CN116612440B - Building engineering safety monitoring method, equipment and medium based on machine vision - Google Patents
Building engineering safety monitoring method, equipment and medium based on machine vision Download PDFInfo
- Publication number
- CN116612440B CN116612440B CN202310896113.3A CN202310896113A CN116612440B CN 116612440 B CN116612440 B CN 116612440B CN 202310896113 A CN202310896113 A CN 202310896113A CN 116612440 B CN116612440 B CN 116612440B
- Authority
- CN
- China
- Prior art keywords
- fire
- construction
- monitoring
- grid
- target
- 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
- 238000012544 monitoring process Methods 0.000 title claims abstract description 116
- 238000000034 method Methods 0.000 title claims abstract description 46
- 238000010276 construction Methods 0.000 claims abstract description 178
- 238000001514 detection method Methods 0.000 claims abstract description 69
- 238000011161 development Methods 0.000 claims abstract description 38
- 230000015654 memory Effects 0.000 claims abstract description 32
- 238000009435 building construction Methods 0.000 claims abstract description 12
- 230000007480 spreading Effects 0.000 claims description 75
- 238000003892 spreading Methods 0.000 claims description 75
- 239000000463 material Substances 0.000 claims description 67
- 230000018109 developmental process Effects 0.000 claims description 36
- 238000005457 optimization Methods 0.000 claims description 16
- 239000000779 smoke Substances 0.000 claims description 10
- 238000004891 communication Methods 0.000 claims description 9
- 230000000007 visual effect Effects 0.000 claims description 7
- 238000005516 engineering process Methods 0.000 claims description 6
- 238000003860 storage Methods 0.000 claims description 6
- 230000008859 change Effects 0.000 claims description 4
- 238000012216 screening Methods 0.000 claims description 4
- 238000013461 design Methods 0.000 claims description 3
- 230000007787 long-term memory Effects 0.000 claims description 3
- 238000012806 monitoring device Methods 0.000 claims description 3
- 210000002569 neuron Anatomy 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 claims description 3
- 230000006403 short-term memory Effects 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 230000008569 process Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 239000012634 fragment Substances 0.000 description 4
- 230000001960 triggered effect Effects 0.000 description 4
- 230000004044 response Effects 0.000 description 3
- 229910000831 Steel Inorganic materials 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000003825 pressing Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 239000010959 steel Substances 0.000 description 2
- 208000027418 Wounds and injury Diseases 0.000 description 1
- 230000005856 abnormality Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000002485 combustion reaction Methods 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 230000003111 delayed effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000010892 electric spark Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000012535 impurity Substances 0.000 description 1
- 208000014674 injury Diseases 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 208000024891 symptom Diseases 0.000 description 1
- 238000012549 training Methods 0.000 description 1
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
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D21/00—Measuring or testing not otherwise provided for
- G01D21/02—Measuring two or more variables by means not covered by a single other subclass
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/13—Architectural design, e.g. computer-aided architectural design [CAAD] related to design of buildings, bridges, landscapes, production plants or roads
-
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield networks
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/08—Construction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B17/00—Fire alarms; Alarms responsive to explosion
- G08B17/10—Actuation by presence of smoke or gases, e.g. automatic alarm devices for analysing flowing fluid materials by the use of optical means
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B17/00—Fire alarms; Alarms responsive to explosion
- G08B17/12—Actuation by presence of radiation or particles, e.g. of infrared radiation or of ions
- G08B17/125—Actuation by presence of radiation or particles, e.g. of infrared radiation or of ions by using a video camera to detect fire or smoke
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B31/00—Predictive alarm systems characterised by extrapolation or other computation using updated historic data
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Multimedia (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Tourism & Hospitality (AREA)
- Computing Systems (AREA)
- Geometry (AREA)
- Emergency Management (AREA)
- Human Resources & Organizations (AREA)
- Computer Hardware Design (AREA)
- Computational Linguistics (AREA)
- General Business, Economics & Management (AREA)
- Artificial Intelligence (AREA)
- Strategic Management (AREA)
- General Engineering & Computer Science (AREA)
- Primary Health Care (AREA)
- Marketing (AREA)
- Economics (AREA)
- Architecture (AREA)
- Mathematical Physics (AREA)
- Structural Engineering (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Databases & Information Systems (AREA)
- Civil Engineering (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Data Mining & Analysis (AREA)
- Biophysics (AREA)
- Development Economics (AREA)
Abstract
The embodiment of the specification discloses a building engineering safety monitoring method, equipment and medium based on machine vision, relates to the technical field of building construction safety, and is used for solving the problem of low reliability of fire safety detection of the existing building engineering. The method comprises the following steps: acquiring detection information acquired by multi-type sensor equipment and monitoring videos acquired by monitoring equipment; extracting a differential image of a key frame image in a monitoring video, inputting the image characteristics of the differential image into a preset classification recognition model, and outputting whether a fire disaster target and a fire disaster target position exist in a construction site of a building engineering; determining a fire monitoring area based on the fire target position, and meshing the fire area to obtain fire state parameters of each boundary grid position; and inputting the fire state parameters into a preset long-short-term memory network to obtain the development trend prediction information of the fire target so as to trigger an alarm instruction of the fire target based on the fire target, the position of the fire target and the development trend prediction information.
Description
Technical Field
The specification relates to the technical field of building construction safety, in particular to a building engineering safety monitoring method, equipment and medium based on machine vision.
Background
With the development of economy and the gradual increase of the pace of urban construction, the number of various buildings in the city is increased. In the construction process of the building engineering, engineering personnel are highly concentrated due to the requirement of the engineering, so that once a fire disaster occurs in the construction process of the building engineering, the fire disaster and the smoke gas spread faster, people are difficult to evacuate, and serious personnel injury is easy to cause. Therefore, the safety detection of the building engineering can timely early warn fire, is beneficial to improving the escape time, and has important significance for guaranteeing the personal and property safety.
In the current construction engineering, fire disaster detection is generally based on front-end camera equipment and rear-end monitoring equipment, video data acquired through front-end acquisition equipment is input to the rear-end monitoring equipment, and is mainly manually observed by monitoring personnel in front of monitors, but for large-scale construction scenes such as the construction of high-rise buildings, a large number of monitors are installed in a monitoring room for displaying monitoring images of various construction positions. At this time, due to the limitation of manual observation, the fire symptoms appearing in the monitoring images of a plurality of construction positions are difficult to timely detect. When the fire situation caused by negligence of personnel is not found in time at this moment, if inflammable materials exist in the current environment, the fire development speed is greatly improved, and the fire situation is difficult to control in time.
Disclosure of Invention
In order to solve the above technical problems, one or more embodiments of the present disclosure provide a method, an apparatus, and a medium for detecting safety of a building engineering based on machine vision.
One or more embodiments of the present disclosure adopt the following technical solutions:
one or more embodiments of the present disclosure provide a machine vision-based construction engineering safety detection method, including:
acquiring a current construction stage of a construction site of a building engineering so as to determine a current construction object and a current building bill of materials of the building engineering based on a construction scheme corresponding to the current construction stage; wherein, the bill of materials includes: material information and material use position;
acquiring detection information acquired by multi-type sensor equipment of a construction site of a building engineering and monitoring videos acquired by monitoring equipment;
extracting a key frame image in the monitoring video in a preset monitoring period, and acquiring adjacent images of the key frame image, so that image features of differential images of the key frame image and the adjacent images are input into a preset classification recognition model, and whether a fire disaster target exists in the construction site of the building engineering and the position of the fire disaster target are output;
Determining a fire monitoring area based on the fire target position, meshing the fire area to obtain detection information corresponding to each boundary grid position, and determining material information corresponding to the detection information based on the material using position to take the detection information and the corresponding material information as fire state parameters of each boundary grid position;
and inputting fire state parameters of each boundary grid position in the fire disaster area into a preset long-short-period memory network to obtain development trend prediction information of the fire target so as to trigger an alarm instruction of the fire target based on the fire target, the fire target position and the development trend prediction information.
Optionally, in one or more embodiments of the present disclosure, the method further includes inputting a fire status parameter of each boundary grid position in the fire area into a preset long-short-period memory network to obtain the development trend prediction information of the fire target, where the method specifically includes:
determining an initial spreading path grid corresponding to the boundary grid position according to the connection line of the boundary grid position and the fire target position, and determining a main air port type of each grid in the initial spreading path grid according to the air port type of each grid in the initial spreading path grid; wherein the tuyere type includes: the upper air port type is positioned at the upper air port of the fire disaster target, and the lower air port type is positioned at the lower air port of the fire disaster target;
Determining grid line intersection points in the connecting line to obtain two initial spreading path grids corresponding to the grid line intersection points as a first spreading path grid and a second spreading path grid;
based on the wind gap types of the first spreading path grid and the second spreading path grid, respectively determining first weight values of the first spreading path grid and the second spreading path grid; the first weight value corresponding to the main tuyere type is larger than the first weight value corresponding to the non-main tuyere type, and the sum of the first weight value and the first weight value is 1;
determining second weight values of the first spreading path grid and the second spreading path grid according to the number of fire-related grids of the preamble grids of the first spreading path grid and the second spreading path grid respectively; wherein the second weight value is proportional to the number of fire-related grids;
acquiring a total weight value based on the product of the first weight value and the second weight value, and determining a to-be-filtered spreading path grid in the first spreading path grid and the second spreading path grid by comparing the total weight value so as to delete the to-be-filtered spreading path grid and obtain a spreading path grid;
Screening fire state parameters of the boundary grid positions and the spreading path grids based on a Pearson correlation coefficient method to obtain data with higher fire development correlation degree as key data;
performing parameter optimization on an initial long-short-term memory network based on a preset optimization algorithm to obtain an optimal parameter combination, and updating the initial long-short-term memory network based on the optimal parameter combination to obtain a preset long-short-term memory network; wherein the parameters include: the number of neurons, the width of a sliding time window and an initial learning value;
and inputting the key data of each boundary grid position into a preset long-term and short-term memory network for iterative prediction so as to obtain development trend prediction information corresponding to the fire target.
Optionally, in one or more embodiments of the present disclosure, the optimizing the parameters of the initial long-short term memory network based on the preset optimizing algorithm to obtain an optimal parameter combination, so as to update the initial long-short term memory network based on the optimal parameter combination to obtain a preset long-short term memory network, which specifically includes:
determining the number of random combined populations of preset numbers of parameters in the initial long-short term network, determining the maximum iteration number of the preset optimization algorithm, and determining the current searching parameter nodes of the random combined populations;
Determining the adaptability of each node in each random combined population based on the distance between the searching parameter node and other nodes in the random combined population, and determining the current optimal position in the random combined population according to the adaptability;
and updating the positions of all parameter combinations in the random combination population according to the preset coefficient vector of the preset optimization algorithm and the maximum iteration times, and acquiring the optimal positions in the updated random combination population, wherein the parameter combinations based on the nodes corresponding to the optimal positions are used as optimal parameter combinations.
Optionally, in one or more embodiments of the present disclosure, the obtaining a current construction stage of a construction site of a building engineering to determine, based on a construction scheme corresponding to the current construction stage, a current construction object and a current building bill of materials of the building engineering specifically includes:
acquiring a construction stage schedule corresponding to the current construction project of the construction site, so as to determine an adjustment value of the construction stage schedule based on a difference value between the end time of the previous construction stage and the planned end time of the previous construction stage in the construction stage schedule;
Adjusting the construction stage schedule according to the adjustment value to obtain a current construction stage schedule of the construction site of the building engineering;
acquiring a current construction stage of the building construction site based on the current construction stage schedule and the current time;
and acquiring a construction scheme corresponding to the current construction stage according to the construction plan of the building engineering so as to determine a current construction object and a current building bill of materials of the building engineering based on the construction scheme.
Optionally, in one or more embodiments of the present disclosure, before the acquiring the detection information collected by the multi-type sensor device at the construction site of the building engineering, the method further includes:
acquiring a building design drawing of the current building engineering to establish a BIM model of the current building engineering based on a Luban civil engineering technology;
determining a typical scattered point position of a construction site of the current building engineering based on a BIM model of the current building engineering; wherein the typical scatter points include: the air vent, the floor slab top, the circuit crossing point and the electric switch;
determining a constructed range and a to-be-constructed range in the BIM based on the construction scheme of the current construction stage, so as to determine the matching relation between each typical scattered point position and sensor equipment according to detection attributes corresponding to each typical scattered point position in the constructed range and the to-be-constructed range;
Determining an initial layout position of the sensor equipment based on the matching relation, acquiring effective ranges of the sensor equipment of each type in the initial layout position, and determining the coverage range of the sensor equipment of each type based on the effective ranges of the sensor equipment of each type;
if the coverage area is determined to not fully cover the constructed range and the to-be-constructed range, acquiring the residual coverage areas of the constructed range and the to-be-constructed range; wherein the residual coverage range is the constructed range and the area range except the coverage range in the to-be-constructed range;
acquiring each communication range of the residual coverage range to acquire a central point of each communication range as a supplementary layout position of the sensor equipment of the same type;
and laying each sensor based on the initial laying position and the supplementary laying position so as to acquire detection information acquired by multi-type sensor equipment on the construction site of the building engineering.
Optionally, in one or more embodiments of the present disclosure, the extracting a key frame image in the monitoring video in a preset monitoring period, and acquiring an adjacent image of the key frame image, so as to input an image feature of a difference image between the key frame image and the adjacent image into a preset classification and identification model, and output whether a fire target exists in the construction site of the building engineering and the location of the fire target, specifically includes:
Segment segmentation is carried out on the monitoring video based on a preset analysis interval to obtain a plurality of monitoring analysis segments, so that the monitoring analysis segments are acquired according to a preset acquisition interval to obtain monitoring images of the monitoring analysis segments;
determining a gray threshold value of each monitoring image according to the gray histogram of each monitoring image so as to determine a binarized image of each monitoring image based on the gray threshold value of each monitoring image in the monitoring analysis segment;
sequentially comparing the binarized images, determining a sudden change image in the monitoring image as a key frame image of the monitoring video to obtain a plurality of adjacent images of the key frame image, sequentially obtaining differential images based on the time sequence of the key frame image and the adjacent images, and taking a communication area of the differential images as a hidden danger area;
converting pixel points of the hidden danger areas into a YCbCr color model to acquire color characteristics of the hidden danger areas, and acquiring local highest points of the hidden danger areas, so as to determine sharp angle characteristics of the hidden danger areas based on comparison between the local highest points and adjacent points of the local highest points, and determine area growth characteristics of the hidden danger areas according to area transformation of each hidden danger area;
Inputting the color characteristics, the sharp angle characteristics, the area increase characteristics and the key frame images into the preset classification and identification model to output whether a fire target exists in the construction site of the building engineering and the position of the fire target; the preset classification recognition model is a least square vector model.
Optionally, in one or more embodiments of the present disclosure, determining a fire monitoring area based on the fire target location, meshing the fire area to obtain detection information corresponding to each boundary mesh location, and determining material information corresponding to the detection information based on the material usage location, so as to take the detection information and the corresponding material information as fire status parameters of each boundary mesh location, including:
determining a fire monitoring area of the construction site of the building engineering based on the fire target position and a preset safety radius;
performing grid division on the fire disaster area based on a preset grid size, determining the corresponding relation between each grid boundary position and the sensor equipment based on the layout position of the sensor equipment and the effective acquisition range of the sensor equipment, and acquiring detection information corresponding to each boundary grid position according to the corresponding relation;
And determining a matching relation between the material information and the detection information according to the effective acquisition range of each sensor device and the use position of the material, and determining the material information corresponding to the detection information according to the matching relation, wherein the detection information and the corresponding material information are used as fire state parameters of each boundary grid position.
Optionally, in one or more embodiments of the present disclosure, the alarm instruction for triggering the fire target based on the fire target, the location of the fire target and the trend prediction information specifically includes:
if the fire type of the fire target is determined to be smoke based on the fire target, and the spreading probability of the fire target is determined to be lower than the preset spreading probability based on the development trend prediction information, triggering an audible and visual alarm instruction;
if the fire type of the fire target is determined to be open fire based on the fire target, or the spreading probability of the fire target is determined to be greater than the preset spreading probability based on the development trend prediction information, triggering an audible and visual alarm instruction and a fire alarm instruction;
and generating a fire alarm instruction based on the fire alarm instruction template and reporting the fire target position, the fire target and the fire target position so as to timely perform fire rescue.
One or more embodiments of the present specification provide a machine vision-based construction engineering safety monitoring apparatus, the apparatus comprising: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform any one of the methods described above.
One or more embodiments of the present specification provide a non-volatile computer storage medium storing computer-executable instructions capable of performing any of the methods described above.
The above-mentioned at least one technical scheme that this description embodiment adopted can reach following beneficial effect:
the fire safety monitoring is realized by combining the detection information and the monitoring video by acquiring the detection information acquired by the multi-type sensor equipment and the monitoring video acquired by the monitoring equipment on the construction site of the building engineering. The problem that detection is unreliable only based on monitoring equipment or only based on sensor equipment during fire detection is solved. The mode of acquiring the differential image to lock the fire disaster target after extracting the key frame image in the monitoring video avoids the problems that the monitoring picture of the multi-monitoring equipment is easy to miss detection and not found timely when being observed manually. The fire state parameters of the boundary grid positions are determined after the fire area is grid-divided, so that the prediction of the prediction information of the fire development trend is facilitated, and the timely alarm response to the fire spread is facilitated.
Drawings
In order to more clearly illustrate the embodiments of the present description or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described below, it being obvious that the drawings in the following description are only some of the embodiments described in the present description, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art. In the drawings:
fig. 1 is a schematic flow chart of a building engineering safety monitoring method based on machine vision according to an embodiment of the present disclosure;
FIG. 2 (a) is a schematic diagram of a fire area grid provided in an embodiment of the present disclosure;
fig. 2 (b) is a schematic diagram of an extension path mesh of a fire area according to an embodiment of the present disclosure;
fig. 3 is a schematic diagram of an internal structure of a machine vision-based building engineering safety monitoring device according to an embodiment of the present disclosure;
fig. 4 is a schematic diagram of an internal structure of a nonvolatile storage medium according to an embodiment of the present disclosure.
Detailed Description
The embodiment of the specification provides a building engineering safety monitoring method, equipment and medium based on machine vision.
In order to make the technical solutions in the present specification better understood by those skilled in the art, the technical solutions in the embodiments of the present specification will be clearly and completely described below with reference to the drawings in the embodiments of the present specification, and it is obvious that the described embodiments are only some embodiments of the present specification, not all embodiments. All other embodiments, which can be made by one of ordinary skill in the art based on the embodiments herein without making any inventive effort, shall fall within the scope of the present disclosure.
As shown in fig. 1, one or more embodiments of the present disclosure provide a machine vision-based construction engineering safety detection method. As can be seen from fig. 1, a machine vision-based construction engineering safety detection method comprises the following steps:
s101: acquiring a current construction stage of a construction site of a building engineering so as to determine a current construction object and a current building bill of materials of the building engineering based on a construction scheme corresponding to the current construction stage; wherein, the bill of materials includes: material information and the use position of the material.
Because different construction stages can build different building structures when the building construction is carried out on the building construction site, and because part of materials in the building process are inflammable materials, the building construction site needs to be monitored to ensure the safety of the building engineering in order to ensure the fire safety in the building construction process. Specifically, in one or more embodiments of the present disclosure, a current construction stage of a building construction site is first obtained, so that a construction object of a building engineering and a current building bill of materials are obtained according to a construction scheme corresponding to the current construction stage. It can be understood that when the construction object, that is, the object to be constructed at present, for example, the current construction stage is the construction of the beam side form, the construction object is the current beam side form, the foot pressing plate, the diagonal bracing, and the like. The building bill of materials includes material information and the use position of the materials, for example: when the beam side template is constructed, the construction scheme indicates that concrete pouring is needed, and then the concrete material information needed in the current construction stage and the position where the materials are used can be determined based on the building bill of materials in the current construction stage.
Specifically, in one or more embodiments of the present disclosure, a current construction stage of a construction site of a building engineering is obtained to determine a current construction object and a current building bill of materials of the building engineering based on a construction scheme corresponding to the current construction stage, which specifically includes:
firstly, a construction stage schedule corresponding to the current construction project in a construction site is obtained, and the construction stage of the construction project can be finished in advance of the time of the schedule or later than the time of the schedule because the construction of the construction project can be influenced by various factors such as weather, sudden accidents and the like. So that the construction phase schedule can be adjusted in time for determination, and the current construction phase is determined according to the current time. According to the method and the device, in the embodiment of the specification, according to the ending time of the last construction stage and the difference value of the planned ending time of the last construction stage in the construction stage schedule, an adjustment value of the construction stage schedule is determined; for example: the end time of the last construction stage is 6 months 10, and the planned end time of the last construction stage in the construction stage schedule is 6 months 8, and it can be determined that the difference is 2 days, that is, the end time of the last construction stage is delayed by 2 days, so that the schedule of the construction stages after the last construction stage in the construction stage schedule needs to be moved backward as a whole. The construction stage schedule is required to be adjusted according to the adjustment value so as to obtain the current construction stage schedule of the construction site of the building engineering.
For example: and when the schedule of the current construction stage before adjustment is 6 months 9, and the schedule of the current construction stage after adjustment is 2 days based on the adjustment value is 6 months 11, and the schedule of the current construction stage after adjustment is 6 months 27. At the moment, the current construction stage of the building construction site can be accurately obtained according to the current construction stage schedule and the current time, and the problem that fire safety detection is unreliable due to the fact that the construction stage is determined to be performed incorrectly due to uncontrollable construction engineering is avoided. That is, if the construction stage is not adjusted, an error occurs when the construction stage of 6 months 26 is judged based on the original construction stage schedule, and thus, the construction scheme and the acquisition of the material information are wrong, so that inflammable materials in different construction stages cannot be fully considered. After the current construction stage is determined, a construction scheme corresponding to the current construction stage can be acquired according to a construction plan of the building engineering, so that a current construction object and a current building bill of materials of the building engineering are determined based on the construction scheme. For example: when the current construction stage is floor template construction, the construction scheme is' guarantee: 18mm of a bottom building template, wherein a first layer of keels (secondary ribs) adopt single purlin b=80 mm, h=80 mm and the spacing is 350mm; the second layer of keels (main keels) adopts double steel pipes phi 48 multiplied by 3.0. At this time, the template is paved from the periphery and the middle is closed. And (5) during edge pressing, the angle template is nailed through a line. After the floor template is paved, whether the bracket is firm or not is carefully checked, the surface of the plate is smooth, the joint is standard, and the whole template surface is clean and free of impurities. After the template is installed, the template should be cleaned, and the next working procedure construction is carried out after the template is qualified through item inspection. "the current bill of materials based on the construction scheme includes: steel pipes, plates of floor templates, and the like.
S102: and acquiring detection information acquired by the multi-type sensor equipment of the construction site of the building engineering and monitoring videos acquired by the monitoring equipment.
In order to avoid the problem that detection is unreliable only based on monitoring equipment or only when fire is detected based on sensor equipment, detection information collected by multiple types of sensor equipment on a construction site of a building engineering and monitoring videos collected by the monitoring equipment are obtained in the embodiment of the specification, so that the detection information and the monitoring videos are combined to carry out reliable monitoring of fire safety.
In order to enable the sensor device to be installed in a proper position, the full coverage of a building construction site is achieved, and the accuracy of monitoring is improved. Further, in one or more embodiments of the present disclosure, before acquiring the detection information collected by the multi-type sensor devices at the construction site of the building engineering, it is further required to determine the layout position of the sensor devices, and the method includes the following steps:
firstly, obtaining a building design drawing of the current building engineering, and then establishing a BIM model of the current building engineering according to the Luban civil engineering technology. The building method is characterized in that the Luban civil engineering technology is a technology in the existing building field, so that a detailed description of the process of building a BIM model by the Luban civil engineering technology is omitted. And determining the typical scattered point position of the construction site of the current building engineering according to the acquired BIM model of the current building engineering. Wherein, it should be noted that typical scattered points are the positions where fire is easy to occur, including: vents, floor tops, circuit intersections, switches, etc. And determining a constructed range and a to-be-constructed range in the BIM according to the construction scheme of the current construction stage, so as to determine the matching relation between each typical scattered point position and the sensor equipment according to the detection attribute corresponding to each typical scattered point position in the constructed range and the to-be-constructed range. It should be noted that the detection attribute is a detection attribute corresponding to a problem that the current position is easy to appear, for example, the circuit intersection position is easy to generate electric sparks so as to generate smoke or cause fire, so that the detection attribute is smoke, and the typical scattered point is matched with the smoke sensor. And determining the initial layout position of the sensor equipment according to the determined matching relation between each typical scattered point position and the sensor equipment. And then acquiring the effective range of each type of sensor equipment in the initial layout position, and determining the coverage range of the type of sensor equipment according to the effective range of each type of sensor equipment. If it is determined that the coverage of this type of sensor device does not entirely cover the already constructed range and the to-be-constructed range of the current building construction site, then in order for the monitoring of the sensor to cover the entire area, it is necessary to acquire the remaining coverage of the already constructed range and the to-be-constructed range; wherein it is understood that the remaining coverage is a construction range and a region range to be constructed out of the coverage. In order to be able to sufficiently monitor the remaining coverage with the minimum of sensor devices, the present embodiment acquires each communication range of the remaining coverage, thereby acquiring the center point of each communication range as a supplementary layout position of the same type of sensor device. And then, arranging each sensor according to the initial arrangement position and the supplementary arrangement position so as to obtain detection information acquired by the multi-type sensor equipment on the construction site of the building engineering.
S103: and extracting a key frame image in the monitoring video in a preset monitoring period, and acquiring adjacent images of the key frame image, so that image features of differential images of the key frame image and the adjacent images are input into a preset classification recognition model, and whether a fire disaster target exists in the construction site of the building engineering and the position of the fire disaster target are output.
In order to lock a fire target based on a monitoring video, the problems that a monitoring picture of a plurality of monitoring devices is easy to miss and find out untimely are avoided. In the embodiment of the present disclosure, in order to achieve targeted monitoring and reduce analysis cost, first, a key frame image in a monitoring video is extracted within a preset detection period. At this time, the adjacent images of the key frame images are acquired, and objects which are obviously different from the background, such as smoke, open fire and the like, can appear when a fire disaster occurs. Therefore, differential images of the key frame images and the adjacent images are acquired, the alien objects appearing in the key frame images can be determined, whether a fire disaster target exists in a construction site of the building engineering can be determined based on the classification recognition model by inputting the image features of the differential images into the preset classification recognition model, and the position information of the fire disaster target can be determined according to the position of the fire disaster target in the key frame images.
Specifically, in one or more embodiments of the present disclosure, the extracting a key frame image in the surveillance video in a preset monitoring period, and acquiring an adjacent image of the key frame image, so as to input an image feature of a difference image between the key frame image and the adjacent image into a preset classification recognition model, and output whether a fire target exists in the construction site of the building engineering and the location of the fire target, specifically includes:
in order to improve analysis efficiency, the monitoring video is segmented according to preset analysis intervals to obtain a plurality of monitoring analysis segments, so that the monitoring analysis segments can be synchronously analyzed and processed based on a parallel analysis thread mode, analysis efficiency is improved, and fire disaster early warning and rescue time is timely carried out. After the plurality of monitoring analysis fragments are obtained by segmentation, the monitoring analysis fragments are collected according to a preset collection interval to obtain monitoring images of the monitoring analysis fragments, and it can be understood that the collection of the monitoring images according to the preset collection detection realizes the downsampling of the monitoring analysis fragments and reduces the data volume of analysis. And then determining the gray threshold value of each monitoring image according to the gray histogram of each monitoring image, so as to carry out binarization processing on the monitoring image according to the gray threshold value of each monitoring image in the monitoring analysis segment to obtain a binarized image of the monitoring image. And then comparing the binarized images in sequence to obtain a sudden change image in the monitoring image as a key frame image of the monitoring video. In order to obtain a sudden change target in a key frame image, in the embodiment of the present disclosure, a plurality of adjacent images of the key frame image are obtained, and according to a time sequence of the key frame image and the adjacent images, a differential image is sequentially obtained, and a connected region of the differential image is used as a hidden danger region. For example, after the key frame image a and the adjacent images thereof are arranged based on time sequence, an image C, an image D, an image a and an image B are formed, and then the differential images 1 and 2 of the images C and D and the images a and B are obtained in sequence, and then the obvious abnormality in the differential image 1 can be further clarified by comparing the differential image 1 and the differential image 2. In this case, the connected region in the differential image 1 is regarded as a hidden trouble region.
After the hidden danger area is determined, whether the hidden danger in the hidden danger area is fire disaster or not is further determined, so that effective fire safety monitoring for the building engineering is realized. In the embodiment of the present disclosure, the pixel points of the hidden danger area are converted into the YCbCr color model, so as to obtain the color characteristics of the hidden danger area. Namely, when the fire hazard exists, the texture color of the smoke or flame is changed, so that the color features of the hidden danger area are extracted as one type of features in the image features. Meanwhile, as the smoke is dispersed or the form expansion of the flame is provided with obvious sharp angles during combustion, the local highest point of the hidden danger area is obtained in the embodiment of the specification, and the sharp angle characteristics of the hidden danger area can be determined by comparing the local highest point with the adjacent points of the local highest point. In addition, as the situation of expansion occurs if the expansion is not processed in time after the occurrence of the fire disaster, the area growth characteristics of the hidden danger areas are determined according to the area transformation of the hidden danger areas. After extracting the image features of the color features, the sharp angle features and the area growth features of the differential image, inputting the image features and the key frame image into a preset classification and identification model, so that the preset classification and identification model identifies the image features and then outputs whether a fire disaster target exists in the construction site of the building engineering, and if the fire disaster exists, identifying the target position of the fire disaster based on the position of a communication area in the differential image. The method is characterized in that a preset classification recognition model is obtained by training a least square vector model.
S104: and determining a fire monitoring area based on the fire target position, meshing the fire area to obtain detection information corresponding to each boundary grid position, and determining material information corresponding to the detection information based on the material using position to take the detection information and the corresponding material information as fire state parameters of each boundary grid position.
After the fire target position is obtained according to the step S103, a timely early warning response is performed in order to timely predict the development of the fire. In the embodiment of the present disclosure, the range of the fire area in the dangerous range is determined according to the obtained location of the fire target, that is, the area in the preset range after the occurrence of the fire is likely to be affected by the fire target, so that the area needs to be used as a fire monitoring area for monitoring and predicting the development. Therefore, in the embodiment of the present disclosure, after determining the fire monitoring area according to the fire target, in order to facilitate prediction of the development of the fire target, the fire area is meshed so as to obtain detection information corresponding to each boundary mesh position of the fire area. And since inflammable materials can influence the development of fire, determining material information corresponding to the detection information according to the material use position, and taking the detection information and the corresponding material information as fire state parameters of each boundary grid position so as to facilitate the subsequent prediction of fire development based on the fire state parameters.
Specifically, in one or more embodiments of the present disclosure, a fire monitoring area is determined based on a fire target location, and the fire monitoring area is meshed to obtain detection information corresponding to each boundary grid location, and material information corresponding to the detection information is determined based on a material usage location, and the detection information and the corresponding material information are used as fire status parameters of each boundary grid location, including the following steps:
firstly, determining an area of a construction site of a building engineering within a preset safety radius as a fire monitoring area according to a fire target position and the preset safety radius. For example, if the preset safety radius is 50m, it is assumed that the area greater than 50m from the fire target position is a safety area, and the area less than or equal to 50m is a fire monitoring area of the construction site. And carrying out grid division on the fire area based on a preset grid size so as to determine the corresponding relation between each grid boundary position and the sensor equipment based on the layout position of the sensor equipment and the effective acquisition range of the sensor equipment, and acquiring detection information corresponding to each boundary grid position according to the corresponding relation. And then determining the matching relation between the material information and the detection information according to the effective acquisition range of each sensor device and the use position of the material, namely, if the use position of the material is within the effective acquisition range of the sensor device, indicating that the material information and the detection information acquired by the sensor device have the matching relation. And after the matching relation is obtained, determining material information corresponding to the detection information according to the matching relation, so that the detection information and the corresponding material information are used as fire state parameters of each boundary grid position.
S105: and inputting fire state parameters of each boundary grid position in the fire disaster area into a preset long-short-period memory network to obtain development trend prediction information of the fire target so as to trigger an alarm instruction of the fire target based on the fire target, the fire target position and the development trend prediction information.
In order to determine that a fire disaster occurs in a construction site of a building engineering based on the steps, as the fire disaster types, such as smoldering fire, bombing fire, non-bombing fire and the like, have different development trends due to the influence of the current environment, early warning response can be timely performed, in the embodiment of the specification, fire disaster state parameters of all boundary grid positions in a fire disaster area are input into a preset long-term and short-term memory network, so that development trend prediction information of a fire disaster target is obtained, and then corresponding fire disaster target warning instructions are triggered according to the fire disaster target, the fire disaster target positions and the development trend prediction information obtained by prediction.
Specifically, in one or more embodiments of the present disclosure, a fire status parameter of each boundary grid position in a fire area is input into a preset long-short-period memory network to obtain development trend prediction information of a fire target, which specifically includes:
In order to predict the development trend of the fire target spreading outside the fire area, the embodiment of the specification determines the initial spreading path grid corresponding to the boundary grid position according to the connection line between the boundary grid position and the fire target position, thereby determining the main air port type of the initial spreading path grid according to the air port type of each grid in the initial spreading path grid. Among these, it is understood that the tuyere types include: as shown in fig. 2 (a), the type of upper tuyere located at the upper tuyere of the fire target and the type of lower tuyere located at the lower tuyere of the fire target. In the case of a plurality of fire targets, the grid tuyere is in a tuyere state as long as the grid tuyere is positioned at the tuyere relative to a certain fire target. Then, determining intersection points of grid lines in the connecting lines as shown in fig. 2 (a), so as to obtain two initial spreading path grids corresponding to the intersection points of the grid lines as a first spreading path grid and a second spreading path grid. And then respectively determining first weight values of the first spreading path grid and the second spreading path grid according to the types of the air openings of the first spreading path grid and the second spreading path grid. It should be understood that the first weight value corresponding to the primary tuyere type is greater than the first weight value corresponding to the non-primary tuyere type, and the sum of the first weight value and the second weight value is 1. Then determining second weight values of the first spreading path grid and the second spreading path grid according to the number of fire-related grids of the preamble grids of the first spreading path grid and the second spreading path grid respectively; it should be noted that the second weight value is proportional to the number of fire-related grids. The number of fire-related grids of the first and second spreading path grids, that is, the number of grids having flames in the paths passing by when reaching the first or second spreading path grids. And then obtaining a total weight value according to the product of the first weight value and the second weight value, and determining a to-be-filtered spreading path grid in the first spreading path grid and the second spreading path grid by comparing the total weight values of the first weight value and the second weight value, so as to delete the to-be-filtered spreading path grid and obtain a spreading path grid. For example, as shown in fig. 2 (b), when the weight value of the first propagation path grid is smaller than that of the second propagation path grid, the first propagation path is deleted, that is, it is determined that the grid passing by the second propagation path grid as the propagation path in the grid crossing points is subjected to subsequent analysis, and the matching degree of the subsequent fire state parameter acquisition is improved by screening the propagation path from the fire point corresponding to the fire target to the boundary grid position, so that the prediction reliability of the development trend prediction information is improved on the premise of reducing the analysis cost. And then screening fire state parameters of boundary grid positions and the spreading path grids based on a Pearson correlation coefficient method to obtain data with higher correlation degree with fire development as key data. Then, in order to further improve the accuracy of prediction, the problem that the prediction result falls into a locally optimal solution is avoided. And carrying out parameter optimization on the initial long-short-term memory network according to a preset optimization algorithm to obtain an optimal parameter combination, so that the initial long-short-term memory network is updated according to the optimal parameter combination to obtain the preset long-short-term memory network. The parameters to be described include: neuron number, sliding time window width, initial learning value, etc. After the preset long-short-period memory network is obtained, key data of each boundary grid position are input into the preset long-short-period memory network for iterative prediction, and development trend prediction information corresponding to a fire disaster target is obtained.
Further, in one or more embodiments of the present disclosure, parameter optimization is performed on an initial long-short-term memory network based on a preset optimization algorithm to obtain an optimal parameter combination, so as to update the initial long-short-term memory network based on the optimal parameter combination to obtain the preset long-short-term memory network, and specifically includes the following steps:
firstly, determining the number of random combined populations of preset numbers of parameters in an initial long-short-term network, and determining the maximum iteration times of a preset optimization algorithm and search parameter nodes of the current random combined populations. And then determining the fitness of each node in each random combination population according to the distance between the search parameter node and other nodes in the random combination population. And then determining the current optimal position in the random combined population according to the acquired fitness. And updating the positions of the parameter combinations in the random combination population according to the preset coefficient vector and the maximum iteration times of a preset optimization algorithm, and obtaining the optimal position in the updated random combination population, wherein the parameter combination based on the node corresponding to the optimal position is used as the optimal parameter combination.
Further, in one or more embodiments of the present disclosure, in order to timely trigger an alarm instruction to improve fire safety of a building engineering, the alarm instruction for triggering a fire target based on the fire target, the location of the fire target, and the prediction information of the trend of development specifically includes the following steps:
If the fire type of the fire target is determined to be smoke based on the fire target, and the spreading probability of the fire target is determined to be lower than the preset spreading probability according to the development trend prediction information, the fact that the probability of the fire is lower at present is indicated, and therefore an audible and visual alarm instruction is triggered, audible and visual alarm is carried out, and constructors are reminded of timely avoiding risks. If the fire type of the fire target is determined to be open fire according to the fire target, or the spreading probability of the fire target is determined to be greater than the preset spreading probability based on the development trend prediction information, the fire alarm instruction is triggered at the same time when the audible and visual alarm instruction is triggered, so that the fire target position, the fire target and the fire target position are conveniently reported by generating the fire alarm instruction according to the template of the fire alarm instruction, the fire rescue is conveniently and timely carried out, the effect of timely reporting the fire situation is realized, and the rescue efficiency is improved.
As shown in fig. 3, an embodiment of the present disclosure provides an internal structure schematic diagram of a machine vision-based construction engineering safety detection device, and as can be seen from fig. 3, in one or more embodiments of the present disclosure, a machine vision-based construction engineering safety detection device includes:
At least one processor 301; the method comprises the steps of,
a memory 302 communicatively coupled to the at least one processor 301; wherein,,
the memory 302 stores instructions executable by the at least one processor 301 to enable the at least one processor 301 to perform any one of the methods described above.
As shown in fig. 4, the embodiment of the present disclosure provides a schematic internal structure of a nonvolatile storage medium, and as can be seen from fig. 4, in one or more embodiments of the present disclosure, a nonvolatile storage medium stores computer executable instructions, where the computer executable instructions can perform any one of the methods described above.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments. In particular, for apparatus, devices, non-volatile computer storage medium embodiments, the description is relatively simple, as it is substantially similar to method embodiments, with reference to the section of the method embodiments being relevant.
The foregoing describes specific embodiments of the present disclosure. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims can be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
The foregoing is merely one or more embodiments of the present description and is not intended to limit the present description. Various modifications and alterations to one or more embodiments of this description will be apparent to those skilled in the art. Any modification, equivalent replacement, improvement, or the like, which is within the spirit and principles of one or more embodiments of the present description, is intended to be included within the scope of the claims of the present description.
Claims (9)
1. A machine vision-based construction engineering safety monitoring method, the method comprising:
acquiring a current construction stage of a construction site of a building engineering so as to determine a current construction object and a current building bill of materials of the building engineering based on a construction scheme corresponding to the current construction stage; wherein, the bill of materials includes: material information and material use position;
Acquiring detection information acquired by multi-type sensor equipment of a construction site of a building engineering and monitoring videos acquired by monitoring equipment;
extracting a key frame image in the monitoring video in a preset monitoring period, and acquiring adjacent images of the key frame image, so that image features of differential images of the key frame image and the adjacent images are input into a preset classification recognition model, and whether a fire disaster target exists in the construction site of the building engineering and the position of the fire disaster target are output;
determining a fire monitoring area based on the fire target position, meshing the fire area to obtain detection information corresponding to each boundary grid position, and determining material information corresponding to the detection information based on the material using position to take the detection information and the corresponding material information as fire state parameters of each boundary grid position;
inputting fire state parameters of each boundary grid position in the fire disaster area into a preset long-short-period memory network to obtain development trend prediction information of the fire target so as to trigger an alarm instruction of the fire target based on the fire target, the fire target position and the development trend prediction information;
The method for obtaining the development trend prediction information of the fire target comprises the following steps of:
determining an initial spreading path grid corresponding to the boundary grid position according to the connection line of the boundary grid position and the fire target position, and determining a main air port type of each grid in the initial spreading path grid according to the air port type of each grid in the initial spreading path grid; wherein the tuyere type includes: the upper air port type is positioned at the upper air port of the fire disaster target, and the lower air port type is positioned at the lower air port of the fire disaster target;
determining grid line intersection points in the connecting line to obtain two initial spreading path grids corresponding to the grid line intersection points as a first spreading path grid and a second spreading path grid;
based on the wind gap types of the first spreading path grid and the second spreading path grid, respectively determining first weight values of the first spreading path grid and the second spreading path grid; the first weight value corresponding to the main tuyere type is larger than the first weight value corresponding to the non-main tuyere type, and the sum of the first weight value and the first weight value is 1;
Determining second weight values of the first spreading path grid and the second spreading path grid according to the number of fire-related grids of the preamble grids of the first spreading path grid and the second spreading path grid respectively; wherein the second weight value is proportional to the number of fire-related grids;
acquiring a total weight value based on the product of the first weight value and the second weight value, and determining a to-be-filtered spreading path grid in the first spreading path grid and the second spreading path grid by comparing the total weight value so as to delete the to-be-filtered spreading path grid and obtain a spreading path grid;
screening fire state parameters of the boundary grid positions and the spreading path grids based on a Pearson correlation coefficient method to obtain data with higher fire development correlation degree as key data;
performing parameter optimization on an initial long-short-term memory network based on a preset optimization algorithm to obtain an optimal parameter combination, and updating the initial long-short-term memory network based on the optimal parameter combination to obtain a preset long-short-term memory network; wherein the parameters include: the number of neurons, the width of a sliding time window and an initial learning value;
And inputting the key data of each boundary grid position into a preset long-term and short-term memory network for iterative prediction so as to obtain development trend prediction information corresponding to the fire target.
2. The machine vision-based building engineering safety monitoring method according to claim 1, wherein the parameter optimization is performed on an initial long-short term memory network based on a preset optimization algorithm to obtain an optimal parameter combination, so as to update the initial long-short term memory network based on the optimal parameter combination to obtain a preset long-short term memory network, and the method specifically comprises the following steps:
determining the number of random combined populations of preset numbers of parameters in the initial long-short term network, determining the maximum iteration number of the preset optimization algorithm, and determining the current searching parameter nodes of the random combined populations;
determining the adaptability of each node in each random combined population based on the distance between the searching parameter node and other nodes in the random combined population, and determining the current optimal position in the random combined population according to the adaptability;
and updating the positions of all parameter combinations in the random combination population according to the preset coefficient vector of the preset optimization algorithm and the maximum iteration times, and acquiring the optimal positions in the updated random combination population, wherein the parameter combinations based on the nodes corresponding to the optimal positions are used as optimal parameter combinations.
3. The machine vision-based building engineering safety monitoring method according to claim 1, wherein the step of obtaining the current construction stage of the construction site of the building engineering to determine the current construction object and the current building bill of materials of the building engineering based on the construction scheme corresponding to the current construction stage comprises the following steps:
acquiring a construction stage schedule corresponding to the current construction project of the construction site, so as to determine an adjustment value of the construction stage schedule based on a difference value between the end time of the previous construction stage and the planned end time of the previous construction stage in the construction stage schedule;
adjusting the construction stage schedule according to the adjustment value to obtain a current construction stage schedule of the construction site of the building engineering;
acquiring a current construction stage of the building construction site based on the current construction stage schedule and the current time;
and acquiring a construction scheme corresponding to the current construction stage according to the construction plan of the building engineering so as to determine a current construction object and a current building bill of materials of the building engineering based on the construction scheme.
4. A machine vision based construction safety monitoring method according to claim 3, wherein before obtaining the detection information collected by the multi-type sensor device at the construction site of the construction project, the method further comprises:
acquiring a building design drawing of the current building engineering to establish a BIM model of the current building engineering based on a Luban civil engineering technology;
determining a typical scattered point position of a construction site of the current building engineering based on a BIM model of the current building engineering; wherein the typical scatter points include: the air vent, the floor slab top, the circuit crossing point and the electric switch;
determining a constructed range and a to-be-constructed range in the BIM based on the construction scheme of the current construction stage, so as to determine the matching relation between each typical scattered point position and sensor equipment according to detection attributes corresponding to each typical scattered point position in the constructed range and the to-be-constructed range;
determining an initial layout position of the sensor equipment based on the matching relation, acquiring effective ranges of the sensor equipment of each type in the initial layout position, and determining the coverage range of the sensor equipment of each type based on the effective ranges of the sensor equipment of each type;
If the coverage area is determined to not fully cover the constructed range and the to-be-constructed range, acquiring the residual coverage areas of the constructed range and the to-be-constructed range; wherein the residual coverage range is the constructed range and the area range except the coverage range in the to-be-constructed range;
acquiring each communication range of the residual coverage range to acquire a central point of each communication range as a supplementary layout position of the sensor equipment of the same type;
and laying each sensor based on the initial laying position and the supplementary laying position so as to acquire detection information acquired by multi-type sensor equipment on the construction site of the building engineering.
5. The machine vision-based building engineering safety monitoring method according to claim 1, wherein the steps of extracting a key frame image in the monitoring video within a preset monitoring period, and acquiring adjacent images of the key frame image, so as to input image features of a difference image between the key frame image and the adjacent images into a preset classification and identification model, and outputting whether a fire target exists in the building engineering construction site and the fire target position comprise:
Segment segmentation is carried out on the monitoring video based on a preset analysis interval to obtain a plurality of monitoring analysis segments, so that the monitoring analysis segments are acquired according to a preset acquisition interval to obtain monitoring images of the monitoring analysis segments;
determining a gray threshold value of each monitoring image according to the gray histogram of each monitoring image so as to determine a binarized image of each monitoring image based on the gray threshold value of each monitoring image in the monitoring analysis segment;
sequentially comparing the binarized images, determining a sudden change image in the monitoring image as a key frame image of the monitoring video to obtain a plurality of adjacent images of the key frame image, sequentially obtaining differential images based on the time sequence of the key frame image and the adjacent images, and taking a communication area of the differential images as a hidden danger area;
converting pixel points of the hidden danger areas into a YCbCr color model to acquire color characteristics of the hidden danger areas, and acquiring local highest points of the hidden danger areas, so as to determine sharp angle characteristics of the hidden danger areas based on comparison between the local highest points and adjacent points of the local highest points, and determine area growth characteristics of the hidden danger areas according to area transformation of each hidden danger area;
Inputting the color characteristics, the sharp angle characteristics, the area increase characteristics and the key frame images into the preset classification and identification model to output whether a fire target exists in the construction site of the building engineering and the position of the fire target; the preset classification recognition model is a least square vector model.
6. The machine vision based construction engineering safety monitoring method according to claim 4, wherein determining a fire monitoring area based on the fire target position, meshing the fire area to obtain detection information corresponding to each boundary mesh position, and determining material information corresponding to the detection information based on the material usage position to take the detection information and the corresponding material information as fire state parameters of each boundary mesh position, comprises:
determining a fire monitoring area of the construction site of the building engineering based on the fire target position and a preset safety radius;
performing grid division on the fire disaster area based on a preset grid size, determining the corresponding relation between each grid boundary position and the sensor equipment based on the layout position of the sensor equipment and the effective acquisition range of the sensor equipment, and acquiring detection information corresponding to each boundary grid position according to the corresponding relation;
And determining a matching relation between the material information and the detection information according to the effective acquisition range of each sensor device and the use position of the material, and determining the material information corresponding to the detection information according to the matching relation, wherein the detection information and the corresponding material information are used as fire state parameters of each boundary grid position.
7. The machine vision-based building engineering safety monitoring method according to claim 1, wherein the alarm instruction for triggering the fire target based on the fire target, the fire target position and the development trend prediction information specifically comprises:
if the fire type of the fire target is determined to be smoke based on the fire target, and the spreading probability of the fire target is determined to be lower than the preset spreading probability based on the development trend prediction information, triggering an audible and visual alarm instruction;
if the fire type of the fire target is determined to be open fire based on the fire target, or the spreading probability of the fire target is determined to be greater than the preset spreading probability based on the development trend prediction information, triggering an audible and visual alarm instruction and a fire alarm instruction;
And generating a fire alarm instruction based on the fire alarm instruction template and reporting the fire target position, the fire target and the fire target position so as to timely perform fire rescue.
8. A machine vision-based construction engineering safety monitoring device, the device comprising:
at least one processor; the method comprises the steps of,
a memory communicatively coupled to the at least one processor; wherein,,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of the preceding claims 1-7.
9. A non-volatile storage medium storing computer executable instructions, characterized in that the computer executable instructions are capable of performing the method of any of the preceding claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310896113.3A CN116612440B (en) | 2023-07-21 | 2023-07-21 | Building engineering safety monitoring method, equipment and medium based on machine vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310896113.3A CN116612440B (en) | 2023-07-21 | 2023-07-21 | Building engineering safety monitoring method, equipment and medium based on machine vision |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116612440A CN116612440A (en) | 2023-08-18 |
| CN116612440B true CN116612440B (en) | 2023-09-29 |
Family
ID=87684076
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310896113.3A Active CN116612440B (en) | 2023-07-21 | 2023-07-21 | Building engineering safety monitoring method, equipment and medium based on machine vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116612440B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117312828B (en) * | 2023-09-28 | 2024-06-14 | 光谷技术有限公司 | Public facility monitoring method and system |
| CN117455270A (en) * | 2023-12-22 | 2024-01-26 | 成都理工大学 | Fire emergency treatment method, system, electronic equipment and medium |
| CN117876969B (en) * | 2024-03-11 | 2024-06-04 | 贵州省公路建设养护集团有限公司 | Safety monitoring method and system for bridge construction |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112287849A (en) * | 2020-10-30 | 2021-01-29 | 武汉理工光科股份有限公司 | Fire early warning method and device for high-rise building |
| CN112972972A (en) * | 2021-03-31 | 2021-06-18 | 湖南现代德雷工程有限公司 | Intelligent fire fighting method and system for building safety construction site |
| CN114757387A (en) * | 2022-03-07 | 2022-07-15 | 南京林业大学 | Forest fire danger state monitoring and analyzing system and method based on artificial intelligence |
| CN115099573A (en) * | 2022-05-25 | 2022-09-23 | 河南鑫安利安全科技股份有限公司 | Safety production management platform |
| CN115983115A (en) * | 2022-12-20 | 2023-04-18 | 厦门大学 | Indoor fire evolution prediction method based on deep neural network |
| CN116012780A (en) * | 2022-12-22 | 2023-04-25 | 广西电网有限责任公司河池供电局 | Fire disaster monitoring method and system based on image recognition |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11966670B2 (en) * | 2018-09-06 | 2024-04-23 | Terrafuse, Inc. | Method and system for predicting wildfire hazard and spread at multiple time scales |
| US11660480B2 (en) * | 2018-11-21 | 2023-05-30 | One Concern, Inc. | Fire forecasting |
| KR102087000B1 (en) * | 2019-08-13 | 2020-05-29 | 주식회사 지에스아이엘 | Method And System for Monitoring Fire |
-
2023
- 2023-07-21 CN CN202310896113.3A patent/CN116612440B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112287849A (en) * | 2020-10-30 | 2021-01-29 | 武汉理工光科股份有限公司 | Fire early warning method and device for high-rise building |
| CN112972972A (en) * | 2021-03-31 | 2021-06-18 | 湖南现代德雷工程有限公司 | Intelligent fire fighting method and system for building safety construction site |
| CN114757387A (en) * | 2022-03-07 | 2022-07-15 | 南京林业大学 | Forest fire danger state monitoring and analyzing system and method based on artificial intelligence |
| CN115099573A (en) * | 2022-05-25 | 2022-09-23 | 河南鑫安利安全科技股份有限公司 | Safety production management platform |
| CN115983115A (en) * | 2022-12-20 | 2023-04-18 | 厦门大学 | Indoor fire evolution prediction method based on deep neural network |
| CN116012780A (en) * | 2022-12-22 | 2023-04-25 | 广西电网有限责任公司河池供电局 | Fire disaster monitoring method and system based on image recognition |
Non-Patent Citations (3)
| Title |
|---|
| 分区域变尺度网格的社区火灾场景数值仿真;张琳;黄晓霞;李红旮;李霞;;中国图象图形学报(第09期);全文 * |
| 城市火灾视频监控目标区域图像准确检测仿真;杨柳;张德;王亚慧;衣俊艳;;计算机仿真(第04期);全文 * |
| 基于YOLOv3算法的高速公路火灾检测;刘俊;张文风;;上海船舶运输科学研究所学报(第04期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116612440A (en) | 2023-08-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116612440B (en) | Building engineering safety monitoring method, equipment and medium based on machine vision | |
| CN101795395B (en) | System and method for monitoring crowd situation | |
| CN108389359B (en) | Deep learning-based urban fire alarm method | |
| CN105763853A (en) | Emergency early warning method for stampede accident in public area | |
| CN101751744B (en) | Detection and early warning method of smoke | |
| CN112132090A (en) | Smoke and fire automatic detection and early warning method based on YOLOV3 | |
| CN106991668B (en) | Evaluation method for pictures shot by skynet camera | |
| CN101835035B (en) | Regional invasion real-time detection method | |
| CN111047818A (en) | Forest fire early warning system based on video image | |
| US20220167142A1 (en) | Digital twin disaster management system customized for underground public areas | |
| CN112488898B (en) | Emergency handling system and method based on intelligent building management and control | |
| US20210150692A1 (en) | System and method for early identification and monitoring of defects in transportation infrastructure | |
| CN112184773A (en) | Helmet wearing detection method and system based on deep learning | |
| CN105574468A (en) | Video flame detection method, device and system | |
| CN107688793A (en) | A kind of outside transformer substation fire automatic monitoring method for early warning | |
| CN2785052Y (en) | Underground fire disaster intelligent monitoring and alarming device | |
| CN114444386A (en) | Fire early warning and post-disaster floor damage prediction method and system based on BIM and deep learning | |
| CN114689058A (en) | Fire evacuation path planning method based on deep learning and hybrid genetic algorithm | |
| CN114677640A (en) | Intelligent construction site safety monitoring system and method based on machine vision | |
| CN112949536B (en) | Fire alarm method based on cloud platform | |
| Wu et al. | Post-flood disaster damaged houses classification based on dual-view image fusion and Concentration-Based Attention Module | |
| KR20240028290A (en) | Device and method for detecting wildfire | |
| CN114882410B (en) | Tunnel dome lamp fault detection method and system based on improved positioning loss function | |
| JP7372446B2 (en) | Image analysis system, image analysis method and image analysis program | |
| CN114519834A (en) | High-rise fire hazard early warning method, device and application |
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 |