates the distribution aggregation degree corresponding to each data point according to the category corresponding to the area where each sensor is located in the park and the position of each data point in the three-dimensional coordinate system, is used for representing the relevance among the data acquired by the sensors, further determines the neighborhood distance corresponding to each data point, can effectively avoid the follow-up interference of inappropriateness of neighborhood size setting on the accuracy of calculation results when the abnormal degree of each data point is calculated, further enables the screening result of the follow-up data to be more accurate, compresses normal data and abnormal data to different degrees respectively, and improves the efficiency and reliability of data compression on the basis of ensuring that important data of a park security system is not lost.
2. When the abnormality degree of each data point is calculated, firstly, the abnormality degree of each data point is obtained based on the data value of each data point and the data value of the data point in the neighborhood distance of each data point, and the pixel points with the abnormality degree larger than an abnormality degree threshold value are marked; in consideration of different characteristics of marked data points and non-marked data points, the abnormal degree of the marked pixel points is larger and attention should be paid to the marked pixel points, so that data points to be corrected are obtained according to the marked data points and the non-marked data points.
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 method for efficiently compressing data of a campus security system provided by the invention.
Detailed Description
In order to further describe the technical means and effects adopted by the present invention to achieve the preset purpose, the following detailed description is given below of a method for efficiently compressing data of a security system in a campus according to the present invention with reference to the accompanying drawings and preferred embodiments.
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 specific scheme of a data efficient compression method for a park security system, which is specifically described below with reference to the accompanying drawings.
An embodiment of a method for efficiently compressing data of a park security system comprises the following steps:
the embodiment provides a method for efficiently compressing data of a campus security system, as shown in fig. 1, the method for efficiently compressing data of the campus security system in the embodiment comprises the following steps:
and S1, acquiring data in the park in real time by utilizing different types of sensors.
In the management process of the campus security system, a plurality of sensors are installed in the campus to collect data information inside the campus, and collected data are transmitted to the management system through the Internet of things technology. However, because the data volume of gathering is comparatively huge to partial data that the sensor gathered does not contain or only contains a small amount of important information, if all data that the sensor gathered all transmit and store, can occupy great space, increase equipment cost, so this embodiment will screen the data that the sensor gathered, select the data that contains more information volume and influence the analysis of garden security protection system risk, reduce the redundancy of data as far as possible.
In this embodiment, a plurality of types of sensor devices are arranged in the campus, and are used for collecting data information in the campus in real time, where the types of the sensor devices include: an infrared sensor, a fire smoke sensor, a gas leakage sensor, a temperature sensor, a humidity sensor and the like; setting sampling frequencies and collecting moments of all sensors to be the same, wherein the embodiment sets all sensors to collect data once per second; the installation position of each sensor can be planned according to the distribution condition in the park, and the specific model and type of the sensor are selected by the implementer according to the specific condition.
So far, the data in each collection time park is obtained by utilizing different types of sensors.
Step S2, constructing a three-dimensional coordinate system corresponding to each type of sensor based on data in a park acquired by each type of sensor, wherein the three-dimensional coordinate system comprises at least two data points; obtaining the category corresponding to the area where each sensor is located based on the park plan and the trained neural network; and obtaining the distribution aggregation degree corresponding to each data point according to the category corresponding to the area where each sensor is located and the position of each data point in the three-dimensional coordinate system.
Because the data collected by the sensor contains no or only a small amount of important information, the data has less effect on the campus security risk analysis, namely, the data contains less information which is beneficial to the campus security risk analysis. Therefore, the embodiment calculates the abnormality degree of each data, is used for representing the information quantity of each data, which is favorable for the analysis of the security risk of the campus, and further screens the data, reduces the redundancy of the data as much as possible, and provides favorable conditions for the compression of the data.
In order to facilitate analysis of the degree of abnormality of each data acquired by the sensor, the embodiment performs data point conversion on the data acquired by the sensor, converts the acquired data into three-dimensional data points by constructing a three-dimensional data coordinate system, and determines the degree of abnormality of the corresponding data by analyzing the degree of abnormality of each three-dimensional data point. In this embodiment, a three-dimensional coordinate system is constructed for the data collected by each type of sensor, and the data in the three-dimensional coordinate system is continuously updated along with the update of the collected data, that is, along with the change of the collection time, specifically, for the gas leakage sensor: acquiring two-dimensional coordinates of each gas leakage sensor in a park plan, and constructing a three-dimensional coordinate system corresponding to each gas leakage sensor at the acquisition time according to the two-dimensional coordinates of each gas leakage sensor in the park plan and the data value acquired by each gas leakage sensor at the acquisition time at any acquisition time, wherein an X axis and a Y axis in the three-dimensional coordinate system represent coordinate values of data points, and a Z axis represents the data value of the data points; since the heights of the sensor installations of the same type are generally not very different, the present embodiment does not consider the height coordinates of the sensor. By adopting the method, the three-dimensional coordinate system corresponding to each type of sensor at each acquisition time can be obtained, and the data points in the three-dimensional coordinate system are the data values acquired by the corresponding type of sensor at the corresponding acquisition time.
Because all data points in each three-dimensional data coordinate system are data in a park collected by all sensors of each type at corresponding collection moments, the embodiment obtains the distribution characteristics of each data point by analyzing the distribution relation of the data points in the three-dimensional coordinate system corresponding to each type of sensor, further judges the abnormal distribution characteristics corresponding to the abnormal data points, and calculates the degree of abnormality corresponding to the abnormal data points. Considering that the sensors are arranged at different locations on the campus and that there is a correlation between the data collected by the same type of sensors in similar areas on the campus, the present embodiment uses the LOF algorithm (adaptive local anomaly factor detection algorithm) to analyze the degree of anomaly of the collected different types of data based on this characteristic.
For any type of sensor:
since the data values of the data points in the three-dimensional coordinate system may change along with the change of the acquisition time, the embodiment takes one acquisition time as an example, and the embodiment calculates the distribution aggregation degree corresponding to all the data points in the three-dimensional coordinate system corresponding to the sensor based on all the data points in the three-dimensional coordinate system corresponding to the sensor, and it should be noted that the distribution aggregation degree corresponding to the data points refers to the distribution aggregation degree of the positions of the sensor for acquiring the data values of the data points, namely the coordinates of the X axis and the coordinates of the Y axis of all the data points in the corresponding three-dimensional coordinate system, and is used for subsequently acquiring the corresponding neighborhood distances of the data points with different distribution characteristics in the LOF algorithm.
Firstly, the positions of all sensors arranged in a park are required to be classified, so that the relevance among the positions of the sensors is acquired, and the distribution aggregation degree among the sensors in the same type of area in the park is acquired.
Since the installation locations of the sensors are related to the distribution of the campus building, the process of classifying all the sensor locations is achieved by classifying the campus plan, the specific processes are: firstly, a camera is used for collecting a park plan, then the park plan is analyzed based on a neural network, the neural network in the embodiment adopts a semantic segmentation network, the purpose of the semantic segmentation network is classification, and the training process of the semantic segmentation network is as follows: the training data set is a sample plan of a plurality of parks; marking different areas in a sample plan of a park by adopting a manual marking mode, marking a grassland area as 0, manually marking a building area as 1, manually marking a road area as 2 and the like; the loss function of the semantic segmentation network is a cross entropy function; in a specific application, the labeling situation implementers of each area in the sample plan of the park can be set according to specific situations; the training process of the semantic segmentation network is the prior art, and is not repeated here; inputting the park plan into a trained semantic segmentation network, and outputting the park plan into categories corresponding to all areas in the park plan; the campus plan after semantic segmentation is obtained, that is, the campus plan is subjected to connected domain analysis, a plurality of connected domains in the campus plan are obtained, each connected domain is used as a region, each region in the campus plan corresponds to a category, and the category corresponding to the region where each sensor is located is obtained based on the position of each sensor and the category corresponding to each region in the campus plan, for example: if a certain sensor is located in a building area, the category corresponding to the area where the sensor is located is the building area category. So far, the category corresponding to the area where all the sensors in the park are located is obtained by adopting the method.
In the above step, the area where the sensor is located is classified, and then the distribution aggregation degree corresponding to each data point is calculated based on the classification result, that is, the distribution aggregation degree of the sensor is calculated, so as to obtain the neighborhood distance of each data point. The Euclidean distance between data points in a three-dimensional coordinate system corresponding to a certain type of sensor can represent the distribution condition of the sensor in a park; the more data are collected in the same-class area by using a certain type of sensor, the more the number of data points in a three-dimensional coordinate system corresponding to the type of sensor is, the more the number of the sensors of the type are arranged in the same-class area in a park; considering that the coordinates of the data points in the three-dimensional coordinate system can represent the distribution positions of the sensors, the embodiment determines the distribution aggregation degree corresponding to each data point according to the category corresponding to the area where each sensor is located and the position of each data point in the three-dimensional coordinate system, that is, calculates the distribution aggregation degree corresponding to each data point according to the category corresponding to each data point in the three-dimensional coordinate system and the position of each data point.
For the ith data point in the three-dimensional coordinate system corresponding to any one type of sensor:
according to the category of the area corresponding to the data point and the position of the data point, calculating the distribution aggregation degree corresponding to the data point, namely:
wherein alpha is i For the degree of distribution aggregation corresponding to the ith data point in the three-dimensional coordinate system corresponding to the sensor of the type, J i D (i, j) is the Euclidean distance between the ith data point in the three-dimensional coordinate system corresponding to the sensor of the type and the jth data point corresponding to the region corresponding to the ith data point in the three-dimensional coordinate system corresponding to the sensor of the type, wherein the ith data point and the jth data point in the three-dimensional coordinate system corresponding to the sensor of the type correspond to the region corresponding to the ith data point belong to the region of the same type, and exp () is an exponential function based on natural constants; the region corresponding to the ith data point is the region where the sensor is located when the data value of the ith data point is acquired; the d (i, j) is obtained based on the coordinates of the data points in the three-dimensional coordinate system on the X axis and the coordinates of the data points on the Y axis, namely the Euclidean distance between the positions of the two sensors for acquiring the data values of the two data points.
If the total number of the data points corresponding to the region belonging to the same category in the three-dimensional coordinate system corresponding to the sensor of the type is larger, the number of the sensors of the type arranged in the corresponding region in the park is larger; if the Euclidean distance between the ith data point and each data point corresponding to the region of the three-dimensional coordinate system corresponding to the sensor of the type, which corresponds to the region of the same category, is relatively close, the denser distribution of the data points around the ith data point in the three-dimensional coordinate system is indicated;characterizing the mean value of Euclidean distances between the ith data point and all data points corresponding to the region of the same class corresponding to the region corresponding to the ith data point in the three-dimensional coordinate system corresponding to the sensor of the type, and when the mean value is larger, describing the ith data pointThe more sparse the distribution of data points around the data points, namely the smaller the distribution aggregation degree corresponding to the ith data point in the three-dimensional coordinate system corresponding to the sensor of the type; when the average value is smaller, the denser the distribution of data points around the ith data point is indicated, that is, the greater the distribution aggregation degree corresponding to the ith data point in the three-dimensional coordinate system corresponding to the sensor of the type is.
By adopting the method, the distribution aggregation degree corresponding to each data point in each three-dimensional coordinate system can be obtained.
Step S3, determining the neighborhood distance corresponding to each data point in the three-dimensional coordinate system based on the distribution aggregation degree and the total number of the data points corresponding to the regions belonging to the same category in the region corresponding to each data point in the three-dimensional coordinate system; obtaining the abnormality degree of each data point based on the data value of each data point and the data value of the data point in the neighborhood distance of each data point; and marking the pixel points with the abnormality degree larger than the abnormality degree threshold, acquiring data points to be corrected according to the marked data points and the non-marked data points, and correcting the abnormality degree of each data point to be corrected to obtain corrected abnormality degree.
The embodiment calculates the degree of abnormality corresponding to each data point based on the idea of the LOF algorithm, wherein the main idea of the LOF algorithm is to judge whether each point is an abnormal point or not by comparing the densities of the points in the neighborhood of each point, and if the density of a certain point is lower, the point is more likely to be considered as the abnormal point; the present embodiment calculates the degree of anomaly for each data point based on the correlation between data collected by the same type of sensor located in similar areas of the campus. According to the embodiment, the semantic segmentation and LOF algorithm combination mode is adopted for analysis, namely, the semantic segmentation network is utilized to obtain the category corresponding to each area in the park plan, the data collected by the sensors arranged in the area of the same category are analyzed, the distribution aggregation degree corresponding to each data point is calculated and used for representing the relevance between the data collected by the sensors of the same type in the area of the same category, the neighborhood size of each data point in the LOF algorithm in the three-dimensional coordinate system is further obtained, the interference of the data collected by the sensors of different types to the judgment of the abnormal degree of the data point when the abnormal degree of the data point is calculated based on the LOF algorithm is eliminated, the calculated abnormal degree of the data point is more accurate, and the follow-up data screening result is more accurate.
Considering that when the abnormality degree of each data point is calculated, the neighbor data point of each data point needs to be analyzed, if the neighbor distance is not properly set, the accuracy of the result is directly affected, and the result of calculating the abnormality degree is inaccurate due to too small or too large neighbor distance, therefore, the embodiment determines the neighbor distance corresponding to each data point first, and further analyzes the abnormality degree of each data point based on the data points in the neighbor distance. When the distribution density of the data points in the three-dimensional coordinate system is higher, the data points around the data points are more densely distributed, and when the data points are analyzed, the neighborhood distance is set to be larger for improving the accuracy; based on the distribution aggregation degree corresponding to each data point in the three-dimensional coordinate system and the total number of the data points corresponding to the regions belonging to the same category in the three-dimensional coordinate system, the neighborhood distance corresponding to each data point is determined.
For the ith data point in the three-dimensional coordinate system corresponding to any one type of sensor, the corresponding neighborhood distance is as follows:
wherein k is i For the neighborhood distance, alpha, corresponding to the ith data point in the three-dimensional coordinate system corresponding to the type of sensor i The degree of distribution aggregation, k, corresponding to the ith data point in the three-dimensional coordinate system corresponding to the sensor of the type 0 For the initial neighborhood distance,is rounded downwards. The present embodiment sets the initial neighborhood distance k 0 Has a value of 2, in pixels, and in particular applications, the practitioner can vary according to particular applicationsThe situation is set.
When the distribution aggregation degree corresponding to the ith data point in the three-dimensional coordinate system is larger and the number of data points corresponding to the region, which corresponds to the ith data point and belongs to the same category, in the three-dimensional coordinate system is larger, the number of neighborhood data points which are referred to in calculating the abnormality degree of the data point is larger, namely the neighborhood distance corresponding to the data point is set to be larger, the number of the data points in the same category around the ith data point is larger.
By adopting the method, the neighborhood distance corresponding to all the data points in the three-dimensional coordinate system corresponding to different types of sensors can be obtained, all the data points in the neighborhood distance corresponding to each data point are used as the neighborhood data points of each data point, the distribution condition of each data point is analyzed, the neighborhood data point of each data point is obtained, and the interference of improper neighborhood size setting on the accuracy of a calculation result in the process of calculating the abnormal degree of each data point in the follow-up process can be effectively avoided, so that the screening result of the follow-up data is more accurate. The embodiment takes the data value of each data point in the three-dimensional coordinate system as the distribution measurement in the LOF algorithm, and calculates the abnormality degree of each data point by referring to the data value of the neighborhood data point in the neighborhood distance corresponding to each data point; the local distribution degree of each data point is obtained based on the numerical value difference of the other data points in the neighborhood distance and is used for measuring the distribution of the data values of the data points, and the main idea of the LOF algorithm is to judge whether each data point is an abnormal point or not by comparing the average value of the density of each data point and the density of the neighborhood of the data point. Thus for the ith data point in the three-dimensional coordinate system corresponding to any one type of sensor, the expression of the corresponding degree of anomaly is:
Wherein ε i Local distribution degree, beta, of the ith data point in the three-dimensional coordinate system corresponding to the sensor of the type i For the degree of abnormality of the ith data point in the three-dimensional coordinate system corresponding to the sensor of the type, M (i) is the number of data points within the neighborhood distance of the ith data point in the three-dimensional coordinate system corresponding to the sensor of the type, M is the mth data point within the neighborhood distance of the ith data point in the three-dimensional coordinate system corresponding to the sensor of the type, Δf (i, M) is the difference between the data value of the ith data point in the three-dimensional coordinate system corresponding to the sensor of the type and the data value of the mth data point within the neighborhood distance thereof, ε m Local distribution degree of the mth data point in the neighborhood distance of the ith data point in the three-dimensional coordinate system corresponding to the sensor of the type; the process for acquiring the difference between the data value of the ith data point in the three-dimensional coordinate system corresponding to the sensor of the type and the data value of the mth data point in the neighborhood distance of the data value is as follows: and taking the absolute value of the difference value between the data value of the ith data point in the three-dimensional coordinate system corresponding to the sensor of the type and the data value of the mth data point in the neighborhood distance of the ith data point in the three-dimensional coordinate system corresponding to the sensor of the type as the difference between the data value of the ith data point in the three-dimensional coordinate system corresponding to the sensor of the type and the data value of the mth data point in the neighborhood distance of the ith data point.
The average value of the local distribution degree of the data points in the neighborhood of the ith data point is represented, namely the density of the data points reflecting the neighborhood of the ith data point is obtained, and whether the ith data point is an abnormal point or not is judged by comparing the density of the ith data point with the density of the neighborhood of the ith data point; the lower the density of the ith data point, the bureau of the ith data pointThe smaller the degree of partial distribution, the greater the degree of abnormality of the i-th data point, the more likely the i-th data point is an abnormal point; the higher the density of the i-th data point, the greater the local distribution degree of the i-th data point, the closer to 1 or less than 1 the value of the abnormality degree of the i-th data point, and the more likely the i-th data point is a normal sample point.
By adopting the method, the degree of abnormality of each data point in the three-dimensional coordinate system corresponding to each type of sensor can be obtained. According to the method, the distribution situation of the data points is measured by utilizing the difference of the numerical values of each data point and other data points in the neighborhood in combination with the thought of the LOF algorithm, the abnormality degree of each data point is represented by utilizing the difference between the local distribution degree of each data point and the local distribution degree mean value of other data points in the neighborhood, namely, the abnormality degree of each data point is judged based on the relevance of each data point and surrounding data points, so that the problem that the judgment result is inaccurate when judging whether the data points are abnormal only according to the difference between the data values and the manually set threshold value in the traditional abnormality detection algorithm is avoided, and the screening result of the data is more accurate by the method provided by the embodiment.
When an abnormal situation occurs in the data of a certain area in a certain time zone, in order to ensure the safety of the zone, the data values of the area and the surrounding area thereof need to be acquired at the time zone, so in order to accurately analyze the data change situation of the surrounding area of the area where the abnormality occurs at certain critical time (namely, the time when the data abnormality occurs), the embodiment corrects the abnormality degree of the acquired partial data points. The data collected at each moment is analyzed when the degree of abnormality of the data is analyzed, the data value of the data points in the three-dimensional coordinate system is changed along with the change of time, the t moment is taken as an example for illustration, and the method provided by the embodiment can be adopted for analysis at other moments; setting an abnormality degree threshold value beta T Taking a sensor of the u type as an example, respectively judging the data value and the abnormality degree threshold beta of each data point in the three-dimensional coordinate system corresponding to the sensor of the u type T If the p-th data in the typeDegree of abnormality of point is greater than beta T Then the data point is marked as 1; if the degree of abnormality of the p-th data point in the type is less than or equal to beta T The data point is not marked; here, the description is given by taking the time t as an example, so that the data collected by all the sensors in the type u sensor at the time t are classified into two types, one type is that the degree of abnormality is greater than β T The data in the category is abnormal data, and the other category is abnormal degree less than or equal to beta T Data of (2); since the closer the value of the degree of abnormality is to 1 or less than 1, the more likely the corresponding data point is a normal sample point, thus β T The value of (2) needs to be greater than 1, beta is set in this embodiment T The value of 1.2, in a specific application, the practitioner can set according to the specific circumstances.
The present embodiment will then correct the degree of abnormality of other types of data points with the degree of abnormality of data points in the three-dimensional coordinate system corresponding to one type of sensor, specifically, for any type of sensor: counting the total number of data points in a three-dimensional coordinate system corresponding to the type of sensors, namely the number of the type of sensors in the whole park; counting the number of marked data points in a three-dimensional coordinate system corresponding to the sensor of the type at the moment t, namely acquiring the number of all abnormal data points acquired by the sensor of the type at the moment t; calculating the ratio of the number of marked data points in the three-dimensional coordinate system corresponding to the sensor of the type to the total number of data points in the three-dimensional coordinate system corresponding to the sensor of the type, and taking the ratio as the duty ratio of abnormal data points in the three-dimensional coordinate system corresponding to the sensor of the type; by adopting the method, the duty ratio of the abnormal data points in the three-dimensional coordinate system corresponding to each type of sensor can be obtained; the sensor with the largest proportion of abnormal data points in the three-dimensional coordinate system is marked as a reference type sensor and is used for correcting the abnormal degree of the data points in the three-dimensional coordinate system corresponding to other types of sensors except the reference type sensor; the three-dimensional coordinate systems corresponding to the other types of sensors except the reference type of sensor are recorded as coordinate systems to be analyzed, and the q-th coordinate system to be analyzed: taking an nth non-marked data point in the coordinate system to be analyzed as an example for explanation, marking a region corresponding to the nth non-marked data point in the coordinate system to be analyzed as a target region, judging whether the region corresponding to the marked data point exists in the three-dimensional coordinate system corresponding to the sensor of the reference type as the target region, if so, correcting the abnormal degree of the nth non-marked data point in the coordinate system to be analyzed, marking the nth non-marked data point in the coordinate system to be analyzed as the data point to be corrected, and marking the region corresponding to the three-dimensional coordinate system corresponding to the sensor of the reference type as the marked data point of the target region; if the data in the target area does not exist, the data in the target area is not abnormal, and correction of the degree of abnormality of the nth non-marked data point in the coordinate system to be analyzed is not needed; next, correcting the degree of abnormality of the data point to be corrected based on the degree of abnormality of the reference data point, considering that the data collected by the sensor closer to the position of the sensor collecting the abnormal data point should be focused, calculating the degree of abnormality of the data point to be corrected after correction according to the Euclidean distance between the sensor collecting the data value of the data point to be corrected and the sensor collecting the data value of each reference data point and the data value of the data point to be corrected, wherein the degree of abnormality of the data point to be corrected after correction in the nth coordinate system to be analyzed is:
Wherein beta is ′ qn For the degree of abnormality after correction of the nth data point to be corrected in the qth coordinate system to be analyzed, W is the number of reference data points, d n,w For Euclidean distance between the sensor collecting the data value of the nth data point to be corrected and the sensor collecting the data value of the w-th reference data point in the q-th coordinate system to be analyzed, exp () is an exponential function based on natural constant, beta qn The degree of abnormality of the nth data point to be corrected in the qth coordinate system to be analyzed; the saidThe Euclidean distance between the sensor for collecting the data value of the nth data point to be corrected and the sensor for collecting the data value of the w reference data point in the q-th coordinate system to be analyzed is the position distance of the two sensors for collecting the data value of the nth data point to be corrected and the data value of the w reference data point in the park, namely the Euclidean distance between the X coordinate and the Y coordinate of the two data points in the three-dimensional coordinate system.
The closer the data point in the coordinate system to be analyzed is to the reference data point, the closer the distance between the sensor of the type and the sensor of the reference type is, and the greater the attention of the data point in the three-dimensional coordinate system corresponding to the sensor of the type is, namely, the greater the correction weight of the abnormality degree of the data point in the three-dimensional coordinate system corresponding to the sensor of the type is.
By analogy to the method, the abnormal degrees of the data points to be corrected in all the coordinate systems to be analyzed can be corrected, and the abnormal degrees of all the data points to be corrected after correction are obtained. According to the method and the device, the relevance of the distance distribution among the sensors of different types at the same moment is considered, and the degree of abnormality acquired in advance is corrected according to the relevance of the positions when the sensors of different types collect data, so that the safety management requirements of park security are met, and the screening precision of follow-up data can be effectively guaranteed.
And S4, acquiring normal data points and abnormal data points based on the abnormal degree of each data point and the abnormal degree of each data point to be corrected, and compressing and storing corresponding data based on the normal data points and the abnormal data points.
In step S3, correcting the degree of abnormality of the data points to be corrected, obtaining the degree of abnormality after correction of each data point to be corrected, and recording all data points except the data points to be corrected as non-corrected data points, and also obtaining the degree of abnormality of each non-corrected data point; the greater the degree of anomaly the more should be of interest, the greater the degree of anomaly of the non-corrected data points will be than the threshold degree of anomaly beta T Taking each non-corrected data point as an abnormal data point, and comparing the abnormal degree of the corrected data point with that of the corrected data pointThreshold value beta of degree of abnormality T To make the corrected abnormality degree greater than the abnormality degree threshold value beta T The data points to be corrected in the three-dimensional coordinate system are used as abnormal data points, namely abnormal data points in all three-dimensional coordinate systems are obtained, all data points except the abnormal data points in all three-dimensional coordinate systems are used as normal data points, and because the abnormal analysis is carried out on the data points based on a single acquisition time, the data points acquired at each acquisition time can be divided into two types by adopting the method, namely the normal data points and the abnormal data points.
The collected data is then compressed, and in order to ensure the integrity of the data, the embodiment not only encodes and stores the data value of the abnormal data point, but also encodes and stores the data value of the normal data point. Specifically, in this embodiment, data collected in each time period is encoded respectively, and run-length encoded data is used for encoding, for any time period: the method comprises the steps that normal data points and abnormal data points possibly exist in the time period, and for the data values of the normal data points, the average value of the data values of the normal data points in the time period is used as data to be coded, and run-length coding is adopted for coding; and for the data value of the abnormal data point, directly taking the data value of the abnormal data point as data to be encoded, and encoding by adopting run-length encoding. In the embodiment, the duration of each time period is set to be 1 minute, and in a specific application, an implementer can set according to specific conditions; and the encoded data is transmitted to a campus security system server for storage, so that the data is convenient for subsequent staff to check. Run length encoding is a well known technique and will not be described in detail herein.
According to the method, a three-dimensional coordinate system corresponding to each type of sensor is built based on data in a park acquired by each type of sensor, and in consideration of certain relevance among data acquired by the same type of sensor arranged in similar areas in the park, the method utilizes a neural network to acquire categories corresponding to each area in a park plan, calculates distribution aggregation degree corresponding to each data point according to the categories corresponding to the areas where each sensor is located in the park and positions of each data point in the three-dimensional coordinate system, is used for representing relevance among data acquired by the sensors, further determines a neighborhood distance corresponding to each data point, can effectively avoid interference of inappropriateness of neighborhood size setting on accuracy of calculation results when the abnormal degree of each data point is calculated, further enables screening results of subsequent data to be more accurate, compresses normal data and abnormal data to different degrees respectively, and improves data compression efficiency and reliability on the basis of guaranteeing that important data of a park security system is not lost. When the abnormality degree of each data point is calculated, firstly, the abnormality degree of each data point is obtained based on the data value of each data point and the data value of the data point in the neighborhood distance of each data point, and the pixel points with the abnormality degree larger than the abnormality degree threshold value are marked; in consideration of different characteristics of marked data points and non-marked data points, the abnormal degree of the marked pixel points is larger and attention should be paid to the marked pixel points, so that data points to be corrected are obtained according to the marked data points and the non-marked data points, the data points to be corrected are corrected again, the abnormal degree of the data points to be corrected is obtained, all data are divided into two types, namely normal data and abnormal data, and the screening precision of the data is improved.
Claims (8)
1. The efficient data compression method for the park security system is characterized by comprising the following steps of:
acquiring data in a park in real time by utilizing different types of sensors;
constructing a three-dimensional coordinate system corresponding to each type of sensor based on data in the park acquired by each type of sensor, wherein the three-dimensional coordinate system comprises at least two data points; obtaining the category corresponding to the area where each sensor is located based on the park plan and the trained neural network; obtaining the distribution aggregation degree corresponding to each data point according to the category corresponding to the area where each sensor is located and the position of each data point in the three-dimensional coordinate system;
determining a neighborhood distance corresponding to each data point in the three-dimensional coordinate system based on the distribution aggregation degree and the total number of the data points corresponding to the regions belonging to the same category in the region corresponding to each data point in the three-dimensional coordinate system; obtaining the abnormality degree of each data point based on the data value of each data point and the data value of the data point in the neighborhood distance of each data point; marking the pixel points with the abnormality degree larger than the abnormality degree threshold, acquiring data points to be corrected according to marked data points and non-marked data points, and correcting the abnormality degree of each data point to be corrected to obtain corrected abnormality degree;
Acquiring normal data points and abnormal data points based on the abnormal degree of each data point and the abnormal degree of each data point to be corrected after correction, and compressing and storing corresponding data based on the normal data points and the abnormal data points;
the construction of the three-dimensional coordinate system corresponding to each type of sensor based on the data in the campus acquired by each type of sensor comprises the following steps:
for any type of sensor: and acquiring the two-dimensional coordinates of each sensor of the type in the park plan, and constructing a three-dimensional coordinate system corresponding to the sensor of the type at the acquisition time according to the two-dimensional coordinates of each sensor of the type in the park plan and the data value acquired by each sensor of the type at the acquisition time at any acquisition time, wherein an X axis and a Y axis in the three-dimensional coordinate system represent coordinate values of data points, and a Z axis represents the data value of the data points.
2. The efficient data compression method for a campus security system of claim 1, wherein the obtaining the data points to be corrected from the marked data points and the unmarked data points comprises:
for any type of sensor: calculating the ratio of the number of marked data points in the three-dimensional coordinate system corresponding to the sensor of the type to the total number of data points in the three-dimensional coordinate system corresponding to the sensor of the type, and taking the ratio as the duty ratio of abnormal data points in the three-dimensional coordinate system corresponding to the sensor of the type;
The sensor with the largest duty ratio of abnormal data points in the three-dimensional coordinate system is marked as a reference sensor; the three-dimensional coordinate system corresponding to the other types of sensors except the reference type of sensor is recorded as a coordinate system to be analyzed;
for the q-th coordinate system to be analyzed:
and marking the area corresponding to the nth non-marked data point in the coordinate system to be analyzed as a target area, judging whether the area corresponding to the marked data point exists in the three-dimensional coordinate system corresponding to the sensor of the reference type or not as the target area, and if so, marking the nth non-marked data point in the coordinate system to be analyzed as a data point to be corrected.
3. The efficient data compression method of the campus security system according to claim 2, wherein correcting the degree of abnormality of each data point to be corrected to obtain the corrected degree of abnormality comprises:
marking the marked data points in which the corresponding area in the three-dimensional coordinate system corresponding to the sensor of the reference type is the target area as the reference data points; and calculating the degree of abnormality of the data point to be corrected according to the Euclidean distance between the sensor for collecting the data value of the data point to be corrected and the sensor for collecting the data value of each reference data point and the data value of the data point to be corrected.
4. The efficient data compression method for a park security system according to claim 3, wherein the abnormal degree of the nth data point to be corrected in the q coordinate system to be analyzed is calculated by adopting the following formula:
wherein beta is ′ qn For the degree of abnormality after correction of the nth data point to be corrected in the qth coordinate system to be analyzed, W is the number of reference data points, d n,w Sensor for acquiring data value of nth data point to be corrected in qth coordinate system to be analyzed and data value of w reference data pointThe Euclidean distance between sensors, exp () is an exponential function with natural constant as the base, β qn The degree of abnormality of the nth data point to be corrected in the qth coordinate system to be analyzed.
5. The efficient data compression method of the campus security system according to claim 1, wherein the obtaining the distribution aggregation degree corresponding to each data point according to the category corresponding to the area where each sensor is located and the position of each data point in the three-dimensional coordinate system includes:
for the ith data point in the three-dimensional coordinate system corresponding to any one type of sensor:
calculating the mean value of Euclidean distances between the ith data point and all data points corresponding to the region belonging to the same category in the region corresponding to the ith data point in the three-dimensional coordinate system corresponding to the sensor of the type, and marking the mean value as a first mean value; the region corresponding to the ith data point is the region where the sensor is located when the data value of the ith data point is acquired;
Taking a natural constant as a base, taking the value of an exponential function taking the negative first mean value as an index as the distribution aggregation degree corresponding to the ith data point.
6. The efficient data compression method of the campus security system according to claim 1, wherein the determining the neighborhood distance corresponding to each data point in the three-dimensional coordinate system based on the distribution aggregation degree and the total number of data points corresponding to the regions belonging to the same category in the region corresponding to each data point in the three-dimensional coordinate system includes:
for the ith data point in the three-dimensional coordinate system corresponding to any one type of sensor: calculating the product of the total number of data points corresponding to the region belonging to the same category and the distribution aggregation degree corresponding to the ith data point in the three-dimensional coordinate system corresponding to the sensor of the type, and marking the product as a first product; rounding down the first product, and marking the result of rounding down as a first distance; and taking the sum of the first distance and the initial neighborhood distance as the neighborhood distance corresponding to the ith data point.
7. The efficient data compression method of the campus security system according to claim 1, wherein the obtaining the abnormality degree of each data point based on the data value of each data point and the data value of the data point in the neighborhood distance of each data point comprises:
For the ith data point in the three-dimensional coordinate system corresponding to any one type of sensor:
obtaining the local distribution degree of the ith data point according to the difference between the data value of the ith data point and the data value of each data point in the neighborhood distance of the ith data point and the data value of the ith data point;
taking the ratio of the average value of the local distribution degree of all data points in the neighborhood distance of the ith data point and the local distribution degree of the ith data point as the abnormality degree of the ith data point.
8. The efficient data compression method of the campus security system according to claim 1, wherein the acquiring the normal data point and the abnormal data point based on the abnormality degree of each data point and the abnormality degree corrected by each data point to be corrected, and the compressing and storing the corresponding data based on the normal data point and the abnormal data point includes:
taking corrected data points with the abnormality degree larger than the abnormality degree threshold value as abnormal data points, taking all data points except the data points to be corrected as abnormal data points, and taking all data points except the abnormal data points in all three-dimensional coordinate systems as normal data points;
and respectively encoding the data value of the normal data point and the data value of the abnormal data point by adopting run-length encoding, and storing the encoded data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211657753.0A CN115941807B (en) | 2022-12-22 | 2022-12-22 | Efficient data compression method for park security system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211657753.0A CN115941807B (en) | 2022-12-22 | 2022-12-22 | Efficient data compression method for park security system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115941807A CN115941807A (en) | 2023-04-07 |
| CN115941807B true CN115941807B (en) | 2024-02-23 |
Family
ID=86652214
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211657753.0A Active CN115941807B (en) | 2022-12-22 | 2022-12-22 | Efficient data compression method for park security system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115941807B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117078048B (en) * | 2023-10-17 | 2024-01-26 | 深圳市福山自动化科技有限公司 | Digital twinning-based intelligent city resource management method and system |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106878995A (en) * | 2017-04-27 | 2017-06-20 | 重庆邮电大学 | A kind of wireless sensor network Exception Type discrimination method based on perception data |
| CN110631624A (en) * | 2019-09-04 | 2019-12-31 | 精英数智科技股份有限公司 | Method, device and system for identifying abnormal operation data of mine sensor |
| CN112256686A (en) * | 2020-11-12 | 2021-01-22 | 北京理工大学 | MES-based fault data compression and prediction method and system |
| KR102321440B1 (en) * | 2020-09-10 | 2021-11-02 | 연세대학교 산학협력단 | Apparatus and Method for Anomaly Detection Using Multivariate Heterogeneous Time-Series Data |
| CN114020811A (en) * | 2021-11-09 | 2022-02-08 | 中国工商银行股份有限公司 | Data anomaly detection method and device and electronic equipment |
| CN114492629A (en) * | 2022-01-25 | 2022-05-13 | 京东城市(北京)数字科技有限公司 | Abnormality detection method, abnormality detection device, electronic apparatus, and storage medium |
| CN114970621A (en) * | 2022-05-20 | 2022-08-30 | 京东城市(北京)数字科技有限公司 | Method and device for detecting abnormal aggregation event, electronic equipment and storage medium |
| CN115294409A (en) * | 2022-10-08 | 2022-11-04 | 南通商翼信息科技有限公司 | Video compression method, system and medium for security monitoring |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE102005030031B4 (en) * | 2005-06-27 | 2007-08-02 | Nec Europe Ltd. | Method for data management in a sensor network |
| DE102020209163A1 (en) * | 2020-07-21 | 2022-01-27 | Robert Bosch Gesellschaft mit beschränkter Haftung | Method and device for detecting anomalies in sensor recordings of a technical system |
-
2022
- 2022-12-22 CN CN202211657753.0A patent/CN115941807B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106878995A (en) * | 2017-04-27 | 2017-06-20 | 重庆邮电大学 | A kind of wireless sensor network Exception Type discrimination method based on perception data |
| CN110631624A (en) * | 2019-09-04 | 2019-12-31 | 精英数智科技股份有限公司 | Method, device and system for identifying abnormal operation data of mine sensor |
| KR102321440B1 (en) * | 2020-09-10 | 2021-11-02 | 연세대학교 산학협력단 | Apparatus and Method for Anomaly Detection Using Multivariate Heterogeneous Time-Series Data |
| CN112256686A (en) * | 2020-11-12 | 2021-01-22 | 北京理工大学 | MES-based fault data compression and prediction method and system |
| CN114020811A (en) * | 2021-11-09 | 2022-02-08 | 中国工商银行股份有限公司 | Data anomaly detection method and device and electronic equipment |
| CN114492629A (en) * | 2022-01-25 | 2022-05-13 | 京东城市(北京)数字科技有限公司 | Abnormality detection method, abnormality detection device, electronic apparatus, and storage medium |
| CN114970621A (en) * | 2022-05-20 | 2022-08-30 | 京东城市(北京)数字科技有限公司 | Method and device for detecting abnormal aggregation event, electronic equipment and storage medium |
| CN115294409A (en) * | 2022-10-08 | 2022-11-04 | 南通商翼信息科技有限公司 | Video compression method, system and medium for security monitoring |
Non-Patent Citations (2)
| Title |
|---|
| 无线传感器网络中异常读数检测算法研究;薛安荣;李明;;计算机应用研究(第09期);3452-3455 * |
| 萤火虫优化和随机森林的WSN异常数据检测;许欧阳;李光辉;;计算机科学与探索(第10期);107-118 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115941807A (en) | 2023-04-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111694879B (en) | Multielement time sequence abnormal mode prediction method and data acquisition monitoring device | |
| CN111209434A (en) | Substation equipment inspection system and method based on multi-source heterogeneous data fusion | |
| CN115941807B (en) | Efficient data compression method for park security system | |
| CN115600513B (en) | Karst collapse monitoring early warning and prevention and control integrated informationized simulation research and judgment system | |
| CN113391239A (en) | Transformer abnormality monitoring method and system based on edge calculation | |
| CN115022187B (en) | Situation awareness method and device for electric-gas comprehensive energy system | |
| CN117114913A (en) | Intelligent agricultural data acquisition system based on big data | |
| CN116644284A (en) | Stratum classification characteristic factor determining method, system, electronic equipment and medium | |
| CN111855653B (en) | Plant drought monitoring method, monitoring module and monitoring device | |
| CN117421684B (en) | Abnormal data monitoring and analyzing method based on data mining and neural network | |
| CN116684878B (en) | 5G information transmission data safety monitoring system | |
| CN117009900A (en) | Internet of things signal anomaly detection method and system based on graph neural network | |
| CN117252459A (en) | Fruit quality evaluation system based on deep learning | |
| CN116861331A (en) | Expert model decision-fused data identification method and system | |
| CN114580472B (en) | Large-scale equipment fault prediction method with repeated cause and effect and attention in industrial internet | |
| CN113408816B (en) | Power grid disaster situation assessment method based on deep neural network | |
| CN113098640B (en) | Frequency spectrum anomaly detection method based on channel occupancy prediction | |
| CN108802845A (en) | A kind of indoor occupant occupation rate method of estimation based on infrared sensor array | |
| CN115062686A (en) | Multi-KPI (Key performance indicator) time sequence abnormity detection method and system based on multi-angle features | |
| CN114003638A (en) | Intelligent interconnected big data processing system | |
| Cuan et al. | Research on TV imaging casing damage detection and classification method based on C4. 5 decision tree | |
| CN113958463A (en) | Online monitoring method, system and device for fan blade | |
| CN117611015B (en) | Real-time monitoring system for quality of building engineering | |
| CN113743355B (en) | Switch device state checking method, device, system and computer equipment | |
| CN117435997B (en) | Elevator early warning and on-demand maintenance method based on transducer and time sequence compression mechanism |
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 |