CN106909490A - A kind of monitoring device data flow assessment and noise cancellation method - Google Patents
A kind of monitoring device data flow assessment and noise cancellation method Download PDFInfo
- Publication number
- CN106909490A CN106909490A CN201710111079.9A CN201710111079A CN106909490A CN 106909490 A CN106909490 A CN 106909490A CN 201710111079 A CN201710111079 A CN 201710111079A CN 106909490 A CN106909490 A CN 106909490A
- Authority
- CN
- China
- Prior art keywords
- point
- data
- test point
- correction value
- construct
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/3055—Monitoring arrangements for monitoring the status of the computing system or of the computing system component, e.g. monitoring if the computing system is on, off, available, not available
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/34—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
- G06F11/3409—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment for performance assessment
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
Abstract
The present invention relates to a kind of assessment of monitoring device data flow and noise cancellation method.The method：First, input data, according to association analysis method, obtains the relevance between the parameter to be analyzed；According to different data-stream form and type, using different counter-measures；Two sequences stronger for correlation：If two sequences produce mutation simultaneously, the principle according to correlation rule thinks that Sudden Changing Rate is not wrong data, test point is not processed；If one of sequence produces mutation, then it is assumed that be wrong data, test point is modified；For the sequence not strong with other sequences correlation：Sudden Changing Rate or the data of separation are considered wrong data, test point is modified.The present invention is realized and the assessment of monitoring device data flow and noise is eliminated, and is easy to grasp the state of monitoring device, and counter-measure is taken in time.
Description
Technical field
The present invention relates to a kind of assessment of monitoring device data flow and noise cancellation method
Background technology
With the continuous progress of ICT, the technology such as Internet of Things, cloud computing is developed rapidly, the number of production and living
Word, information-based degree more and more higher, the data volume in the whole world about will every two years be doubled, the data for not only producing daily
Amount rapid growth, data structure also becomes to become increasingly complex, including various non-structured data, traditional data processing method
Seem unable to do what one wishes, " big data epoch " have arrived, become the focus of IT industry discussion.In this background
Under, the treatment of stream data just seems more and more important.Data flow is a kind of real-time time series, data representation therein
The state of current time object, with ageing, and being continually changing over time, the change of data have again high-speed type and
Uncertainty, so need to be analyzed treatment in time to the finite data reached in data flow, the change in prediction data stream future
Change and trend, be easy to grasp the state of object, take measures in time.
The content of the invention
It is an object of the invention to provide a kind of assessment of monitoring device data flow and noise cancellation method, realize to monitoring
Device data stream is assessed and noise is eliminated, and is easy to grasp the state of monitoring device, and counter-measure is taken in time.
To achieve the above object, the technical scheme is that：A kind of monitoring device data flow assessment and noise elimination side
Method, comprises the following steps,
S1：Input data, according to association analysis method, obtains the relevance between the parameter to be analyzed；According to difference
Data-stream form and type, using different counter-measures；
S2：Two stronger sequences of correlation：If two sequences produce mutation simultaneously, the principle according to correlation rule thinks
Sudden Changing Rate is not wrong data, test point is not processed；If one of sequence produces mutation, then it is assumed that be error number
According to being modified to test point；
S3：The not strong sequence with other sequences correlation：The data of Sudden Changing Rate or separation are considered wrong data, it is right
Test point is modified.
In an embodiment of the present invention, the modification method in step S2 is：When detecting that test point needs amendment, first will
The data value of the point is calculated as zero, the correction value of the point is then calculated using backward interpolation method and forward interpolation method, specifically：If
The sequence for needing cleaning is X, and the one piece of data in sequence is { xi-5,xi-4,xi-3,xi-2,xi-1,xi,xi+1,xi+2,xi+3,xi+4,
xi+5, wherein the test point for needing amendment is xi；
(1) correction value of the point is tried to achieve first by forward interpolation method：
Construct preposition multinomialWherein Ij=1,2 ..., 5 }；
Construct preposition difference functionsJ=1,2 ..., 5；
Therefore obtaining the forward modified value of test point is
(2) correction value of the point and then using backward interpolation method is tried to achieve：
Construct rearmounted multinomialWherein Ij=1,2 ..., 5 }；
Construct rearmounted difference functionsJ=1,2 ..., 5；
Therefore obtaining the backward correction value of test point is
(3) the final correction value for trying to achieve test point is
In an embodiment of the present invention, the modification method in step S3 is：When detecting that test point needs amendment, first will
The data value of the point is calculated as zero, the correction value of the point is then calculated using backward interpolation method and forward interpolation method, specifically：If
The sequence for needing cleaning is X, and the one piece of data in sequence is { xi-5,xi-4,xi-3,xi-2,xi-1,xi,xi+1,xi+2,xi+3,xi+4,
xi+5, wherein the test point for needing amendment is xi；
(1) correction value of the point is tried to achieve first by forward interpolation method：
Construct preposition multinomialWherein Ij=1,2 ..., 5 }；
Construct preposition difference functionsJ=1,2 ..., 5；
Therefore obtaining the forward modified value of test point is
(2) correction value of the point and then using backward interpolation method is tried to achieve：
Construct rearmounted multinomialWherein Ij=1,2 ..., 5 }；
Construct rearmounted difference functionsJ=1,2 ..., 5；
Therefore obtaining the backward correction value of test point is
(3) the final correction value for trying to achieve test point is
In an embodiment of the present invention, the association analysis method for being used in step S1 is DBSCAN clustering algorithms.
In an embodiment of the present invention, in step S1, before using association analysis method, also need to lack input data
Mistake value is detected, if detecting missing values, is replaced using average value.
Compared to prior art, the invention has the advantages that：The present invention realizes and monitoring device data flow is commented
Estimate and eliminated with noise, be easy to grasp the state of monitoring device, counter-measure is taken in time.
Brief description of the drawings
Fig. 1 is the inventive method flow chart.
Fig. 2 is DBSCAN examples.
Fig. 3 is total active and three-phase total current wash result (strong correlation) figure of one embodiment of the invention three-phase.
Fig. 4 is that one embodiment of the invention three-phase is total active and three-phase total current wash result (strong correlation, there is noise) figure.
Specific embodiment
Below in conjunction with the accompanying drawings, technical scheme is specifically described.
As shown in figure 1, a kind of monitoring device data flow assessment of the invention and noise cancellation method, comprise the following steps,
S1：Input data, according to association analysis method, obtains the relevance between the parameter to be analyzed；According to difference
Data-stream form and type, using different counter-measures；
S2：Two stronger sequences of correlation：If two sequences produce mutation simultaneously, the principle according to correlation rule thinks
Sudden Changing Rate is not wrong data, test point is not processed；If one of sequence produces mutation, then it is assumed that be error number
According to being modified to test point；
S3：The not strong sequence with other sequences correlation：The data of Sudden Changing Rate or separation are considered wrong data, it is right
Test point is modified.
Modification method in step S2, S3 is：When detecting that test point needs amendment, first the data value of the point is calculated as
Zero, the correction value of the point is then calculated using backward interpolation method and forward interpolation method, specifically：If the sequence for needing cleaning is
X, the one piece of data in sequence is { xi-5,xi-4,xi-3,xi-2,xi-1,xi,xi+1,xi+2,xi+3,xi+4,xi+5, wherein needing amendment
Test point be xi；
(1) correction value of the point is tried to achieve first by forward interpolation method：
Construct preposition multinomialWherein Ij=1,2 ..., 5 }；
Construct preposition difference functionsJ=1,2 ..., 5；
Therefore obtaining the forward modified value of test point is
(2) correction value of the point and then using backward interpolation method is tried to achieve：
Construct rearmounted multinomialWherein Ij=1,2 ..., 5 }；
Construct rearmounted difference functionsJ=1,2 ..., 5；
Therefore obtaining the backward correction value of test point is
(3) the final correction value for trying to achieve test point is
The association analysis method used in step S1 is DBSCAN clustering algorithms.In step S1, association analysis side is being used
Before method, also need to carry out missing values detection to input data, if detecting missing values, replaced using average value.
It is below specific implementation process of the invention.
DBSCAN (Density-based Spatial Clustering of Application with Noise) is calculated
Method is belonging to a kind of Spatial Data Clustering method based on density mode, is most initially to propose one by Ester Martin et al.
Plant algorithm.The algorithm can originally have highdensity region division into different clusters, and can be for " noise
" Clustering and similar cluster of arbitrary shape can be also found in spatial data.The thought of the core the most of DBSCAN algorithms
It is：For each analysis object in each cluster, in the neighborhood of given radius (conventional Eps is represented)
(neighborhood) interior data object number must be greater than the set-point of initial setting.That is neighborhood density must be greater than one
Fixed threshold value (conventional MinPts is represented).DBSCAN algorithms only use query search one by retrieving consecutive points to all of data
It is secondary to can be obtained by last result, so its speed of service is quickly.And DBSCAN also have one it is very big be a little it
The cluster property of arbitrary shape can be processed, is not disturbed by noise, and also can removed according to threshold value MinPts and contain
Some noises.
The definition involved by DBSCAN algorithms is given below：
Define 1 point of Eps- neighborhoods：The Eps neighborhoods of any point p refer to p as the center of circle, with Eps as radius in space
The set of the point included in region.
Define 2 density：The density of any point p is with point p as the center of circle, in the round region with Eps as radius in space
Number comprising point.
Define 3 core points and boundary point：The density of certain point then claims if greater than a certain given threshold value MinPts in space
The point is core point.Otherwise the point is called boundary point.
Define 4 direct density reachable：Point p is reachable from the point direct density of q, if they meet following two conditions：P is in neighbour
In domain；Q is core point.
Define 5 density reachable：Point p is reachable from point q density, if (p1,p2,…,pn, wherein p1=p, pn=q) and have piFrom
pi+1Direct density is reachable.
Define the connection of 6 density：Point p and electricity q are density connections, if being p and q all reachable from o density to arbitrary o.
7 are defined to cluster：The nonempty set A of database D is a class, A meets following condition that and if only if "
(1) for p and q, if p ∈ A, and q is can reach from p density, then q ∈ A；
(2) for p and q, if p ∈ A and q ∈ A, p and q are density connections.
Define 8 noises：The point that any class is not belonging in database D is noise.
DBSCAN is different from other some clustering algorithms, and DBSCAN algorithms can be found that some exceptions for being unsatisfactory for cluster
Point.As shown in Fig. 2 what is represented in fig. 2 is the cluster obtained with the clustering method that radius neighborhood Eps is input.DBSCAN is calculated
Method depends on another parameter minimal amount Minpts simultaneously, if minimal amount Minpts be set be equal to 4, cluster 3 and cluster
Group 5 is considered isolated abnormity point.
The most representational algorithm that DBSCAN algorithms can be classified as in clustering method.The class that it will can be analyzed
The maximum set body of the connected point of density is defined as, at the same time also there is noise resisting ability very high.So being dug to data
Pick can usually use the algorithm of DBSCAN during finding isolated point.But DBSCAN itself also has certain limitation.It is first
First DBSCAN depends on two parameters of input, that is, need that parameter radius neighborhood Eps and minimal amount Minpts is determined in advance,
For the selection of different parameters, the cluster result of final data can be caused different.Secondly in DBSCAN algorithms, due to variable half
Neighborhood Eps and minimal amount Minpts are arranged to global variable in footpath, and with uniqueness, so working as data distribution not
Clustering Effect is general when uniform.But due to being to apply in power system power transmission and transforming equipment Monitoring Data DBSCAN algorithms,
In view of its monitoring and the data characteristicses that collect, so latter problem for existing can be without the concern for.
According to the advantage of above-mentioned DBSCAN algorithms, as shown in figure 1, a kind of monitoring device data flow of the invention is assessed and made an uproar
Sound removing method, comprises the following steps,
S1：Input data, according to association analysis method, obtains the relevance between the parameter to be analyzed；According to difference
Data-stream form and type, using different counter-measures；Wherein, the association analysis method for using is calculated for DBSCAN is clustered
Method, separately, before using association analysis method, also needs to carry out missing values detection to input data, if detecting missing values, uses
Average value replaces.
S2：Two stronger sequences of correlation：If two sequences produce mutation simultaneously, the principle according to correlation rule thinks
Sudden Changing Rate is not wrong data, test point is not processed；If one of sequence produces mutation, then it is assumed that be error number
According to being modified to test point；
S3：The not strong sequence with other sequences correlation：The data of Sudden Changing Rate or separation are considered wrong data, it is right
Test point is modified.
Above-mentioned modification method is：When detecting that test point needs amendment, the data value of the point is first calculated as zero, then made
The correction value of the point is calculated with backward interpolation method and forward interpolation method, specifically：If the sequence for needing cleaning is X, in sequence
One piece of data be { xi-5,xi-4,xi-3,xi-2,xi-1,xi,xi+1,xi+2,xi+3,xi+4,xi+5, wherein needing the test point of amendment
It is xi；
(1) correction value of the point is tried to achieve first by forward interpolation method：
Construct preposition multinomialWherein Ij=1,2 ..., 5 }；
Construct preposition difference functionsJ=1,2 ..., 5；
Therefore obtaining the forward modified value of test point is
(2) correction value of the point and then using backward interpolation method is tried to achieve：
Construct rearmounted multinomialWherein Ij=1,2 ..., 5 }；
Construct rearmounted difference functionsJ=1,2 ..., 5；
Therefore obtaining the backward correction value of test point is
(3) the final correction value for trying to achieve test point is
It is below specific embodiment of the invention.
By taking power grid environment temperature data as an example.Wherein initial data overall length is the sequence number n=10, q in 480, FCM analyses
=48.
As shown in figure 3, first subgraph is input initial data, wherein what blue line represented is the total wattful power of three-phase circuit
Rate, what green line was represented is the total current of three-phase circuit.Second subgraph is revised image, wherein the expression three-phase of black
The total current of circuit, the expression of pink colour is the total active power of three-phase circuit.Blue circles mark is obtained by DBSCAN algorithms
The isolated point of the total active power for going out, the corresponding time is t=28 and t=84.Red circle mark is the isolated of total current
Point, the corresponding time is t=28 and t=205.It can be seen that both are identicals in the position that first isolated point occurs.Again
Because total have strong correlation between active power and total current, the data to this time point are not processed.And for
There is the isolated point at the t=84 moment in three phase power, three-phase current does not occur exception, it is possible to think that this point is abnormal
Point does to replace, it is necessary to be modified by data cleansing to the data of the point.Similarly, occur three-phase current in t=
The point at 205 moment is needed also exist for being cleaned and corrected.Both revised images are as shown in second subgraph, it can be seen that first
Individual point has been retained, and part analysis and treatment are entered for follow-up, and remaining point is processed.
As shown in figure 4, first subgraph is input initial data, wherein what blue line represented is the total wattful power of three-phase circuit
Rate, what green line was represented is the total current of three-phase circuit.Second subgraph is revised image, wherein the expression three-phase of black
The total current of circuit, the expression of pink colour is the total active power of three-phase circuit.Near wherein to total active power t=300 of three-phase
It has been superimposed a noise.Find that this part has substantial amounts of abnormity point by DBSCAN algorithms, then entered by cleaning algorithm
Go preliminary cleaning, reduce the influence of noise.But still there is certain noise jamming, it is necessary to follow-up further place
Reason and analysis.
Above is presently preferred embodiments of the present invention, all changes made according to technical solution of the present invention, produced function work
During with scope without departing from technical solution of the present invention, protection scope of the present invention is belonged to.
Claims (5)
1. a kind of monitoring device data flow is assessed and noise cancellation method, it is characterised in that：Comprise the following steps,
S1：Input data, according to association analysis method, obtains the relevance between the parameter to be analyzed；According to different numbers
According to manifold formula and type, using different counter-measures；
S2：Two stronger sequences of correlation：If two sequences produce mutation simultaneously, the principle according to correlation rule thinks mutation
Amount is not wrong data, test point is not processed；If one of sequence produces mutation, then it is assumed that be wrong data,
Test point is modified；
S3：The not strong sequence with other sequences correlation：Sudden Changing Rate or the data of separation are considered wrong data, to detection
Point is modified.
2. a kind of monitoring device data flow according to claim 1 is assessed and noise cancellation method, it is characterised in that：Step
Modification method in S2 is：When detecting that test point needs amendment, the data value of the point is first calculated as zero, then using backward
Interpolation method and forward interpolation method calculate the correction value of the point, specifically：If the sequence for needing cleaning is X, one section in sequence
Data are { xi-5,xi-4,xi-3,xi-2,xi-1,xi,xi+1,xi+2,xi+3,xi+4,xi+5, wherein the test point for needing amendment is xi；
(1) correction value of the point is tried to achieve first by forward interpolation method：
Construct preposition multinomialWherein Ij=1,2 ..., 5 }；
Construct preposition difference functions
Therefore obtaining the forward modified value of test point is
(2) correction value of the point and then using backward interpolation method is tried to achieve：
Construct rearmounted multinomialWherein Ij=1,2 ..., 5 }；
Construct rearmounted difference functions
Therefore obtaining the backward correction value of test point is
(3) the final correction value for trying to achieve test point is
3. a kind of monitoring device data flow according to claim 1 is assessed and noise cancellation method, it is characterised in that：Step
Modification method in S3 is：When detecting that test point needs amendment, the data value of the point is first calculated as zero, then using backward
Interpolation method and forward interpolation method calculate the correction value of the point, specifically：If the sequence for needing cleaning is X, one section in sequence
Data are { xi-5,xi-4,xi-3,xi-2,xi-1,xi,xi+1,xi+2,xi+3,xi+4,xi+5, wherein the test point for needing amendment is xi；
(1) correction value of the point is tried to achieve first by forward interpolation method：
Construct preposition multinomialWherein Ij=1,2 ..., 5 }；
Construct preposition difference functions
Therefore obtaining the forward modified value of test point is
(2) correction value of the point and then using backward interpolation method is tried to achieve：
Construct rearmounted multinomialWherein Ij=1,2 ..., 5 }；
Construct rearmounted difference functions
Therefore obtaining the backward correction value of test point is
(3) the final correction value for trying to achieve test point is
4. a kind of monitoring device data flow according to claim 1 is assessed and noise cancellation method, it is characterised in that：Step
The association analysis method used in S1 is DBSCAN clustering algorithms.
5. a kind of monitoring device data flow according to claim 1 is assessed and noise cancellation method, it is characterised in that：Step
In S1, before using association analysis method, also need to carry out missing values detection to input data, if detecting missing values, use
Average value replaces.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710111079.9A CN106909490B (en) | 2017-02-28 | 2017-02-28 | Monitoring equipment data flow evaluation and noise elimination method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710111079.9A CN106909490B (en) | 2017-02-28 | 2017-02-28 | Monitoring equipment data flow evaluation and noise elimination method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106909490A true CN106909490A (en) | 2017-06-30 |
| CN106909490B CN106909490B (en) | 2020-05-05 |
Family
ID=59209189
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710111079.9A Active CN106909490B (en) | 2017-02-28 | 2017-02-28 | Monitoring equipment data flow evaluation and noise elimination method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106909490B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110083593A (en) * | 2019-04-23 | 2019-08-02 | 中国大唐集团科学技术研究院有限公司华东电力试验研究院 | The cleaning of power station operating parameter and restorative procedure, repair system |
| CN110188810A (en) * | 2019-05-23 | 2019-08-30 | 西北工业大学 | A kind of detection of online outlier, identification and modification method based on Density Clustering |
| CN112699113A (en) * | 2021-01-12 | 2021-04-23 | 上海交通大学 | Industrial manufacturing process operation monitoring system driven by time sequence data stream |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105488736A (en) * | 2015-12-02 | 2016-04-13 | 国家电网公司 | Data processing method for photovoltaic power station data acquisition system |
| CN106295850A (en) * | 2016-07-15 | 2017-01-04 | 广东电网有限责任公司电力科学研究院 | A kind of bad data recognition method and system based on historical metrology |
-
2017
- 2017-02-28 CN CN201710111079.9A patent/CN106909490B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105488736A (en) * | 2015-12-02 | 2016-04-13 | 国家电网公司 | Data processing method for photovoltaic power station data acquisition system |
| CN106295850A (en) * | 2016-07-15 | 2017-01-04 | 广东电网有限责任公司电力科学研究院 | A kind of bad data recognition method and system based on historical metrology |
Non-Patent Citations (2)
| Title |
|---|
| 戴慧等: "多变量时间序列滑动窗口异常点的检测", 《南京信息工程大学学报：自然科学版》 * |
| 王志勇: "数据挖掘方法在短期负荷预测中的应用研究", 《中国博士学位论文全文数据库工程科技II辑》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110083593A (en) * | 2019-04-23 | 2019-08-02 | 中国大唐集团科学技术研究院有限公司华东电力试验研究院 | The cleaning of power station operating parameter and restorative procedure, repair system |
| CN110083593B (en) * | 2019-04-23 | 2023-02-10 | 中国大唐集团科学技术研究院有限公司华东电力试验研究院 | Power station operation parameter cleaning and repairing method and repairing system |
| CN110188810A (en) * | 2019-05-23 | 2019-08-30 | 西北工业大学 | A kind of detection of online outlier, identification and modification method based on Density Clustering |
| CN112699113A (en) * | 2021-01-12 | 2021-04-23 | 上海交通大学 | Industrial manufacturing process operation monitoring system driven by time sequence data stream |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106909490B (en) | 2020-05-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Liu et al. | Filter design for autoregressive moving average graph filters | |
| CN106909490A (en) | A kind of monitoring device data flow assessment and noise cancellation method | |
| JP2012530319A5 (en) | ||
| WO2016165378A1 (en) | Energy storage power station mass data cleaning method and system | |
| JP5733530B2 (en) | Data similarity calculation method and data similarity calculation device | |
| KR20080021592A (en) | Image comparison by metric embeddings | |
| CN110569957A (en) | Optimization method based on artificial bee colony algorithm | |
| CN103455534B (en) | Document clustering method and device | |
| CN113746798B (en) | Cloud network shared resource abnormal root cause positioning method based on multi-dimensional analysis | |
| Wu et al. | Reducing the search space of a linear fractional programming problem under fuzzy relational equations with max-Archimedean t-norm composition | |
| CN112101765A (en) | Abnormal data processing method and system for operation index data of power distribution network | |
| CN110995153A (en) | Abnormal data detection method and device for photovoltaic power station and electronic equipment | |
| Fan et al. | Zoning search using a hyper-heuristic algorithm | |
| CN106780633B (en) | A kind of method for correcting image, device and binocular vision system | |
| CN105354228A (en) | Similar image searching method and apparatus | |
| CN110505322A (en) | A kind of IP address section lookup method and device | |
| Liu et al. | Research on the strategy of locating abnormal data in IOT management platform based on improved modified particle swarm optimization convolutional neural network algorithm | |
| CN107193979B (en) | Method for searching homologous images | |
| CN201156248Y (en) | Hardware construction for computing number of multiple data ordering | |
| WO2019087122A1 (en) | Mutually repulsing centroids for segmenting a vast social graph | |
| Ruchay et al. | Removal of impulsive noise from color images with cascade switching algorithm | |
| JP6079738B2 (en) | Histogram intersection calculation device, data collation device, similar data search device, histogram intersection calculation method and program | |
| Zhao et al. | Removal and Selection: Improving RGB-Infrared Object Detection via Coarse-to-Fine Fusion | |
| Demetrovics et al. | An efficient method to reduce the size of consistent decision tables | |
| CN108428018B (en) | Dimension-variable progressive gray prediction method for short-circuit current peak value |
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 |