CN114782988B - Multistage safety early warning method oriented to construction environment - Google Patents
Multistage safety early warning method oriented to construction environment Download PDFInfo
- Publication number
- CN114782988B CN114782988B CN202210322348.7A CN202210322348A CN114782988B CN 114782988 B CN114782988 B CN 114782988B CN 202210322348 A CN202210322348 A CN 202210322348A CN 114782988 B CN114782988 B CN 114782988B
- Authority
- CN
- China
- Prior art keywords
- data
- wearing
- early warning
- behavior
- data set
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000010276 construction Methods 0.000 title claims abstract description 67
- 238000000034 method Methods 0.000 title claims abstract description 61
- 238000001514 detection method Methods 0.000 claims abstract description 35
- 230000008569 process Effects 0.000 claims abstract description 13
- 230000006399 behavior Effects 0.000 claims description 74
- 206010000117 Abnormal behaviour Diseases 0.000 claims description 46
- 238000003708 edge detection Methods 0.000 claims description 21
- YBJHBAHKTGYVGT-ZKWXMUAHSA-N (+)-Biotin Chemical compound N1C(=O)N[C@@H]2[C@H](CCCCC(=O)O)SC[C@@H]21 YBJHBAHKTGYVGT-ZKWXMUAHSA-N 0.000 claims description 18
- FEPMHVLSLDOMQC-UHFFFAOYSA-N virginiamycin-S1 Natural products CC1OC(=O)C(C=2C=CC=CC=2)NC(=O)C2CC(=O)CCN2C(=O)C(CC=2C=CC=CC=2)N(C)C(=O)C2CCCN2C(=O)C(CC)NC(=O)C1NC(=O)C1=NC=CC=C1O FEPMHVLSLDOMQC-UHFFFAOYSA-N 0.000 claims description 18
- 238000009826 distribution Methods 0.000 claims description 16
- 230000002159 abnormal effect Effects 0.000 claims description 11
- 238000004364 calculation method Methods 0.000 claims description 11
- 230000033001 locomotion Effects 0.000 claims description 10
- 238000013528 artificial neural network Methods 0.000 claims description 9
- 230000005856 abnormality Effects 0.000 claims description 8
- 230000007613 environmental effect Effects 0.000 claims description 8
- 238000002372 labelling Methods 0.000 claims description 8
- 230000004913 activation Effects 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 6
- 230000001788 irregular Effects 0.000 claims description 6
- 238000004140 cleaning Methods 0.000 claims description 4
- 231100001261 hazardous Toxicity 0.000 claims description 4
- 238000012549 training Methods 0.000 claims description 4
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 230000009194 climbing Effects 0.000 claims description 3
- 238000013135 deep learning Methods 0.000 claims description 3
- 238000010801 machine learning Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 230000007246 mechanism Effects 0.000 abstract description 8
- 238000004458 analytical method Methods 0.000 abstract description 5
- 230000000007 visual effect Effects 0.000 abstract description 3
- 230000000694 effects Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 238000013461 design Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000007547 defect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000008447 perception Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 230000000391 smoking effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/088—Non-supervised learning, e.g. competitive learning
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Abstract
The invention discloses a multistage safety early warning method oriented to a construction environment, and belongs to the field of safety early warning. The multistage safety early warning method facing the construction environment adopts a hierarchical early warning mechanism, so that the systematicness, the accuracy and the reliability of the multistage safety early warning of the construction big data are improved; the collected data are labeled and dangerously divided, and the association relation among the multidimensional multi-element data is deeply excavated, so that the early warning cost is effectively saved on the premise of ensuring the accuracy and timeliness; in the detection process, a visual depth analysis scheme suitable for project construction scenes is provided by combining with the existing advanced algorithm, the occurrence of dead angles in safety detection is effectively prevented, and the system is endowed with the capability of dynamically adjusting the early warning level.
Description
Technical Field
The invention belongs to the field of safety early warning, and particularly relates to a multistage safety early warning method oriented to a construction environment.
Background
Along with the continuous advancement of industrial informatization technology, how to perform early warning and prevention and control in advance for safety in project construction process in multiple directions, multiple angles and multiple levels has become one of the problems to be solved in safety production. The data of the early warning mechanism mainly originate from large data of a construction site, focus on the safety control of the construction site and surround key elements such as personnel, machinery, materials and environment, however, the existing related early warning model has the problems of single data, insufficient data linkage, small data processing scale, poor technical expansibility, unclear hierarchical control limit, high cost and the like, so that the system has insufficient mobility and expansibility, low accuracy and reliability, and poor timeliness and linkage.
Different from a general early warning system, the reasons for the inconspicuous effect and poor usability of the project construction early warning mechanism are mainly divided into the following three aspects: firstly, construction scenes are complex and changeable, the sources are single only by means of data collected by a sensor, and loopholes exist in detection due to self-limitation, so that the effective information acquisition rate is low, the data redundancy is high, and the dangerous level of early warning information cannot be accurately divided, extracted and judged; secondly, the relations between people and environment, objects (materials, equipment and the like) and environment and people and objects (materials, equipment and the like) are changed in many ways, and the construction scene and the state are different, so that the conventional early warning system is difficult to flexibly adjust aiming at the complex environment, and the difficulty of the early warning system is increased due to multi-element and multi-dimensional access; finally, under the construction scene, abnormal behaviors such as personnel alarm, smoking, chat and the like are difficult to detect in a conventional mode, and the conventional early warning system lacks global monitoring and large-scale image capturing, processing and analyzing capabilities, so that the safety early warning effect of construction big data is seriously influenced.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a multistage safety early warning method facing construction environment.
In order to achieve the purpose, the invention is realized by adopting the following technical scheme:
a multistage safety early warning method facing construction environment comprises the following steps;
s1, acquiring multi-dimensional construction big data C, wherein the multi-dimensional construction big data C comprises an image video and text data C of environmental temperature and humidity text The image video is composed of historical data C to be marked old And new data C collected in real time new Constructing;
historical data C to be annotated old Labeling different dimensions to obtain a label data set C L The method comprises the steps of carrying out a first treatment on the surface of the Label dataset C L Inner bagContaining hazardous area dataset C Da And dangerous behavior data set C Db ；
S2, based on label data set C L Predicting the wearing condition of the safety equipment by utilizing a neural network to obtain a wearing prediction result;
based on the hazardous area dataset C Da Dividing the dangerous area D area Calculating to obtain the edge detection confidence coefficient C border The method comprises the steps of carrying out a first treatment on the surface of the Predicting the personnel position information by using a neural network to obtain the personnel position detection confidence coefficient C location ；
New data C based on real-time acquisition new Generating a historical moving image MHI based on a dangerous behavior data set C Db Acquiring space-time information of normal human body movement, constructing and updating an abnormal behavior detection model by using a neural network based on the space-time information and a historical moving image MHI, and predicting personnel behaviors;
text data C according to ambient temperature and humidity text Calculating the mean value and standard deviation of the data;
s3, converting the wearing prediction result into wearing and non-wearing situations according to a preset threshold value, combining the wearing situations of all the safety equipment, and dividing the wearing situations into 8 levels of dangerous behaviors based on the combined result;
according to the edge detection confidence C border Confidence of person position detection C location Dividing the combination result of 4 objects and the construction environment into grades;
according to the personnel behavior prediction result and the dangerous behavior data set C occupied by each abnormal behavior Db Determining the abnormal behavior category;
s4, calculating a basic security score S of the personnel P Environmental security score S E And a global security score S G Dynamically adjusting the activation parameters of each security score to realize the dynamic adjustment of the early warning level and obtain the early warning scoreAnd simultaneously, early warning is implemented.
Further, the step S1 specifically includes:
cleaning the collected data by adopting a machine learning algorithm, deleting redundant data and treating historical data C to be marked old Labeling to obtain a safety helmet label L hat Safety belt label L belt Mask label L mask With construction vehicle label L car Conversion of data set C into tag data set C L ；
Marking data acquired by terminal equipment in a dangerous area in the data set C and data with dangerous behaviors respectively to obtain the dangerous area data set C Da And dangerous behavior data set C Db 。
Further, the step S2 specifically includes:
s201, utilizing YOLO deep learning network to perform label data set C L Training and modeling to obtain a safety wearing detection model, and simultaneously predicting the wearing conditions of the safety helmet, the wearing of the safety belt and the wearing conditions of the mask in the same scene to obtain a safety helmet wearing prediction score P helmet Belt wearing prediction score P belt And mask wearing prediction score P mask ；
S202, integrating the dangerous area data set C Da The data in the dangerous area D is subjected to graying, filtering and edge detection, then Hough linear transformation is adopted for the linear boundary, a sliding window is adopted for the curve boundary, and an irregular area is replaced by an irregular contour circumscribed minimum rotation rectangle, so that the dangerous area D is determined area Computing edge detection confidence C border The method comprises the steps of carrying out a first treatment on the surface of the Person position calculating person position detection confidence C obtained through adjacent frame relation and safe wearing detection model location ；
S203, deducting the dangerous behavior data set C from the data set C Db Obtaining a normal behavior data setNormal behavior data set +.>I-th original gesture Pi in (b)Normalizing to obtain spatial information; by ∈>Calculating to obtain the motion velocity v of the ith skeleton joint point i Using Rayleigh distribution to distribute v discretely i Modeling into probability distribution, namely normal behavior probability distribution of the human body, and taking the normal behavior probability distribution of the human body as time information; new data C acquired from real time new Generating a historical moving picture MHI;
utilizing time information and space information to construct regularity of human body posture track learned by a space-time-based transducer model, introducing a historical moving image MHI and a non-negative sparse automatic encoder, enabling the transducer model to learn in a self-supervision incremental mode, and predicting by taking the trained transducer model as an abnormal behavior detection model to obtain an abnormal behavior score F;
s204, text data C based on ambient temperature and humidity text And (3) determining the data of the uncontrolled state in the batch by utilizing the SPC statistical process judgment standard, and calculating the data abnormality index E.
Further, in S202, the edge detection confidence C border And object position confidence C location The calculation process of (2) is as follows:
wherein k is 1 Is a filtering influence factor; kernel_size is the size using gaussian kernels; k (k) 2 An influence factor is detected for the edge; t is the threshold of edge detection; h is a SUB Minimum contour height for the object obtained by the difference between adjacent frames; cur (cur) h The predicted height obtained for the safe wearing detection model; beta is a high scale factor.
Further, the person movement velocity v in S203 i The method comprises the following steps:
wherein, (x) i，j ，y i，j ) And (x) i+1，j ，y i+1，j ) Representing the positions of the jth node at the ith and the (i+1) th frames, respectively. 6. The construction environment-oriented multistage safety precaution method according to claim 3, characterized in that the data abnormality index E in S204 is calculated as follows:
N normal representing the number of normal data, N abnormal The number of data representing the uncontrolled state.
Further, the step S3 specifically includes:
s301, combining wearing prediction results according to importance degrees of different wearing devices in construction, dividing the wearing prediction results into 8 different levels of dangerous behaviors, and specifically comprising the following steps:
wherein w is h Wearing safety helmet for representative，Representing an unworn helmet; w (w) h Representing wearing safety belt, the->Representing an unworn seat belt; w (w) m Representing wearing mask, the wearer is about to wear the mask>Represents a mask which is not worn, w h 、w b 、w m ∈{0，1}；
S302, the position coordinates of the personnel are combined with D area The boundary coordinates are judged to obtain initial out-of-range behavior; according to the edge detection confidence C border Dividing the boundary decision factor S by the size of (a) b Confidence C is detected according to personnel position location Dividing the position determination factor S by the size of (a) l Based on the boundary determination factor S b And a position determination factor S l The combination of (a) classifies initial out-of-range behavior into class 4 out-of-range confidence levels B i The method is characterized by comprising the following steps:
boundary determination factor S b Position determination factor S l And out-of-range behavior B i The calculation is as follows:
s303, calculating three abnormal behaviors of alarm, fall and climb of personnel in dangerous behavior data set C Db Determining an abnormal behavior class A based on the abnormal behavior probability and the abnormal behavior score F j ，A j The method comprises the following steps:
wherein c fight Data set C representing dangerous behavior of alarm abnormal behavior Db The number of (3)；c fall Data set C representing abnormal falling behaviors in dangerous behaviors Db Is a number of (a) in (b); c sneak Data set C representing climbing abnormal behaviors in dangerous behaviors Db Is a number of (a) in (b); f represents an abnormal behavior score;
s304, setting an alarm threshold T alarm Data abnormality index and alarm threshold T alarm Comparing, and early warning the abnormal state, wherein the abnormal state is defined as follows:
wherein S is k Representing the kth state.
Further, a person basic security score S in step S4 P Environmental security score S E Global security score S G And an early warning scoreThe calculation method of (1) is as follows:
S P ＝D 1 *P helmet +D 2 *P belt +D 3 *P mask
S E ＝C border *C location
S G ＝F
wherein D is 1 、D 2 、D 3 Is a risk factor; f is an abnormal behavior score; a, a 1 、a 2 、a 3 To activate the parameters a 1 ，a 2 ，a 3 ∈{0，1}；w 1 、w 2 、w 3 Is a detection factor; s is S i Is the firsti activation parameter conditions under the early warning level.
Further, risk factor D 1 、D 2 、D 3 The method comprises the following steps:
wherein N is i Represents L i Behavior level in tag dataset C L Is a statistical quantity of (a) in the database.
Compared with the prior art, the invention has the following beneficial effects:
the invention provides a multistage safety early warning method facing to a construction environment, which adopts a hierarchical early warning mechanism around the difference between construction projects and general projects, and improves the systematicness, accuracy and reliability of multistage safety early warning of construction big data; the collected data are labeled and dangerously divided, and the association relation among the multidimensional multi-element data is deeply excavated, so that the early warning cost is effectively saved on the premise of ensuring the accuracy and timeliness; in the detection process, a visual depth analysis scheme suitable for project construction scenes is provided by combining with the existing advanced algorithm, the occurrence of dead angles in safety detection is effectively prevented, and the system is endowed with the capability of dynamically adjusting the early warning level.
Furthermore, in the step S1, processing operations such as labeling and danger division are performed through collected real-time data of a construction site, so that the original data can provide richer semantic information in multi-stage early warning, subsequent operation and processing are facilitated, and data perception and collection are completed.
Further, step S2 focuses on the semantics of the original data, and different tag data sets are processed and analyzed in a mode of combining an advanced algorithm with a traditional computer vision technology, so that extraction of key features of the data and discrimination of basic information are realized, and the early warning cost of constructing big data is greatly reduced.
Further, step S3 focuses on normative worn by constructors, association relation between objects and environment and personnel behavior distribution, meanwhile gives consideration to real-time states of materials in construction sites, builds early warning behaviors, provides regular guarantee for multi-stage early warning, and further improves early warning capability while realizing systematicness.
Further, in step S4, the early warning grade is dynamically adjusted according to the calculated personnel safety, the environment safety and the global safety score to effectively adapt to various occasions, prevent the early warning from occurring dead angles, combine the human skeleton information according to the normal distribution of the normal behaviors to effectively discriminate various abnormal behaviors, and realize the early warning of the abnormal behaviors by judging the rationality of the human behaviors in the whole to effectively prevent the early warning from occurring dead angles.
In summary, the early warning method fully realizes the multi-stage safety early warning mechanism suitable for different construction scenes, more systematically relates to the three-stage early warning module, improves the utilization rate of large construction data by extracting and training different tagged data, improves the adaptability of the model by an incremental learning mode, reduces the early warning cost by visual realization, and effectively prevents the occurrence of dead angles of early warning through the multi-stage early warning mechanism.
Drawings
FIG. 1 is a flow chart of a multi-stage three-dimensional project construction safety pre-warning method applied to industrial big data;
FIG. 2 is a schematic diagram of the end-edge-cloud architecture of the present invention;
fig. 3 is an implementation effect diagram.
Detailed Description
In order that those skilled in the art will better understand the present invention, a technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.
It should be noted that the terms "first," "second," and the like in the description and the claims of the present invention and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the invention described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
The invention is described in further detail below with reference to the attached drawing figures:
the invention provides a construction environment-oriented multistage safety early warning method, which carries out association and hierarchical analysis on primary data, designs a multistage early warning method based on industrial big data, designs a multistage early warning module aiming at the primary data collected by edge intelligent equipment, and firstly predicts the normative behavior of personnel wearing to realize primary early warning; aiming at a complex construction site, the aim of secondary early warning is fulfilled by analyzing the association relation between the primary data and the construction environment; aiming at the global state, the abnormal behaviors of personnel are analyzed in a three-dimensional mode, and therefore three-level early warning for large construction data is achieved.
Referring to fig. 1, fig. 1 is a flowchart of a construction environment-oriented multi-stage safety early warning method, and the specific steps of the construction environment-oriented multi-stage safety early warning method are as follows:
s1, collecting multi-dimensional construction big dataC, the multi-dimensional construction big data C comprises an image video and text data C of ambient temperature and humidity text The image video is composed of historical data C to be marked old And new data C collected in real time new Constructing;
historical data C to be annotated old Labeling different dimensions to obtain a label data set C L The method comprises the steps of carrying out a first treatment on the surface of the Label dataset C L Containing a data set C of dangerous areas Da And dangerous behavior data set C Db ；
The step S1 specifically comprises the following steps:
s101, cleaning the collected data by adopting a machine learning algorithm, deleting redundant data and treating historical data C to be marked old Labeling to obtain a safety helmet label L hat Safety belt label L belt Mask label L mask With construction vehicle label L car Conversion of data set C into tag data set C L ；
Marking data acquired by terminal equipment in a dangerous area in the data set C and data with dangerous behaviors respectively to obtain the dangerous area data set C Da And dangerous behavior data set C Db 。
S2, based on label data set C L Predicting the wearing condition of the safety equipment by utilizing a neural network to obtain a wearing prediction result;
based on the hazardous area dataset C Da Dividing the dangerous area D area And obtaining the edge detection confidence coefficient C by using the related parameters border The method comprises the steps of carrying out a first treatment on the surface of the Predicting the personnel position information by using a neural network to obtain the personnel position detection confidence coefficient C location ；
New data C based on real-time acquisition new Generating a historical moving image MHI based on a dangerous behavior data set C Db Acquiring space-time information of normal human body movement, constructing and updating an abnormal behavior detection model by using a neural network based on the space-time information and a historical moving image MHI, and predicting personnel behaviors;
text data C according to ambient temperature and humidity text Calculate dataMean and standard deviation of (a);
the step S2 specifically comprises the following steps:
s201, utilizing YOLO deep learning network to perform label data set C L Training and modeling to obtain a safety wearing detection model, and simultaneously predicting the wearing conditions of the safety helmet, the wearing of the safety belt and the wearing conditions of the mask in the same scene to obtain a safety helmet wearing prediction score P helmet Belt wearing prediction score P belt And mask wearing prediction score P mask 。
S202, integrating the dangerous area data set C Da The data in the dangerous area D is subjected to graying, filtering and edge detection, then Hough linear transformation is adopted for the linear boundary, a sliding window is adopted for the curve boundary, and an irregular area is replaced by an irregular contour circumscribed minimum rotation rectangle, so that the dangerous area D is determined area And calculating the edge detection confidence coefficient C by using parameters with strong relevance border In addition, person position detection confidence C is calculated through adjacent frame relation and person position obtained by the safe wearing detection model location 。
Further, edge detection confidence C border Confidence of object position C location The calculation process is as follows:
wherein k is 1 Is a filtering influence factor; kernel_size is the size using gaussian kernels; k (k) 2 An influence factor is detected for the edge; t is the threshold of edge detection; h is a SUB Minimum contour height for the object obtained by the difference between adjacent frames; cur (cur) h The predicted height obtained for the safe wearing detection model; beta is a high scale factor.
S203, deducting the dangerous behavior data set C from the data set C Db ObtainingNormal behavior data setNormal behavior data set +.>Normalizing the ith original gesture Pi as spatial information; by ∈>Calculating to obtain the motion velocity v of the ith skeleton joint point i Using Rayleigh distribution to distribute v discretely i Modeling into probability distribution, namely normal behavior probability distribution of the human body, and taking the normal behavior probability distribution of the human body as time information; new data C acquired from real time new Generating a historical moving picture MHI;
and constructing regularity of human body posture tracks learned by a space-time-based transducer model by utilizing time information and space information, introducing a historical moving image MHI and a non-negative sparse automatic encoder, enabling the transducer model to learn in a self-supervision incremental mode, and predicting by taking the trained transducer model as an abnormal behavior detection model to obtain an abnormal behavior score F.
Further, the movement speed v of the person i The calculation process of (2) is as follows:
wherein, (x) i，j ，y i，j ) And (x) i+1，j ，y i+1，j ) Respectively representing the positions of the jth joint point in the ith frame and the (i+1) th frame;
s204, text data C based on ambient temperature and humidity text And (3) determining the data of the uncontrolled state in the batch by utilizing the SPC statistical process judgment standard, and calculating the data abnormality index E.
Further, the data anomaly index E is calculated as follows:
wherein N is normal Representing the number of normal data, N abnormal The number of data representing the uncontrolled state.
S3, converting the wearing prediction result into wearing and non-wearing situations according to a preset threshold value, combining the wearing situations of all the safety equipment, and dividing the wearing situations into 8 levels of dangerous behaviors based on the combined result;
according to the edge detection confidence C border Confidence of person position detection C location Dividing the combination result of 4 objects and the construction environment into grades;
according to the personnel behavior prediction result and the dangerous behavior data set C occupied by each abnormal behavior Db Determining the abnormal behavior category;
the step S3 specifically comprises the following steps:
s301, combining wearing prediction results according to importance degrees of different wearing devices in construction, dividing the wearing prediction results into 8 different levels of dangerous behaviors, and specifically comprising the following steps:
wherein w is h Is represented by wearing a safety helmet,representing an unworn helmet; w (w) h Representing wearing safety belt, the->Representing an unworn seat belt; w (w) m Representing wearing mask, the wearer is about to wear the mask>Represents a mask which is not worn, w h 、w b 、w m ∈{0，1}；
S302, the position coordinates of the personnel are combined with D area The boundary coordinates are judged to obtain initial out-of-range behavior; according to the edge detection confidence C border Dividing the boundary decision factor S by the size of (a) b Confidence C is detected according to personnel position location Dividing the position determination factor S by the size of (a) l Based on the boundary determination factor S b And a position determination factor S l The combination of (a) classifies initial out-of-range behavior into class 4 out-of-range confidence levels B i The method is characterized by comprising the following steps:
boundary determination factor S b Position determination factor S l And out-of-range behavior B i The calculation is as follows:
s303, calculating three abnormal behaviors of alarm, fall and climb of personnel in dangerous behavior data set C Db Determining an abnormal behavior class A based on the abnormal behavior probability and the abnormal behavior score F j ，A j The method comprises the following steps:
wherein c fight Data set C representing dangerous behavior of alarm abnormal behavior Db Is a number of (a) in (b); c fall Representing fall-down differenceNormal behavior in dangerous behavior data set C Db Is a number of (a) in (b); c sneak Data set C representing climbing abnormal behaviors in dangerous behaviors Db Is a number of (a) in (b); f represents an abnormal behavior score;
s304, setting an alarm threshold T alarm Data abnormality index and alarm threshold T alarm Comparing, and early warning the abnormal state, wherein the abnormal state is defined as follows:
wherein S is k Representing the kth state.
S4, calculating a basic security score S of the personnel P Environmental security score S E And a global security score S G Dynamically adjusting the activation parameters of each security score to realize the dynamic adjustment of the early warning level and obtain the early warning scoreAnd simultaneously, early warning is implemented.
The step S4 specifically comprises the following steps:
s401 personnel basic Security score S P Environmental security score S E Global security score S G And an early warning scoreThe calculation method of (1) is as follows:
S P ＝D 1 *P helmet +D 2 *P belt +D 3 *P mask
S E ＝C border *C location
S G ＝F
wherein D is 1 、D 2 、D 3 Is a risk factor; f is an abnormal behavior score; a, a 1 、a 2 、a 3 To activate the parameters a 1 ，a 2 ，a 3 ∈{0，1}；w 1 、w 2 、w 3 Is a detection factor; s is S i The activation parameter condition under the ith early warning level.
Further, D 1 、D 2 、D 3 The calculation process is as follows:
where Ni represents the statistical quantity of the L_i behavior level in the tag dataset C_L.
In summary, by using the multi-stage safety early warning method facing the construction environment, which is provided by the invention, the capability of the newly acquired data lifting model can be continuously utilized, and meanwhile, the multi-scene and three-dimensional efficient dynamic early warning can be realized by means of the three-stage early warning method.
Referring to fig. 2, fig. 2 is an end-side-cloud architecture pattern diagram, which uses data sensing and acquisition-summarization and integration-analysis and reasoning one early warning as a system, uses an end-side-cloud mode facing industrial big data in a complex construction scene as an architecture, uses regional autonomy and core early warning analysis and regulation as a design thought, mainly forms a novel three-dimensional safety early warning mechanism by three-level early warning, and is specifically developed to monitor early warning original data in real time, such as personnel operation and wearing normalization, so as to realize one-level early warning; aiming at complex construction site environments, analyzing association relation between primary data related to safety and the construction environments, such as out-of-range detection of personnel and vehicles, combined detection of terminal temperature and humidity intelligent sensor data and the like, so as to realize secondary early warning; finally, starting from the global situation, three-dimensionally analyzing the abnormal behaviors of personnel, establishing an abnormal behavior rule base under a construction scene, and defining the abnormal behaviors by adopting an incremental learning method according to the normal behavior probability distribution of skeleton motion statistics, thereby realizing an industrial big data-oriented project construction safety three-level early warning mechanism. Meanwhile, in order to overcome the defects of high construction scene complexity, large range, wide distribution and multiple types, small-range autonomy formed by intelligent terminal equipment is researched and designed, a three-dimensional safety early warning 'end-side-cloud' architecture model regulated and controlled by a cloud end as a whole is developed based on the 'artificial' safety aim at moment, the method specifically comprises the steps of finishing the perception and collection of construction data by a terminal intelligent equipment (end), finishing the data preprocessing and preliminary early warning of data cleaning, labeling of main body labels and the like by an edge server (side), and establishing and dynamic early warning of a multi-stage early warning model rule base by the cloud end (cloud end).
Examples
The invention has the effects in construction site and railway construction scenes, as shown in fig. 3, the first column is an original picture acquired from intelligent terminal equipment, the second column is a datamation representation obtained through a detection model, and the third column is a qualitative description obtained through a rule base.
The above is only for illustrating the technical idea of the present invention, and the protection scope of the present invention is not limited by this, and any modification made on the basis of the technical scheme according to the technical idea of the present invention falls within the protection scope of the claims of the present invention.
Claims (9)
1. A multistage safety early warning method facing construction environment is characterized by comprising the following steps of;
s1, acquiring multi-dimensional construction big data C, wherein the multi-dimensional construction big data C comprises an image video and text data C of environmental temperature and humidity text The image video is composed of a history to be markedData C old And new data C collected in real time new Constructing;
historical data C to be annotated old Labeling different dimensions to obtain a label data set C L The method comprises the steps of carrying out a first treatment on the surface of the Label dataset C L Containing a data set C of dangerous areas Da And dangerous behavior data set C Db ；
S2, based on label data set C L Predicting the wearing condition of the safety equipment by utilizing a neural network to obtain a wearing prediction result;
based on the hazardous area dataset C Da Dividing the dangerous area D area Calculating to obtain the edge detection confidence coefficient C border The method comprises the steps of carrying out a first treatment on the surface of the Predicting the personnel position information by using a neural network to obtain the personnel position detection confidence coefficient C location ；
New data C based on real-time acquisition new Generating a historical moving image MHI based on a dangerous behavior data set C Db Acquiring space-time information of normal human body movement, constructing and updating an abnormal behavior detection model by using a neural network based on the space-time information and a historical moving image MHI, and predicting personnel behaviors;
text data C according to ambient temperature and humidity text Calculating the mean value and standard deviation of the data;
s3, converting the wearing prediction result into wearing and non-wearing situations according to a preset threshold value, combining the wearing situations of all the safety equipment, and dividing the wearing situations into 8 levels of dangerous behaviors based on the combined result;
according to the edge detection confidence C border Confidence of person position detection C location Dividing the combination result of 4 objects and the construction environment into grades;
according to the personnel behavior prediction result and the dangerous behavior data set C occupied by each abnormal behavior Db Determining the abnormal behavior category;
s4, calculating a basic security score S of the personnel P Environmental security score S E And a global security score S G Dynamically adjusting the activation parameters of each security score to achieve dynamic tuningThe early warning grade is integrated to obtain early warning scoreAnd simultaneously, early warning is implemented.
2. The construction environment-oriented multistage safety early warning method according to claim 1, wherein the step S1 is specifically:
cleaning the collected data by adopting a machine learning algorithm, deleting redundant data and treating historical data C to be marked old Labeling to obtain a safety helmet label L hat Safety belt label L belt Mask label L mask With construction vehicle label L car Conversion of data set C into tag data set C L ；
Marking data acquired by terminal equipment in a dangerous area in the data set C and data with dangerous behaviors respectively to obtain the dangerous area data set C Da And dangerous behavior data set C Db 。
3. The construction environment-oriented multistage safety early warning method according to claim 1, wherein the step S2 is specifically:
s201, utilizing YOLO deep learning network to perform label data set C L Training and modeling to obtain a safety wearing detection model, and simultaneously predicting the wearing conditions of the safety helmet, the wearing of the safety belt and the wearing conditions of the mask in the same scene to obtain a safety helmet wearing prediction score P helmet Belt wearing prediction score P belt And mask wearing prediction score P mask ；
S202, integrating the dangerous area data set C Da The data in the dangerous area D is subjected to graying, filtering and edge detection, then Hough linear transformation is adopted for the linear boundary, a sliding window is adopted for the curve boundary, and an irregular area is replaced by an irregular contour circumscribed minimum rotation rectangle, so that the dangerous area D is determined area Computing edge detection confidence C border The method comprises the steps of carrying out a first treatment on the surface of the Character obtained by adjacent frame relation and safe wearing detection modelConfidence C for position detection of position calculator location ；
S203, deducting the dangerous behavior data set C from the data set C Db Obtaining a normal behavior data setData set of normal behaviorNormalizing the ith original gesture Pi as spatial information; by ∈>Calculating to obtain the motion velocity v of the ith skeleton joint point i Using Rayleigh distribution to distribute v discretely i Modeling into probability distribution, namely normal behavior probability distribution of the human body, and taking the normal behavior probability distribution of the human body as time information; new data C acquired from real time new Generating a historical moving picture MHI;
utilizing time information and space information to construct regularity of human body posture track learned by a space-time-based transducer model, introducing a historical moving image MHI and a non-negative sparse automatic encoder, enabling the transducer model to learn in a self-supervision incremental mode, and predicting by taking the trained transducer model as an abnormal behavior detection model to obtain an abnormal behavior score F;
s204, text data C based on ambient temperature and humidity text And (3) determining the data of the uncontrolled state in the batch by utilizing the SPC statistical process judgment standard, and calculating the data abnormality index E.
4. The construction environment-oriented multistage safety precaution method according to claim 3, characterized in that the edge detection confidence degree C in S202 border And object position confidence C location The calculation process of (2) is as follows:
wherein k is 1 Is a filtering influence factor; kernel_size is the size using gaussian kernels; k (k) 2 An influence factor is detected for the edge; t is the threshold of edge detection; h is a SUB Minimum contour height for the object obtained by the difference between adjacent frames; cur (cur) h The predicted height obtained for the safe wearing detection model; beta is a high scale factor.
5. The construction environment-oriented multistage safety precaution method according to claim 3, characterized in that the personnel movement speed v in S203 i The method comprises the following steps:
wherein, (x) i，j ，y i，j ) And (x) i+1，j ，y i+1，j ) Representing the positions of the jth node at the ith and the (i+1) th frames, respectively.
6. The construction environment-oriented multistage safety precaution method according to claim 3, characterized in that the data abnormality index E in S204 is calculated as follows:
N normal representing the number of normal data, N abnormal The number of data representing the uncontrolled state.
7. The construction environment-oriented multistage safety precaution method according to claim 3, wherein the step S3 is specifically:
s301, combining wearing prediction results according to importance degrees of different wearing devices in construction, dividing the wearing prediction results into 8 different levels of dangerous behaviors, and specifically comprising the following steps:
wherein w is h Is represented by wearing a safety helmet,representing an unworn helmet; w (w) h Representing wearing safety belt, the->Representing an unworn seat belt; w (w) m Representing wearing mask, the wearer is about to wear the mask>Represents a mask which is not worn, w h 、w b 、w m ∈{0，1}；
S302, the position coordinates of the personnel are combined with D area The boundary coordinates are judged to obtain initial out-of-range behavior; according to the edge detection confidence C border Dividing the boundary decision factor S by the size of (a) b Checking according to personnel positionMeasurement confidence C location Dividing the position determination factor S by the size of (a) l Based on the boundary determination factor S b And a position determination factor S l The combination of (a) classifies initial out-of-range behavior into class 4 out-of-range confidence levels B i The method is characterized by comprising the following steps:
boundary determination factor S b Position determination factor S l And out-of-range behavior B i The calculation is as follows:
s303, calculating three abnormal behaviors of alarm, fall and climb of personnel in dangerous behavior data set C Db Determining an abnormal behavior class A based on the abnormal behavior probability and the abnormal behavior score F j ，A j The method comprises the following steps:
wherein c fight Data set C representing dangerous behavior of alarm abnormal behavior Db Is a number of (a) in (b); c fall Data set C representing abnormal falling behaviors in dangerous behaviors Db Is a number of (a) in (b); c sneak Data set C representing climbing abnormal behaviors in dangerous behaviors Db Is a number of (a) in (b); f represents an abnormal behavior score;
s304, setting an alarm threshold T alarm Data abnormality index and alarm threshold T alarm Comparing, and early warning the abnormal state, wherein the abnormal state is defined as follows:
wherein S is k Representing the kth state.
8. The construction environment-oriented multi-component system of claim 1The level safety early warning method is characterized in that in step S4, the basic safety score S of personnel is as follows P Environmental security score S E Global security score S G And an early warning scoreThe calculation method of (1) is as follows:
S P ＝D 1 *P helmet +D 2 *P belt +D 3 *P mask
S E ＝C border *C location
S G ＝F
wherein D is 1 、D 2 、D 3 Is a risk factor; f is an abnormal behavior score; a, a 1 、a 2 、a 3 To activate the parameters a 1 ，a 2 ，a 3 ∈{0，1}；w 1 、w 2 、w 3 Is a detection factor; s is S i The activation parameter condition under the ith early warning level.
9. The construction environment-oriented multistage safety precaution method according to claim 8, characterized in that the risk factor D 1 、D 2 、D 3 The method comprises the following steps:
wherein N is i Represents L i Behavior level in tag dataset C L Is a statistical quantity of (a) in the database.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210322348.7A CN114782988B (en) | 2022-03-29 | 2022-03-29 | Multistage safety early warning method oriented to construction environment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210322348.7A CN114782988B (en) | 2022-03-29 | 2022-03-29 | Multistage safety early warning method oriented to construction environment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114782988A CN114782988A (en) | 2022-07-22 |
| CN114782988B true CN114782988B (en) | 2024-04-02 |
Family
ID=82427554
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210322348.7A Active CN114782988B (en) | 2022-03-29 | 2022-03-29 | Multistage safety early warning method oriented to construction environment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114782988B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116503797B (en) * | 2023-04-10 | 2023-10-27 | 盐城工学院 | Medical waste treatment personnel protection tool wearing detection method based on target detection |
| CN117041502B (en) * | 2023-10-10 | 2023-12-08 | 湖南睿图智能科技有限公司 | Dangerous scene analysis and monitoring system and method based on machine vision |
| CN117168633B (en) * | 2023-10-20 | 2024-02-02 | 南通豪强电器设备有限公司 | High-low voltage complete equipment protection method and system based on temperature monitoring |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110121053A (en) * | 2018-02-07 | 2019-08-13 | 中国石油化工股份有限公司 | A kind of video monitoring method of situ of drilling well risk stratification early warning |
| CN110119656A (en) * | 2018-02-07 | 2019-08-13 | 中国石油化工股份有限公司 | Intelligent monitor system and the scene monitoring method violating the regulations of operation field personnel violating the regulations |
| CN112785798A (en) * | 2020-12-04 | 2021-05-11 | 国网江苏省电力工程咨询有限公司 | Behavior analysis method for construction project constructors of electric power substation engineering |
| CN113052125A (en) * | 2021-04-09 | 2021-06-29 | 内蒙古科电数据服务有限公司 | Construction site violation image recognition and alarm method |
| KR20220007460A (en) * | 2020-07-10 | 2022-01-18 | 주식회사 포스코 | Method for determining whether to wear personal protective equipment and server for performing the same |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11301683B2 (en) * | 2018-10-10 | 2022-04-12 | Autodesk, Inc. | Architecture, engineering and construction (AEC) construction safety risk analysis system and method for interactive visualization and capture |
-
2022
- 2022-03-29 CN CN202210322348.7A patent/CN114782988B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110121053A (en) * | 2018-02-07 | 2019-08-13 | 中国石油化工股份有限公司 | A kind of video monitoring method of situ of drilling well risk stratification early warning |
| CN110119656A (en) * | 2018-02-07 | 2019-08-13 | 中国石油化工股份有限公司 | Intelligent monitor system and the scene monitoring method violating the regulations of operation field personnel violating the regulations |
| KR20220007460A (en) * | 2020-07-10 | 2022-01-18 | 주식회사 포스코 | Method for determining whether to wear personal protective equipment and server for performing the same |
| CN112785798A (en) * | 2020-12-04 | 2021-05-11 | 国网江苏省电力工程咨询有限公司 | Behavior analysis method for construction project constructors of electric power substation engineering |
| CN113052125A (en) * | 2021-04-09 | 2021-06-29 | 内蒙古科电数据服务有限公司 | Construction site violation image recognition and alarm method |
Non-Patent Citations (2)
| Title |
|---|
| 基于BIM与机器视觉技术结合的建筑施工危险区域入侵预警研究;王伟;吕山可;张雨果;赵楚楠;何华刚;;安全与环境工程;20200330(02);全文 * |
| 山区高速公路安全生产视频监控系统关键技术研究;郭亨聪;黄玲;何斌;;湖南交通科技;20180925(03);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114782988A (en) | 2022-07-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114782988B (en) | Multistage safety early warning method oriented to construction environment | |
| CN111898514B (en) | Multi-target visual supervision method based on target detection and action recognition | |
| EP3633615A1 (en) | Deep learning network and average drift-based automatic vessel tracking method and system | |
| CN108564069B (en) | Video detection method for industrial safety helmet | |
| CN106128022B (en) | A kind of wisdom gold eyeball identification violent action alarm method | |
| CN110414400B (en) | Automatic detection method and system for wearing of safety helmet on construction site | |
| Junaedi et al. | Driver drowsiness detection based on face feature and PERCLOS | |
| CN104616438A (en) | Yawning action detection method for detecting fatigue driving | |
| CN107320115B (en) | Self-adaptive mental fatigue assessment device and method | |
| CN110728252A (en) | Face detection method applied to regional personnel motion trail monitoring | |
| CN106127814A (en) | A kind of wisdom gold eyeball identification gathering of people is fought alarm method and device | |
| CN111931573A (en) | Helmet detection and early warning method based on YOLO evolution deep learning model | |
| CN114387662A (en) | Method for identifying unsafe behaviors of underground personnel in coal mine | |
| CN112183532A (en) | Safety helmet identification method based on weak supervision collaborative learning algorithm and storage medium | |
| CN117423157A (en) | Mine abnormal video action understanding method combining migration learning and regional invasion | |
| Lee et al. | Modeling of human walking trajectories for surveillance | |
| AlKishri et al. | Enhanced image processing and fuzzy logic approach for optimizing driver drowsiness detection | |
| CN114495421B (en) | Intelligent open type road construction operation monitoring and early warning method and system | |
| CN113379247B (en) | Modeling method and system for enterprise potential safety hazard tracking model | |
| CN115169673A (en) | Intelligent campus epidemic risk monitoring and early warning system and method | |
| CN114359831A (en) | Risk omen reasoning-oriented intelligent identification system and method for worker side-falling | |
| Beiyi et al. | Detection and tracking of safety helmet in construction site | |
| CN117058627B (en) | Public place crowd safety distance monitoring method, medium and system | |
| CN117423210B (en) | Nursing is with disease anti-drop intelligent response alarm system | |
| Phayde et al. | Real-Time Drowsiness Diagnostic System Using Opencv Algorithm |
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 |