CN107145851A - Constructions work area dangerous matter sources intelligent identifying system - Google Patents
Constructions work area dangerous matter sources intelligent identifying system Download PDFInfo
- Publication number
- CN107145851A CN107145851A CN201710292232.2A CN201710292232A CN107145851A CN 107145851 A CN107145851 A CN 107145851A CN 201710292232 A CN201710292232 A CN 201710292232A CN 107145851 A CN107145851 A CN 107145851A
- Authority
- CN
- China
- Prior art keywords
- work area
- matter sources
- dangerous matter
- constructions work
- intelligent
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/254—Analysis of motion involving subtraction of images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/255—Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
- H04N7/181—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast for receiving images from a plurality of remote sources
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/467—Encoded features or binary features, e.g. local binary patterns [LBP]
Abstract
The invention discloses a kind of constructions work area dangerous matter sources intelligent identifying system, characterized in that, mainly including a wide range of operation area scene information obtains foreground target and ROI extraction systems, dangerous matter sources INTELLIGENT IDENTIFICATION system, early warning issue response and interactive system under system, complex background.It is mainly used in dangerous matter sources Intelligent Recognition in constructions work area, and the main operating personnel completed in constructions work area screens, safety cap wears detection and invades limit problem determination.On the one hand authentic and valid video data is provided to administrative staff, on the other hand mitigate the difficulty and cost of comprehensive supervision job site, while its fully-automatic intelligent completes identification, have wide range of applications, and compatible existing monitoring system to a certain extent, greatly save system operation cost.
Description
Technical field
, can be in complicated construction the present invention relates to a kind of constructions work area dangerous matter sources intelligent identifying system of view-based access control model method
Main hazard source under operating environment in Intelligent Recognition operation area, belongs to vision measurement field.
Background technology
In recent years, with the continuous propulsion of Chinese Urbanization, how being continuously increased for various infrastructure constructions is strengthened applying
The safety management at work scene becomes unit in charge of construction, the Important Problems of each side of government department concern.At present, China job site is pacified
The main cause of full Frequent Accidents has:
(1)Job site environment is complicated, it is outdoor work under hard conditions, operation is various, work is wide, technique intersects complicated, safe skill
Art involves a wide range of knowledge, and result in job site and includes diversified potential safety hazard;
(2)Workmen's quality is uneven, awareness of safety is thin, and quantity is more, source is complicated, mobility is big, lacks necessary
Safety training, add the safety management difficulty of job site;
(3)Job site is often in semi-enclosed environment, and unrelated person, which is swarmed into, causes the possibility of security incident larger;
(4)At present, the security monitoring to job site is nearly all responsible for by safety manager, it is difficult to realize comprehensive, real-time
Monitoring, monitoring efficiency is low.
With the continuous growth of the network bandwidth, the development of computer technology, the raising of image processing techniques, video monitoring skill
Art also has significant progress, and video surveillance applications penetrate into security protection more and more widely, educates, amusement, the various fields such as medical treatment.
The visual monitor system of current commercialization is mainly used in scene recording and transmission of video, but with image procossing and pattern-recognition
The continuous progress of technology, the intelligent visual surveillance system designed according to real needs is increasingly widely used.Intelligent vision
Monitoring system is not only able to catch the target in monitoring scene, and can carry out discriminance analysis to a certain extent to target,
Car plate detection, pedestrian detection in such as intelligent transportation, target following.
Because intelligent visual surveillance system can provide monitoring in real time, and automatically analyze shooting incessantly in 24 hours
The data that capture of head, alarm etc. is sent to safety manager according to analysis result in time, this just with job site
Safety management demand is consistent.It is that non-operating personnel enters operation for the principal element for triggering security incident in current building trade
The not good safe wearing equipment of area, operating personnel and foreign body intrusion(Specific region workmen or operation are set in work progress
It is standby to enter region into non-permitted)Problem, with reference to smart camera correlation technique, a kind of design constructions work area dangerous matter sources are intelligently known
Other system, is monitored in real time using this system to the safe condition of job site and the behavior of associated construction personnel, on the one hand
Authentic and valid video data is provided to administrative staff, on the other hand mitigates the difficulty and cost of comprehensive supervision job site.
Therefore set up a set of significant suitable for the dangerous matter sources intelligent identifying system in construction operation area.
The content of the invention
In order to solve the above problems, for the demand of safety monitoring in constructions work area, with reference to current software and hardware condition,
View-based access control model sensor is studied, the identification of potential safety hazard in constructions work area, early warning is realized and is published to each terminal.For convenience of area
Divide and be referred to as Zhong Kong centers with description, the part that control whole system is run, the warning information that Zhong Kong centers are sent will be received
Each several part is referred to as terminal.
Technical scheme is as follows:
Constructions work area dangerous matter sources intelligent identifying system, is mainly comprised the steps of:
(1)The large-scale scene visual information in constructions work area is obtained by distributed group net mode, and carried out in each detail point
Image information is pre-processed, and will after processing image transmitting to Zhong Kong centers;
(2)For step(1)The scene visual information got, in constructions work area complex background, according to the danger that need to be recognized
The difference in dangerous source, intelligent extraction goes out foreground target and ROI interested(Region of Interest)Region;
(3)According to the information in the foreground target information and ROI region of acquisition, image procossing and machine learning related side are utilized
Method, intelligently completes the identification in hazard source in the operation area;
(4)For hazard source in the scene that has been acknowledged, remote management platform, Zhong Kong centers, enforcement division are distributed to
The terminals such as door.
A kind of constructions work area dangerous matter sources intelligent identifying system of use above method, the system includes:
(1)A wide range of constructions work area image information acquisition system:The system is mainly the dangerous matter sources recognized as needed not
Together, constructions work zoning is divided into multiple subregions, all subregion is obtained in the subregion by wireless routing networking mode pacifies
The image information for each camera put, the image information of all subregion is transmitted to middle control platform by optical fiber cable, and middle control platform leads to
Cross full gigabit switch and all subregion image information is sent to image processing equipment(Image workstation);
(2)Foreground target and ROI extraction systems under complex background:The system mainly by image processing method, is made in building
Under the complex background in industry area, foreground target information and ROI region are extracted, interference of the background information to dangerous identifing source is eliminated;
(3)Dangerous matter sources INTELLIGENT IDENTIFICATION system:The system is mainly the intelligent personnel's examination that fulfils assignment, safety cap and wears detection and invade
Limit problem;
(4)Early warning issue response and interactive system:The part be mainly the information that confirms is sent to remote management platform, in
The terminals such as control center, executive arm, and well carry out corresponding information interworking.
Beneficial effects of the present invention, the present invention is for the dangerous matter sources Intelligent Recognition in constructions work area, on the one hand to management
Personnel provide authentic and valid video data, on the other hand mitigate the difficulty and cost of comprehensive supervision job site, while its
Fully-automatic intelligent completes identification, has wide range of applications, and compatible existing monitoring system to a certain extent, greatly saves system fortune
Seek cost.
Brief description of the drawings
Fig. 1 is the general structure schematic diagram of constructions work area dangerous matter sources intelligent identifying system of the present invention;
Fig. 2 is that dangerous matter sources intelligent identifying system distributed networking in constructions work area of the present invention obtains scene information schematic diagram;
Fig. 3 is the solid geometric pattern schematic diagram that dangerous matter sources intelligent identifying system binocular camera in constructions work area of the present invention is set up.
Embodiment
Technical scheme is described in more detail with reference to the accompanying drawings and detailed description.
Shown in reference picture 1, Fig. 2 and Fig. 3, constructions work area dangerous matter sources intelligent identifying system, including a wide range of constructions work
Foreground target and ROI extraction systems, dangerous matter sources INTELLIGENT IDENTIFICATION system, early warning hair under area's image information acquisition system, complex background
Cloth is responded and interactive system.
(1)A wide range of constructions work area image information acquisition
Needing to carry out the fixed position placement camera in the scene of safety precaution, the camera number and installation site of placement first
And setting angle need to be determined according to actual conditions, IMAQ or video acquisition are carried out to the scene for needing to monitor.But build
Build regional extent too big, operation area is then divided into different subregions, the camera of all subregion is obtained by wireless routing should
The image information of each camera in subregion, while distinct device is numbered, and is carried out image information in the branch node
The pretreatments such as filtering, denoising, contrast enhancing, reduce transmitted data amount.Because each region is indefinite from a distance from middle control platform, and
And disturbing factor is numerous, then the image information after each regional processing is transmitted to middle control platform by optical fiber cable mode, in
Each regional image information is sent to image processing equipment by control platform by full gigabit switch again(Image workstation).
(2)Foreground target and ROI are extracted under complex background
For triggering the principal element of security incident not good into operation area, operating personnel for non-operating personnel in building trade
Safe wearing measure and foreign body intrusion problem, and problems, are extracted respective from the complex background in constructions work area
Foreground target is the basis of identification of dangerous source.For the foreground target extracted required for dangerous identifing source in constructions work area, nothing
Non- is personnel and implement.The requirement managed according to site safety, the staff into operation area must whole body wearing system
One utility uniform provided, therefore whether work clothes is worn by testing staff, to be carried out to the illegal personnel for entering construction area
Screen, its foreground target is the personnel for entering operation area.Similarly, for the operating personnel into operation area, it need to wear as requested
Wear a safety helmet, therefore its foreground target is also personnel.And for invading limit problem, be divided into implement and invade limit and invade limit with personnel, before it
Scape target is implement and operating personnel.
1)Personnel's foreground extraction
Under the complex background of constructions work scene, because target and background is likely that there are similar appearance information and not bright
Aobvious border, individually uses the segmentation based on spatial domain(Foreground extraction)Technology will produce larger detection error;And when being based on
The cutting techniques in domain make use of inter-frame information, can be used to the pixel for judging to be changed on a timeline, but be due to time domain
Cutting techniques are poor for the Detection results of some changes, and such as fortune is not present in some continuous frames in obvious, target to texture
It is dynamic or change, individually using the cutting techniques based on time domain it is possible that target area is discontinuous, inside has cavity etc.;Cause
This, not only make use of the continuity of Moving Objects in time based on the united cutting techniques of space-time, also use the sky of image
Between cutting techniques, on the basis of frame in does still image segmentation, be aided with timeline information, segmentation result can be made more to manage
Think.
Wherein, background modeling method performance in the moving target of the outdoor complex condition of detection is best, Non-parameter modeling
For MAP-MRF frameworks are compared with the mixed Gauss model of parameter model, precision is more preferably but speed is excessively slow.Frame differential method speed is most
It hurry up, but Detection results are poor, be only applicable to requirement of real-time very high occasion.Consider accuracy of detection and real-time will
Ask, and mixed Gauss model, in the outstanding behaviours of background modeling aspect, this proposes a kind of based on the united mesh of space time information
Detection algorithm is marked, i.e., in time-domain information(Mixed Gauss model)On the basis of increase spatial-domain information(Color Statistical histogram)Come
Improve verification and measurement ratio.
Next it is typically that target to be measured is been described by after target to be identified is partitioned into using object detection method
With expression, original pixels are made to be converted to certain form for being more suitable for further calculating processing.HOG features can fully extract the shape of people
Shape information and appearance information, possess and distinguish people and other targets or the distinguishing ability of background.The characteristics of LBP features is to calculate high
Effect, identification are strong and have consistency to dull gray level change.This proposes the mode for combining HOG and LBP features, that is, exists
LBP histograms are extracted in small lattice, the HOG features and LBP histograms in window are combined, also just combine shape facility and
Textural characteristics, can be obviously improved detection performance.When noise edge composition is more in background, the performance of HOG features is not enough, utilizes
LBP consistent sexual norm can filter out this noise like, make up HOG this defect, and specific practice is by each consistent sexual norm pair
It should arrive histogram, and all nonuniformity patterns are all classified as a post, then all LBP histograms connected in window are i.e. complete
The description of texture, generates the LBP features of window, obtains HOG+LBP finally joint group with HOG feature strings altogether in paired window
Feature.
2)Implement foreground extraction
Some specific region workmens and implement do not allow access into specific region in work progress, and such as railway station is constructed
Railway region, high voltage power transmission networking, the operation road in bridge erection engineering etc..In these regions, mechanical rotary head, reversing or
Fouling of clearance gauge can cause great security incident during intersection, and construction machinery slides the security incident caused to Business Line and is all
Belong to machinery and invade limit accident, therefore can be used uniformly in method to invade limit scope based on binocular stereo vision and divide and invade limit and sentence
Not.Extracted to entering the equipment in restricted area, just can well determine whether to invade limit.
Due to implement feature in the picture(Color characteristic, textural characteristics, edge feature, linearity etc.)It is relatively solid
It is fixed, it can be separated from background pixel by partitioning algorithm, the judgement on limit border is invaded in subsequent step on the basis of it,
Therefore implement is carried out to the basis that segmentation is subsequent step from background.First method can be split by threshold value,
Edge feature detection is such as carried out, the determination of edge feature can be detected by relevant edge operator, be detected by boundary operator
Implement edge out includes substantial amounts of noise, it is therefore desirable to carry out denoising to it, the step can pass through connected region
Domain detects, edge contour area, and the operation such as corrosion carries out noise reduction process, it can be split by region method for second,
3rd class can be by carrying out contours segmentation based on movable contour model method to it.
(3)Identification of dangerous source
1)Operating personnel screens
According to the requirement of constructions work safety management, the necessary whole body of staff into operation area dresses the work of unified payment
Uniform, by whether wearing work clothes, whether the personnel that just can be screened into operating area are staff.If do not worn
Work clothes, system then points out the target to be probably the non-working person into operation area, reminds relevant departments to confirm it
And do corresponding processing.When there is personnel to enter in monitoring visual field, system can be extracted by Background difference discussed above
Foreground target, it is contemplated that the work clothes in building site has a unified color standard, thus can by the color to motion target area and
Template is matched, and discriminates whether to wear work clothes.
2)Safety cap wears detection
Safety cap as a kind of personal cephalic protection articles for use, can effectively prevent and mitigate operating personnel in production operation by
Dropped object or injury when falling certainly to head.Workmen's safe wearing cap is a kind of necessary safety in constructions work area
Measure, the personal safety of workmen can be ensured to a certain extent.
By a frame video camera as monitor area gateway, adjust its angular field of view and cover whole gate area.Work as prison
The each two field picture obtained when control equipment starts monitoring to camera draws binaryzation prospect using Background difference and binaryzation
Figure, and the pretreatment operations such as medium filtering, gaussian filtering, burn into expansion are carried out to it, suppress ambient noise as much as possible with before
Scape noise, accurately to obtain foreground target;Then the prospect produced will be moved in sequence image by moving object detection to become
Change region to extract from background, it is on the scene to obtain the size and shape people that is consistent according to size filter and shape filtering wherein
The target of feature in scape, it is believed that be the target for representing workmen, finds out the boundary rectangle of its profile, and in former frame all generations
Look for whether to be more than certain proportion and the two difference in areas with its boundary rectangle overlapping area in the moving target of table workmen
The different target in the range of given threshold, it is same target that the two is thought if having, and it is tracked and marked, if without if
It is considered fresh target, assigns its new trace labelling.Moving target is traced, and its information can also be transmitted therewith, including
Position that the target occurs first, the target enters or gone out, the target connects the information such as the color of rectangle frame.
After obtaining commissarial moving target, to avoid the interference of redundancy, it is contemplated that the position of safe wearing cap
For the head of people, i.e., commissarial moving target scope leans on tip portion, so selection detects safety cap in the part.Experiment
Show, take the commissarial region of moving target top 1/3 more suitable, can farthest reduce amount of calculation.Therefore, in monitoring
During, the boundary rectangle to reaching the commissarial significant notation moving target between two detection lines is drawn and takes its top 1/3
Region is used as the potential region of safety cap.
In pixel of the latent chromatic value of searching in the zone of safety cap in preset security cap color gamut, and with its color
Corresponding numbering is that these pixels are marked.The profile of the point marked with same color is found out, cast out wherein shape,
Size, dispersion do not meet the profile of safety cap profile predetermined threshold value, remaining, are considered as the profile of safety cap, and according to
The color mark at its midpoint obtains the color mark of the safety cap, that is, thinks that the workman has worn the safety cap of corresponding color.If
The profile of safety cap can be represented by not obtaining any one wherein, then it is assumed that not comprising safety cap in the moving target, i.e.,
This person does not wear a safety helmet, and sends safety cap wearing early warning.
3)Invade limit
For invading limit problem, it is necessary first to it is determined that invading limit scope, limit model is invaded as set forth above, it is possible to be divided using binocular stereo vision
Enclose and invade limit and differentiate.Binocular Stereo Vision System platform by video camera × 2, optical lens × 2, image pick-up card, computer and
Scaling board is constituted.Parallax range between left and right cameras is 10cm, and binocular camera intrinsic parameter and outer parameter have been demarcated first, and
Keep camera parameter constant.Scaling board image is used for the inside and outside ginseng demarcation of video camera, and the solid geometric pattern that binocular camera is set up is such as
Shown in Fig. 3.
Invade limit scope and determine that flow can be divided into:One is border detection(The rail of railway, the borderline region of Ordinary Work,
Or by spilling the artificial restrained boundary such as pulverized limestone), the part can be by rim detection, and the method such as connected region detection extracts side
Edge;Two be to divide to invade limit scope, and this part can obtain two-dimension picture using binocular camera and incline line index point three-dimensional coordinate, and foundation is inclined
The plane equation of line scope;Three be implement detection, and the part is mainly moving Object Detection(Exclude the fire that passed through on railway
It can be excluded on car, operation road normal through the normal devices such as vehicle, such method by its geometric properties), detection
During may relate to multiple Moving target detections.It is contemplated that with background subtraction, the method detection such as Stereo matching.
Because implement to be determined is in the distance of space length limited boundary, therefore only it can not obtain it with a camera
Three-dimensional coordinate.The division of limit scope is invaded it is contemplated that binocular camera is determined.Camera is angled to be set up, parallel to limited boundary side
To being shot, by binocular stereo vision principle, the two dimensional image of two cameras in left and right can just uniquely determine viewing field of camera
The three-dimensional information at any point in space.Camera is set accurately to react three of plant equipment and limited boundary in X-Y scheme
Dimension coordinate.Binocular camera demarcation is carried out first, obtains the inner parameter and the position relationship between them of two video cameras.So
After can according to disparity map carry out three-dimensional reconstruction, obtain the three-dimensional coordinate of extraterrestrial target.Main include is set up camera model and asked
The step of camera parameter two is solved, suitable camera calibration method is selected for various specific cameras, its method is also not quite similar.Determine phase
After machine imaging model, by setting up the transformational relation of the world coordinates of arbitrfary point and two dimensional image pixel coordinate in view field space,
And solve its parameter.So that it is determined that plant equipment and border three-dimensional coordinate.The division for invading limit scope is need to determine up and down four
The equation of individual boundary plane.Point coordinates can be demarcated by extracting boundary perimeter, three-dimensional side is determined according to multiple demarcation point coordinates
Boundary plane equation.The three-dimensional coordinate that the determination on border is exactly multiple coordinate points by having determined to determine, and pass through these
Coordinate points are fitted the left and right plane equation perpendicular to road surface, then determine line scope of inclining by left and right plane equation.
Plant equipment be constructions work area mainly invade limit object, mainly there are two class methods to detect plant equipment,
Individually it is split, it is extracted from background and is tracked, partitioning algorithm refers to foreground target extraction step,
Tracing algorithm refers to Kalman's tracking, light stream, template matches, Camshift scheduling algorithms.Part detection is main still using double
Mesh obtains mobile object three-dimensional coordinate and judges whether the target touches line, can specifically be divided into three parts, and one is that limited boundary is known
Not, two be plant equipment identification and tracking, three be limited boundary and plant equipment Distance Judgment.Border detection mainly passes through instruction
Practice corresponding grader it is identified, its training characteristics considers its color characteristic(Color moment), textural characteristics(LBP,
HOG), geometric properties etc..Border is judged with plant equipment Distance Judgment by binocular camera.Mechanically moving equipment invades limit inspection
Survey and detection of classifier is individually individually carried out to border and plant equipment first, then using in different detection of classifier video flowings
Plant equipment and limited boundary, and border and plant equipment three-dimensional coordinate are obtained by binocular camera in real time, finally calculate both
Space length carry out early warning differentiation.
Personnel invade limit and are divided into two kinds, and one kind is that personnel are appeared under mechanical arm and its in the range of radius of turn, and the problem is needed
Want camera frame on mechanical arm, shooting angle is downward, by detecting that whetheing there is personnel in viewing field of camera enters under mechanical arm.It is another
It is that personnel enter off-limits specific region in operation process to plant, and the problem is mainly invaded limit by the said equipment, passed through first
Binocular, which is obtained, limits scope, then whether there is personnel in the range of detection restriction.Two kinds of main solutions of problem are all pedestrian detections.
Pedestrian detection scheme can be detected by pedestrian's textural characteristics such as HOG features.
The brief handling process of Human detection is:First, original video data is gathered by vision sensor, system is used as
Input information;Secondly, in order to improve picture quality in order to subsequent treatment, noise, gray-scale map are removed by pretreatment module
As treatment technology etc., make the image after conversion closer to true picture;Again, module of target detection is provided using video sequence
Effective information, the area-of-interest in image is split;Then, characteristic extracting module is to the area-of-interest that detects
Initial data enter line translation, obtain a computation complexity low and representative(So that difference is as small as possible in class, between class
Difference is as larger as possible)Characteristic vector the target is described;Then classified in feature space with the grader trained
Identification, the design of grader needs to miss by selection discriminant function, renewal discriminant function and detection repeatedly on training sample set
Three steps of rate, it is contemplated that the real-time of application, designed grader needs to take into account verification and measurement ratio and detection speed;Finally,
Early warning is carried out after being carried out according to the demand of application to the target detected.
(4)Early warning issue response and interaction
As shown in figure 1, according to the dangerous identifing source needs in constructions work area, many cameras to be respectively disposed on to the fixed bit in scene
Put and IMAQ or video acquisition are carried out to scene, and based on VS (Visual Studio) development environment, programmed by C# language
Each several part algorithm is realized, so that the image or video data of collection are handled and corresponding result is returned,
The secondary development by C# language programming realization to Revit softwares, generates corresponding plug-in unit to therein three under Revit simultaneously
Dimension model of place is called and changed, then by Fuzor softwares realize to the rendering of scene, roam and multiple terminals hair
Cloth.
The system is needing to carry out the fixed position placement camera in the scene of safety precaution first, to the field for needing to monitor
Scape carries out IMAQ or video acquisition, and the image or video information that camera is gathered are crossed network and be back to after computer progress
Continuous processing and analysis, the camera number and installation site and setting angle of placement need to be determined according to actual conditions.In Revit
Threedimensional model corresponding with whole scene is set up under software, wherein the camera model of fixed position in scene is included, and to each phase
Machine model is numbered.Based on VS development environments, corresponding feature card is generated under Revit softwares by C# language, realized
Calling and change to threedimensional model, mainly realizes when occurring potential safety hazard in monitored picture, to the camera model in Revit
Triggered and recalled the monitored picture of correspondence camera.Because Revit softwares and Fuzor softwares can be associated, therefore can basis
There is the particular location of monitoring camera correspondence camera model under Revit softwares of the scene of potential safety hazard, determine that the camera exists
The correspondence position of threedimensional model under Fuzor softwares, when occur potential safety hazard when, system by the position according to camera in a model,
The scene location that the camera is monitored under automatic roaming to Fuzor softwares, and remote management department is distributed to by Fuzor, is held
Row department, and its information is published to Zhong Kong centers, to timely respond to.
The foregoing is only a preferred embodiment of the present invention, the application scope of application not limited to this of the present invention, appoints
What those familiar with the art is in the technical scope of present disclosure, the technical scheme that can be become apparent to
Simple change or equivalence replacement each fall within the present invention the application scope of application in.
Claims (3)
1. constructions work area dangerous matter sources intelligent identifying system, including a wide range of operation area scene information obtain system, complex background
Lower foreground target and ROI extraction systems, dangerous matter sources INTELLIGENT IDENTIFICATION system, early warning issue response and interactive system.
2. constructions work area dangerous matter sources intelligent identifying system according to claim 1, it is characterised in that the dangerous matter sources are distinguished
Know intelligent automation, and good combination BIM(Building Information Model)Technology.
3. constructions work area dangerous matter sources intelligent identifying system according to claim 1, it is characterised in that the monitoring system
In addition to limit regional monitoring system is invaded, the compatible existing monitoring system of remainder.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710292232.2A CN107145851A (en) | 2017-04-28 | 2017-04-28 | Constructions work area dangerous matter sources intelligent identifying system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710292232.2A CN107145851A (en) | 2017-04-28 | 2017-04-28 | Constructions work area dangerous matter sources intelligent identifying system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107145851A true CN107145851A (en) | 2017-09-08 |
Family
ID=59775210
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710292232.2A Pending CN107145851A (en) | 2017-04-28 | 2017-04-28 | Constructions work area dangerous matter sources intelligent identifying system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107145851A (en) |
Cited By (26)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106448202A (en) * | 2016-10-31 | 2017-02-22 | 长安大学 | Video based curve early warning system and early warning method |
CN107729842A (en) * | 2017-10-18 | 2018-02-23 | 中国石油大学(北京) | Oil-gas pipeline damage from third-party dangerous discernment method, apparatus and system based on machine vision |
CN107993389A (en) * | 2017-11-29 | 2018-05-04 | 湖南大学 | A kind of construction site safe early warning method and system based on BIM models |
CN108564069A (en) * | 2018-05-04 | 2018-09-21 | 中国石油大学(华东) | A kind of industry safe wearing cap video detecting method |
CN109147204A (en) * | 2018-08-16 | 2019-01-04 | 青海华职康健康科技有限公司 | A kind of automatic distribution method of personal protective equipment and system |
CN109255798A (en) * | 2018-09-18 | 2019-01-22 | 中国矿业大学 | Monitoring alarm method is used in a kind of safety in production |
CN109272197A (en) * | 2018-08-21 | 2019-01-25 | 吉林建筑大学 | A kind of identification of dangerous source method based on building BIM model |
CN109460719A (en) * | 2018-10-24 | 2019-03-12 | 四川阿泰因机器人智能装备有限公司 | A kind of electric operating safety recognizing method |
CN109557935A (en) * | 2018-11-01 | 2019-04-02 | 广州圣安环保科技有限公司 | A kind of safety monitoring during construction method and system based on unmanned plane |
CN110009800A (en) * | 2019-03-14 | 2019-07-12 | 北京京东尚科信息技术有限公司 | A kind of recognition methods and equipment |
CN110443976A (en) * | 2019-08-14 | 2019-11-12 | 深圳市沃特沃德股份有限公司 | Safety prompt function method, apparatus and storage medium based on safety cap |
CN111091535A (en) * | 2019-11-22 | 2020-05-01 | 三一重工股份有限公司 | Factory management method and system based on deep learning image semantic segmentation |
CN111416964A (en) * | 2020-05-09 | 2020-07-14 | 黄河勘测规划设计研究院有限公司 | Remote image intelligent identification method for hydraulic engineering deformation |
CN111445524A (en) * | 2020-03-31 | 2020-07-24 | 清华大学 | Scene understanding-based construction site worker unsafe behavior identification method |
CN112200043A (en) * | 2020-09-30 | 2021-01-08 | 中邮通建设咨询有限公司 | Intelligent danger source identification system and method for outdoor construction site |
CN112329584A (en) * | 2020-10-29 | 2021-02-05 | 深圳技术大学 | Method, system and equipment for automatically identifying foreign matters in power grid based on machine vision |
CN112465873A (en) * | 2020-12-10 | 2021-03-09 | 深圳市普汇智联科技有限公司 | Method and system for determining position of moving object in machine vision |
CN113392713A (en) * | 2021-05-20 | 2021-09-14 | 上海可深信息科技有限公司 | Work clothes wearing identification method |
CN113762171A (en) * | 2021-09-09 | 2021-12-07 | 赛思沃德(武汉)科技有限公司 | Method and device for monitoring safety of railway construction site |
CN113781655A (en) * | 2020-06-09 | 2021-12-10 | 国网新疆电力有限公司建设分公司 | Virtual reality simulation construction method and device, equipment and storage medium |
CN114296029A (en) * | 2021-12-24 | 2022-04-08 | 宁夏广天夏电子科技有限公司 | Method and system for positioning personnel on mining face |
CN114722445A (en) * | 2022-06-08 | 2022-07-08 | 山东高速德建集团有限公司 | Rivers and Dynamo-based dangerous engineering identification method |
CN114783144A (en) * | 2022-06-17 | 2022-07-22 | 深圳市易智博网络科技有限公司 | Intelligent building site safety monitoring method and device and computer equipment |
CN114979611A (en) * | 2022-05-19 | 2022-08-30 | 国网智能科技股份有限公司 | Binocular sensing system and method |
CN116152735A (en) * | 2023-01-09 | 2023-05-23 | 中国科学院空天信息创新研究院 | Security monitoring dangerous source identification system and method based on edge cloud computing |
CN117217547A (en) * | 2023-11-08 | 2023-12-12 | 安徽宜介科技股份有限公司 | Intelligent identification visual monitoring device and method for major hazard sources of construction engineering |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103108159A (en) * | 2013-01-17 | 2013-05-15 | 新疆电力公司乌鲁木齐电业局 | Electric power intelligent video analyzing and monitoring system and method |
CN103714325A (en) * | 2013-12-30 | 2014-04-09 | 中国科学院自动化研究所 | Left object and lost object real-time detection method based on embedded system |
CN104484524A (en) * | 2014-12-09 | 2015-04-01 | 张家驰 | Method and system based on BIM (building information modeling) simulation model and in-site construction interaction |
CN205005181U (en) * | 2015-09-16 | 2016-01-27 | 四川弘和通讯有限公司 | Video intelligent analysis identification system based on DM8168 |
CN205005179U (en) * | 2015-09-07 | 2016-01-27 | 安徽芜湖海螺建筑安装工程有限责任公司 | Long -range real -time monitoring system is built by laying bricks or stones to cement rotary kiln |
CN105334862A (en) * | 2015-10-28 | 2016-02-17 | 上海同筑信息科技有限公司 | BIM-based unmanned aerial vehicle monitoring method and system |
CN105404231A (en) * | 2016-01-12 | 2016-03-16 | 西北工业大学 | Internet of things-based intelligent building monitoring managing system |
CN105931411A (en) * | 2016-06-14 | 2016-09-07 | 广州东亚保安服务有限公司 | Firefighting remote monitoring and early warning platform of multistoried building and realization method of platform |
JP2017060052A (en) * | 2015-09-17 | 2017-03-23 | 株式会社日立ビルシステム | Video monitoring system |
-
2017
- 2017-04-28 CN CN201710292232.2A patent/CN107145851A/en active Pending
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103108159A (en) * | 2013-01-17 | 2013-05-15 | 新疆电力公司乌鲁木齐电业局 | Electric power intelligent video analyzing and monitoring system and method |
CN103714325A (en) * | 2013-12-30 | 2014-04-09 | 中国科学院自动化研究所 | Left object and lost object real-time detection method based on embedded system |
CN104484524A (en) * | 2014-12-09 | 2015-04-01 | 张家驰 | Method and system based on BIM (building information modeling) simulation model and in-site construction interaction |
CN205005179U (en) * | 2015-09-07 | 2016-01-27 | 安徽芜湖海螺建筑安装工程有限责任公司 | Long -range real -time monitoring system is built by laying bricks or stones to cement rotary kiln |
CN205005181U (en) * | 2015-09-16 | 2016-01-27 | 四川弘和通讯有限公司 | Video intelligent analysis identification system based on DM8168 |
JP2017060052A (en) * | 2015-09-17 | 2017-03-23 | 株式会社日立ビルシステム | Video monitoring system |
CN105334862A (en) * | 2015-10-28 | 2016-02-17 | 上海同筑信息科技有限公司 | BIM-based unmanned aerial vehicle monitoring method and system |
CN105404231A (en) * | 2016-01-12 | 2016-03-16 | 西北工业大学 | Internet of things-based intelligent building monitoring managing system |
CN105931411A (en) * | 2016-06-14 | 2016-09-07 | 广州东亚保安服务有限公司 | Firefighting remote monitoring and early warning platform of multistoried building and realization method of platform |
Cited By (34)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106448202A (en) * | 2016-10-31 | 2017-02-22 | 长安大学 | Video based curve early warning system and early warning method |
CN107729842A (en) * | 2017-10-18 | 2018-02-23 | 中国石油大学(北京) | Oil-gas pipeline damage from third-party dangerous discernment method, apparatus and system based on machine vision |
CN107993389A (en) * | 2017-11-29 | 2018-05-04 | 湖南大学 | A kind of construction site safe early warning method and system based on BIM models |
CN108564069A (en) * | 2018-05-04 | 2018-09-21 | 中国石油大学(华东) | A kind of industry safe wearing cap video detecting method |
CN108564069B (en) * | 2018-05-04 | 2021-09-21 | 中国石油大学(华东) | Video detection method for industrial safety helmet |
CN109147204A (en) * | 2018-08-16 | 2019-01-04 | 青海华职康健康科技有限公司 | A kind of automatic distribution method of personal protective equipment and system |
CN109272197A (en) * | 2018-08-21 | 2019-01-25 | 吉林建筑大学 | A kind of identification of dangerous source method based on building BIM model |
CN109272197B (en) * | 2018-08-21 | 2021-03-23 | 吉林建筑大学 | Danger source identification method based on building BIM model |
CN109255798A (en) * | 2018-09-18 | 2019-01-22 | 中国矿业大学 | Monitoring alarm method is used in a kind of safety in production |
CN109460719A (en) * | 2018-10-24 | 2019-03-12 | 四川阿泰因机器人智能装备有限公司 | A kind of electric operating safety recognizing method |
CN109557935A (en) * | 2018-11-01 | 2019-04-02 | 广州圣安环保科技有限公司 | A kind of safety monitoring during construction method and system based on unmanned plane |
CN110009800A (en) * | 2019-03-14 | 2019-07-12 | 北京京东尚科信息技术有限公司 | A kind of recognition methods and equipment |
CN110443976A (en) * | 2019-08-14 | 2019-11-12 | 深圳市沃特沃德股份有限公司 | Safety prompt function method, apparatus and storage medium based on safety cap |
CN110443976B (en) * | 2019-08-14 | 2021-05-28 | 深圳市沃特沃德股份有限公司 | Safety reminding method and device based on safety helmet and storage medium |
CN111091535A (en) * | 2019-11-22 | 2020-05-01 | 三一重工股份有限公司 | Factory management method and system based on deep learning image semantic segmentation |
CN111445524A (en) * | 2020-03-31 | 2020-07-24 | 清华大学 | Scene understanding-based construction site worker unsafe behavior identification method |
CN111416964A (en) * | 2020-05-09 | 2020-07-14 | 黄河勘测规划设计研究院有限公司 | Remote image intelligent identification method for hydraulic engineering deformation |
CN111416964B (en) * | 2020-05-09 | 2021-08-24 | 黄河勘测规划设计研究院有限公司 | Remote image intelligent identification method for hydraulic engineering deformation |
CN113781655A (en) * | 2020-06-09 | 2021-12-10 | 国网新疆电力有限公司建设分公司 | Virtual reality simulation construction method and device, equipment and storage medium |
CN112200043A (en) * | 2020-09-30 | 2021-01-08 | 中邮通建设咨询有限公司 | Intelligent danger source identification system and method for outdoor construction site |
CN112329584A (en) * | 2020-10-29 | 2021-02-05 | 深圳技术大学 | Method, system and equipment for automatically identifying foreign matters in power grid based on machine vision |
CN112465873B (en) * | 2020-12-10 | 2021-06-15 | 深圳市普汇智联科技有限公司 | Method and system for determining position of moving object in machine vision |
CN112465873A (en) * | 2020-12-10 | 2021-03-09 | 深圳市普汇智联科技有限公司 | Method and system for determining position of moving object in machine vision |
CN113392713A (en) * | 2021-05-20 | 2021-09-14 | 上海可深信息科技有限公司 | Work clothes wearing identification method |
CN113762171A (en) * | 2021-09-09 | 2021-12-07 | 赛思沃德(武汉)科技有限公司 | Method and device for monitoring safety of railway construction site |
CN114296029A (en) * | 2021-12-24 | 2022-04-08 | 宁夏广天夏电子科技有限公司 | Method and system for positioning personnel on mining face |
CN114296029B (en) * | 2021-12-24 | 2022-12-02 | 宁夏广天夏科技股份有限公司 | Method and system for positioning personnel on mining face |
CN114979611A (en) * | 2022-05-19 | 2022-08-30 | 国网智能科技股份有限公司 | Binocular sensing system and method |
CN114722445A (en) * | 2022-06-08 | 2022-07-08 | 山东高速德建集团有限公司 | Rivers and Dynamo-based dangerous engineering identification method |
CN114722445B (en) * | 2022-06-08 | 2022-09-02 | 山东高速德建集团有限公司 | Risk engineering identification method based on Revit and Dynamo |
CN114783144A (en) * | 2022-06-17 | 2022-07-22 | 深圳市易智博网络科技有限公司 | Intelligent building site safety monitoring method and device and computer equipment |
CN116152735A (en) * | 2023-01-09 | 2023-05-23 | 中国科学院空天信息创新研究院 | Security monitoring dangerous source identification system and method based on edge cloud computing |
CN117217547A (en) * | 2023-11-08 | 2023-12-12 | 安徽宜介科技股份有限公司 | Intelligent identification visual monitoring device and method for major hazard sources of construction engineering |
CN117217547B (en) * | 2023-11-08 | 2024-01-26 | 安徽宜介科技股份有限公司 | Intelligent identification visual monitoring device and method for major hazard sources of construction engineering |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107145851A (en) | Constructions work area dangerous matter sources intelligent identifying system | |
CN111967393B (en) | Safety helmet wearing detection method based on improved YOLOv4 | |
CN104951775B (en) | Railway highway level crossing signal region security intelligent identification Method based on video technique | |
CN109308447A (en) | The method of equipment operating parameter and operating status is automatically extracted in remote monitoriong of electric power | |
CN105373135B (en) | A kind of method and system of aircraft docking guidance and plane type recognition based on machine vision | |
CN103069434B (en) | For the method and system of multi-mode video case index | |
CN101389004B (en) | Moving target classification method based on on-line study | |
CN104637058B (en) | A kind of volume of the flow of passengers identify statistical methods based on image information | |
CN112396658B (en) | Indoor personnel positioning method and system based on video | |
CN108491758A (en) | A kind of track detection method and robot | |
CN104821056A (en) | Intelligent guarding method based on radar and video integration | |
CN106951889A (en) | Underground high risk zone moving target monitoring and management system | |
CN101883261A (en) | Method and system for abnormal target detection and relay tracking under large-range monitoring scene | |
CN108764167A (en) | A kind of target of space time correlation recognition methods and system again | |
CN112183472A (en) | Method for detecting whether test field personnel wear work clothes or not based on improved RetinaNet | |
CN109948455A (en) | One kind leaving object detecting method and device | |
CN109241847A (en) | The Oilfield Operation District safety monitoring system of view-based access control model image | |
CN112184773A (en) | Helmet wearing detection method and system based on deep learning | |
Liang et al. | Methods of moving target detection and behavior recognition in intelligent vision monitoring. | |
CN111681382A (en) | Method for detecting temporary fence crossing in construction site based on visual analysis | |
CN113436229A (en) | Multi-target cross-camera pedestrian trajectory path generation method | |
CN103049748B (en) | Behavior monitoring method and device | |
Tao et al. | Smoky vehicle detection based on multi-feature fusion and ensemble neural networks | |
CN116846059A (en) | Edge detection system for power grid inspection and monitoring | |
CN105095891A (en) | Human face capturing method, device and system |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170908 |