CN104902220A - Equipment running state cloud monitoring device based on improved GMM (Gaussian Mixture Model) method - Google Patents
Equipment running state cloud monitoring device based on improved GMM (Gaussian Mixture Model) method Download PDFInfo
- Publication number
- CN104902220A CN104902220A CN201410743853.4A CN201410743853A CN104902220A CN 104902220 A CN104902220 A CN 104902220A CN 201410743853 A CN201410743853 A CN 201410743853A CN 104902220 A CN104902220 A CN 104902220A
- Authority
- CN
- China
- Prior art keywords
- equipment
- cloud
- running state
- personnel
- video
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses an equipment running state cloud monitoring device based on an improved GMM (Gaussian Mixture Model) method, which is mainly used for fulfilling the aims of background real-time monitoring and inquiry of machine faults and equipment running capability by certain leasing companies and equipment operation companies. Meanwhile, the aim of monitoring the movement of personnel and the running of equipment can be fulfilled, so that the job responsibility of the personnel, the running state of the equipment and fault evaluation can be judged. Meanwhile, a parallel computation framework based on an SOA (service-oriented architecture) is designed and implemented on the background to complete different video intelligent analysis tasks in order to fulfill the aim of remotely monitoring the working states of the equipment and personnel. A solution is provided, and the improved GMM and optimization algorithm in image processing is adopted for the equipment running state and a personnel identification part, so that the identification accuracy is increased while the system timeliness is ensured, and the working efficiency is increased.
Description
technical field:
Patent of the present invention relates to the equipment running status cloud supervising device based on improving GMM method.
background technology:
Design and implementation completes different video intelligent analysis task based on the parallel computation frame of SOA framework.This platform utilizes the cloud computing system based on SOA framework to realize massive video intellectual analysis, the great market demand of video analysis can be aimed at, make up the blank in this field, following cloud computing, cloud store and large intelligent data analysis market competition in occupy a tiny space.
summary of the invention:
1. based on the cloud vision platform of SOA framework
Services Oriented Achitecture (service-oriented architecture, SOA) is a component model, and it defines good interface between being served by these by the different function units (being called service) of application program and contract connects.Interface adopts neutral mode to carry out defining, and it should independent of hardware platform, operating system and the programming language realizing serving.This makes the service be structured in various such system can carry out alternately with a kind of unification and general mode.
Adopt the cloud vision framework based on SOA as shown in Figure 1.This framework for core, comprises data acquisition module, control desk module, file data memory module, picture recognition module and business diagnosis module composition with Service Bus.Wherein data acquisition module is divided into terminal control unit collection and terminal control unit agent acquisition two kinds of patterns, to tackle the situation that can remote camera directly access; Control desk module is for the treatment of API, timer-triggered scheduler, manual situation such as triggering (UI) and failed operation process etc.; File data stores based on HDFS framework, realizes file read-write by http mode.
1. the equipment running status based on mixed Gauss model background modeling is monitored
Gauss model accurately quantizes things with Gaussian probability-density function exactly, a things is decomposed into some models formed based on Gaussian probability-density function.Image background is set up to principle and the process of Gauss model: image grey level histogram reflection be the frequency that in image, certain gray value occurs, also can think it is the estimation of gradation of image probability density.
Mixed Gauss model uses K Gauss model to carry out the feature of each pixel in token image, mixed Gauss model is upgraded after a new two field picture obtains, mate with mixed Gauss model with each pixel in present image, if success, judge that this point is as background dot, otherwise be foreground point.Take an overall view of whole Gauss model, it mainly has variance and average two parameters to determine, to the study of average and variance, take different study mechanisms, will directly have influence on the stability of model, accuracy and convergence.
The probability density function that mixed Gauss model provides is actually the weighted sum of several Gaussian probability-density function.Each GMM is distributed by K Gaussian and forms, and each Gaussian is called one " Component ", and the linear addition of these Component just constitutes the probability density function of GMM together:
（1）
According to formula above, if we will get a point randomly from the distribution of GMM, in fact can be divided into two steps: among this K Component, select one first randomly, in fact the selected probability of each Component is exactly its coefficient π
k, after have selected Component, then it is just passable to consider from the distribution of this Component, choose a point individually.Suppose there is now N number of data point, we think that these data points are produced by certain GMM model, and we will need to determine π now
k, μ
k, σ
kthese parameters.Very natural, we expect utilizing maximal possibility estimation to determine these parameters, and the likelihood function of GMM is as follows:
(2)
Add owing to having again inside logarithmic function and, we the direct differentiate way of solving an equation directly cannot try to achieve maximum.In order to address this problem, we utilize EM algorithm, distribution iteration try to achieve maximum, and the value of parameters when acquisition obtains maximum.Specifically can be divided into the following steps:
1) initiation parameter π
k, μ
k, Σ
k, a kind of popular way first carries out cluster by K-means algorithm to data point, according to the initial value of cluster result Selecting All Parameters.
2) E step: probability that data estimator is generated by each Component (be not each Component selected probability): for each data x
i, the probability that it is generated by a kth Component is
(3)
But, due to the π in formula
k, μ
k, Σ
kneed the parameter that we estimate just, we adopt iterative method.
3) M step: differentiate is carried out to formula (3), obtains the parameter value corresponding to maximum likelihood:
(4)
(5)
4) calculate the value (formula (2)) of likelihood function, check whether likelihood function restrains.If restrained, illustrate that likelihood function obtains maximum, the value that now parameter is corresponding is the maximal possibility estimation of each parameter.Otherwise then iteration carries out E step, M step.
accompanying drawing illustrates:
Fig. 1 is based on the cloud vision overall framework of SOA; The cloud vision hardware that Fig. 2 gives based on SOA connects block diagram; Fig. 3 is the connection block diagram of data memory module.Fig. 4 utilizes mixed Gauss model background modeling to obtain equipment running status monitored results;
specific implementation method:
In Fig. 4, red rectangle frame is the moving object detected in scene, is used for the running status of marking equipment.Lower left side " Work " " Idle " of every frame scene is used for indication equipment is to run or idle.Red font is running status, and green font is idle state.Time is below the data processing time of the every frame in forms data source, and average out to 16ms, this means, a unit can support real-time, the concurrent processing of 10 data sources simultaneously.Because cloud platform adopts time-division processing mechanism, assuming that access in an every 5 minutes data source, cloud vision has the concurrent processing ability that 120 data sources supported by unit.
Above-mentioned video intelligent analysis module is applied to the performance verification of the cloud vision based on SOA framework, there is following characteristic:
1) more easy care
The loose couplings relation of business service supplier and business service user and the employing of open standard be ensure that to the realization of this characteristic.Be based upon with the information system on SOA basis, when demand changes time, do not need to revise the interface providing business service, only need adjustment business service flow process or retouching operation, whole application system is also more easily maintained.
2) higher availability
This feature is that the loose couplings of ISP and service user closes to fasten to be played and embody.The tool that user need not understand supplier not realizes details.
3) better retractility
The framework model relying on business service design, develop and field etc. to adopt realizes retractility.ISP can be adjusted mutually independently of one another, to meet new demand for services.
Embodiment:
1. based on an equipment maintenance device for cloud computing, its composition comprises: the algoritic module of the equipment maintenance device of video camera, cloud vision is embedded in DSP module, watch-dog and acousto-optic alarm.It is characterized in that: in real time the equipment video information collected is sent in the ROM of equipment maintenance device of cloud vision by the communication of RJ45 network interface by supervision video camera, and by the algorithm be embedded in DSP module, the data collected are carried out image processing and analyzing, the idle determination of equipment is shown on a monitor, and just can Real Time Observation data message online by the mode of SMS or mobile terminal, can remote supervisory equipment, reduce a large amount of human and material resources.Design and implementation completes different video intelligent analysis task based on the parallel computation frame of SOA framework.
2. the equipment maintenance device based on cloud vision platform according to patent requirements 1, it is characterized in that: the realtime graphic arrived by camera acquisition, the Algorithm Analysis of image procossing is carried out in DSP, by the Gaussian Background modeling technique improved, whether interpretation is realized to equipment operation, and monitored in real time now by wireless and wired two kinds of patterns. utilize the cloud computing system based on SOA framework to realize massive video intellectual analysis, the great market demand of video analysis can be aimed at, make up the blank in this field, in following cloud computing, occupy a tiny space in the market competition of cloud storage and large intelligent data analysis.
Claims (5)
1. based on the equipment running status cloud supervising device improving GMM method, its composition comprises: algoritic module, watch-dog and acousto-optic alarm that the equipment of video camera, the cloud vision be embedded in DSP module is supervised.
2. it is characterized in that: in real time the equipment video information collected is sent in the ROM of equipment maintenance device of cloud vision by the communication of RJ45 network interface by supervision video camera, and by the algorithm be embedded in DSP module, the data collected are carried out image processing and analyzing, the free time of equipment is judged, and show on watch-dog or manual terminal, and can Real Time Observation data message online by the mode of SMS or mobile terminal, can remote supervisory equipment, reduce a large amount of human and material resources.
3. Design and implementation completes different video intelligent analysis task based on the parallel computation frame of SOA framework.
4. the equipment running status cloud supervising device based on improvement GMM method according to patent requirements 1, it is characterized in that: the realtime graphic arrived by camera acquisition, the Algorithm Analysis of image procossing is carried out in DSP module, by the Gaussian Background modeling technique improved, whether interpretation is realized to equipment operation, and monitored in real time online by wireless and wired two kinds of patterns, and realize sound and light alarm at server end.
5. utilize the cloud computing system based on SOA framework to realize massive video intellectual analysis, the great market demand of video analysis can be aimed at, make up the blank in this field, following cloud computing, cloud store and large intelligent data analysis market competition in occupy a tiny space.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410743853.4A CN104902220A (en) | 2014-12-09 | 2014-12-09 | Equipment running state cloud monitoring device based on improved GMM (Gaussian Mixture Model) method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410743853.4A CN104902220A (en) | 2014-12-09 | 2014-12-09 | Equipment running state cloud monitoring device based on improved GMM (Gaussian Mixture Model) method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104902220A true CN104902220A (en) | 2015-09-09 |
Family
ID=54034584
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410743853.4A Pending CN104902220A (en) | 2014-12-09 | 2014-12-09 | Equipment running state cloud monitoring device based on improved GMM (Gaussian Mixture Model) method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104902220A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110263811A (en) * | 2019-05-21 | 2019-09-20 | 上海应势信息科技有限公司 | A kind of equipment running status monitoring method and system based on data fusion |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101093603A (en) * | 2007-07-03 | 2007-12-26 | 北京智安邦科技有限公司 | Module set of intellective video monitoring device, system and monitoring method |
| CN101179710A (en) * | 2007-11-30 | 2008-05-14 | 浙江工业大学 | Intelligent video monitoring apparatus of railway crossing |
| CN102663358A (en) * | 2012-03-29 | 2012-09-12 | 山西省电力公司晋中供电分公司 | Video image identification method of operating state of secondary protection equipment of substation |
-
2014
- 2014-12-09 CN CN201410743853.4A patent/CN104902220A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101093603A (en) * | 2007-07-03 | 2007-12-26 | 北京智安邦科技有限公司 | Module set of intellective video monitoring device, system and monitoring method |
| CN101179710A (en) * | 2007-11-30 | 2008-05-14 | 浙江工业大学 | Intelligent video monitoring apparatus of railway crossing |
| CN102663358A (en) * | 2012-03-29 | 2012-09-12 | 山西省电力公司晋中供电分公司 | Video image identification method of operating state of secondary protection equipment of substation |
Non-Patent Citations (2)
| Title |
|---|
| 张龙等: "基于TESPAR与GMM的滚动轴承性能退化评估", 《仪器仪表学报》 * |
| 王恒等: "基于KPCA-GMM的球磨机状态监测与评估研究", 《矿山机械》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110263811A (en) * | 2019-05-21 | 2019-09-20 | 上海应势信息科技有限公司 | A kind of equipment running status monitoring method and system based on data fusion |
| CN110263811B (en) * | 2019-05-21 | 2021-03-26 | 上海应势信息科技有限公司 | Equipment running state monitoring method and system based on data fusion |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107566781B (en) | Video monitoring method and video monitoring equipment | |
| US20190080274A1 (en) | Tracking and/or analyzing facility-related activities | |
| CN108255605A (en) | Image recognition cooperative computing method and system based on neural network | |
| CN109657564A (en) | A kind of personnel detection method, device, storage medium and terminal device on duty | |
| CN103477352A (en) | Gesture recognition using depth images | |
| CN108009497A (en) | Image recognition monitoring method, system, computing device and readable storage medium storing program for executing | |
| DE112021002820T5 (en) | DYNAMIC AUTOMATION OF A SELECTION OF PIPELINE ARTIFACTS | |
| CN109962789A (en) | Method and apparatus based on network data construction working application label system | |
| CN110991871A (en) | Risk monitoring method, device, equipment and computer readable storage medium | |
| CN117389742B (en) | Edge computing method, device and storage medium for machine vision | |
| CN117240887B (en) | Wisdom thing networking energy management platform system | |
| CN112486788A (en) | Workload evaluation method and device, electronic equipment and readable storage medium | |
| CN109743286A (en) | A kind of IP type mark method and apparatus based on figure convolutional neural networks | |
| CN109284735A (en) | Mouse feelings monitoring method, device, processor and storage medium | |
| CN107277100A (en) | The system and method realized and managed near real-time cloud infrastructure strategy | |
| JP2023002768A (en) | Image processing method, device, electronic equipment, storage medium, and computer program | |
| WO2019062404A1 (en) | Application program processing method and apparatus, storage medium, and electronic device | |
| CN115329265A (en) | Method, device and equipment for determining graph code track association degree and storage medium | |
| US20090018813A1 (en) | Using quantitative models for predictive sla management | |
| CN106502856B (en) | A kind of information processing method and device | |
| CN110390226A (en) | Crowd's event recognition method, device, electronic equipment and system | |
| CN110781950B (en) | Message processing method and device | |
| CN104902220A (en) | Equipment running state cloud monitoring device based on improved GMM (Gaussian Mixture Model) method | |
| Wen et al. | Cloud-computing-based framework for multi-camera topology inference in smart city sensing system | |
| Tundo et al. | An energy-aware approach to design self-adaptive ai-based applications on the edge |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20150909 |