CN113850562A - Intelligent side station supervision method and system - Google Patents
Intelligent side station supervision method and system Download PDFInfo
- Publication number
- CN113850562A CN113850562A CN202111031812.9A CN202111031812A CN113850562A CN 113850562 A CN113850562 A CN 113850562A CN 202111031812 A CN202111031812 A CN 202111031812A CN 113850562 A CN113850562 A CN 113850562A
- Authority
- CN
- China
- Prior art keywords
- construction
- image data
- data
- side station
- basic information
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- 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
- G06Q10/00—Administration; Management
- G06Q10/10—Office automation; Time management
- G06Q10/103—Workflow collaboration or project management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06K—GRAPHICAL DATA READING; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
- G06K17/00—Methods or arrangements for effecting co-operative working between equipments covered by two or more of main groups G06K1/00 - G06K15/00, e.g. automatic card files incorporating conveying and reading operations
- G06K17/0022—Methods or arrangements for effecting co-operative working between equipments covered by two or more of main groups G06K1/00 - G06K15/00, e.g. automatic card files incorporating conveying and reading operations arrangements or provisious for transferring data to distant stations, e.g. from a sensing device
- G06K17/0025—Methods or arrangements for effecting co-operative working between equipments covered by two or more of main groups G06K1/00 - G06K15/00, e.g. automatic card files incorporating conveying and reading operations arrangements or provisious for transferring data to distant stations, e.g. from a sensing device the arrangement consisting of a wireless interrogation device in combination with a device for optically marking the record carrier
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Evolutionary Biology (AREA)
- Strategic Management (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Computer Networks & Wireless Communication (AREA)
- Economics (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to an intelligent side station supervision method and system. The method comprises the following steps: inputting basic information of the inspection objects involved in side station supervision, and generating and printing object two-dimensional codes of the inspection objects; pasting the object two-dimensional code on the corresponding inspection object; during project acceptance, scanning an object two-dimensional code of an inspection object related to a construction project through an intelligent terminal, acquiring basic information of each inspection object, acquiring field acceptance image data, filling the basic information of the inspection object and the field acceptance image data into corresponding positions of a preset table, and generating a related record table; acquiring construction process data in real time, and monitoring the construction process; and acquiring the image data of the construction site in real time, identifying the non-conforming items in the image data of the construction site, and automatically generating a notice sheet of the non-conforming items. The invention utilizes informatization and automation means to assist the side station supervision personnel to improve the side station supervision quality, improve the working efficiency, reduce the labor cost and save the human resources.
Description
Technical Field
The invention relates to the technical field of intelligent management of engineering construction projects, in particular to an intelligent side station supervision method and system.
Background
Side station supervision refers to supervision activities of a project supervision organization on construction quality of key parts or key processes of a project, and supervision of side stations is a common method for project management. Supervision typically performs side-station supervision on more important construction processes.
At present, the side station supervision usually comprises that a side station supervision person utilizes communication facilities such as a mobile phone, the construction unit shoots or records the construction process of a key process, retains image data, uses and records the record or mobile phone shooting key technical parameters in real time, fills a side station record form after the side station is finished to form data signature and retention, the side station supervision person needs to supervise the construction person all the time, the side station supervision person can check and accept after the whole process is finished, if a non-conforming item is found, a non-conforming item notice is required to be additionally filled, and the transfer and approval of the notice need to be manually transmitted.
The side station supervision labor cost is extremely high, supervision personnel are limited, and a large number of personnel are required to participate in the side station; limited personnel can not be in place at any time to carry out side standing, slow the project progress and reduce the project efficiency; meanwhile, the dependence of a construction unit is caused and depends on the management of site supervision, the side station supervision is limited by human factors, the quality of the side station depends on the quality and level of supervision personnel, and once the responsibility of the side station supervision personnel and related operating personnel is lost, the inestimable result is easily caused. Besides, the storage and non-conformity paper notice circulation examination and approval time of the records of the supervision side station is long, the occupied space is large, and the requirement on the storage place is high.
Disclosure of Invention
The invention aims to solve the technical problem in the prior art and provides an intelligent side station supervision method and system.
In order to solve the above technical problem, an embodiment of the present invention provides an intelligent side station supervision method, including the following steps:
inputting basic information of inspection objects related to side station supervision, and generating and printing object two-dimensional codes of the inspection objects according to the basic information of the inspection objects; pasting the object two-dimensional code on a corresponding inspection object;
acquiring construction process data in real time through a data acquisition instrument, and monitoring the construction process according to the construction process data; acquiring construction site image data in real time through an image acquisition device, identifying non-conforming items in the construction site image data by using a non-conforming item identification algorithm, and automatically generating a non-conforming item notice;
when the project is accepted, the intelligent terminal scans the object two-dimensional codes of the inspection objects related to the construction project, acquires the basic information of each inspection object, acquires the field acceptance image data, fills the basic information of the inspection objects and the field acceptance image data into the corresponding positions of the preset table, and generates the related record table.
In order to solve the above technical problem, an embodiment of the present invention further provides an intelligent side station supervision system, including: the system comprises an intelligent terminal, a data acquisition instrument, an image acquisition device and a project management platform;
the intelligent terminal is used for inputting basic information of the inspection objects related to side station supervision, and generating and printing object two-dimensional codes of the inspection objects according to the basic information of the inspection objects; pasting the object two-dimensional code on a corresponding inspection object; acquiring basic information of each inspection object through an object two-dimensional code of the inspection object related to a construction project, acquiring field acceptance image data, filling the basic information of the inspection object and the field acceptance image data into corresponding positions of a preset table, generating a related record table, and sending the related record table to a project management platform;
the data acquisition instrument is used for acquiring construction process data in real time; the image acquisition device is used for acquiring image data of a construction site in real time; sending the construction process data and the construction site image data to a project management platform;
and the project management platform is used for monitoring the construction process according to the construction process data, identifying the non-conforming items in the construction site image data by using a non-conforming item identification algorithm, and automatically generating a non-conforming item notice.
The invention has the beneficial effects that: on will contain the two-dimensional code of inspection object basic information to examine the object, when carrying out the project and receiving, can conveniently acquire and objectively input the basic information of the various inspection objects that the construction project relates to, such as personnel, equipment, material and construction code etc., conveniently automatic generation record form, record form sends to project management platform, realize functions such as online circulation approval, acquire work process data and construction site image data in real time, can realize the remote monitoring to work progress and nonconformity item, thereby supplementary side station supervision personnel improve side station supervision quality, promote work efficiency, the cost of labor has been reduced, human resources have been saved.
Additional aspects of the invention and its advantages will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention.
Drawings
Fig. 1 is a flowchart of an intelligent side station supervision method according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a crater record form of a welding project according to an embodiment of the present invention;
fig. 3 is a block diagram of an intelligent wayside supervision system according to an embodiment of the present invention.
Detailed Description
The principles and features of this invention are described below in conjunction with the following drawings, which are set forth by way of illustration only and are not intended to limit the scope of the invention.
Fig. 1 is a flowchart of an intelligent side station supervision method according to an embodiment of the present invention. As shown in fig. 1, the method includes:
s110, inputting basic information of the inspection objects related to side station supervision, and generating and printing object two-dimensional codes of the inspection objects according to the basic information of the inspection objects; and pasting the object two-dimensional code on the corresponding inspection object.
Specifically, the inspection object may include a person, equipment, material, construction code, and the like. The construction code is in the form of number or English letter abbreviation and represents the number of the construction part. Comprises the following steps: single project, unit project, subsection project, inspection lot name, construction professional category and procedure and the like.
In the construction preparation stage, main constructor information, equipment information, material information, construction code information and the like are input by constructors through an intelligent terminal (such as a mobile phone, a tablet personal computer and the like) or a PC (personal computer) end, a two-dimensional code is formed to be printed, the two-dimensional code is printed and pasted on the inspection object, supervision and inspection personnel can acquire construction information through scanning the code, and related construction information directly enters a system background after being input at an APP (application) end, so that construction data can be conveniently formed. The method uses informatization means to replace handwritten data in the original side station process to record basic information such as constructor information, mode position and the like.
S120, acquiring construction process data in real time through a data acquisition instrument, and monitoring the construction process according to the construction process data; the construction site image data are acquired in real time through an image acquisition device, non-conforming items in the construction site image data are identified through a non-conforming item identification algorithm, and a non-conforming item notice is automatically generated.
Wherein, the data acquisition instrument can adopt acquisition devices such as a current sensor, a voltage sensor, a gas flow sensor, a temperature sensor, a humidity sensor and the like. Taking a welding project as an example, the welding data of the welding machine in the welding process can be collected through the welding data collector, and the welding data (parameters such as welding current, welding voltage, welding line energy, wire feeding speed and the like) is transmitted to the welding front cloud server through the infinite AP and then to the project management platform through the infinite AP by the welding data collector. The project management platform compares the real-time collected data with the set limit values such as heat input quantity, wire feeding speed and the like, if abnormity occurs, welding is intelligently interrupted, a reminder is sent to a side station supervisor, and the supervisor receives the reminder and then goes to the process.
The image acquisition device can adopt fixed cameras, handheld mobile recorders, unmanned aerial vehicles and other acquisition devices. And acquiring pictures or video data in the construction process by using equipment such as a fixed camera, a handheld mobile camera and the like, and transmitting the pictures or the video data to the project management platform in real time. Building a video monitoring system capable of being remotely controlled and a server for receiving data according to actual conditions on a construction site, wherein video monitoring images are not lower than the requirement of a D4 standard, and data exchange of camera data and a transmitting terminal in a camp is realized through a wireless receiving terminal; the data of the hard disk video recorder are exchanged with the data of the transmitting terminal through the switch, wired optical cables can be used for transmission, AI intelligent technology can be loaded on obtained images and video data, an intelligent identification system is loaded on the video data shot by the camera, violation behaviors in the video data can be identified, snapshot imaging is carried out, the identified violation data are formed into inconsistent item notification sheets through the project management platform and are pushed to relevant personnel, meanwhile, side station supervision personnel can remind the content to go to the site in time according to pushing to carry out supervision and inspection, voice equipment can also be arranged on the video monitoring equipment, and the violation behaviors are found and are directly shout through the voice system on the video monitoring equipment in time to stop the violation behaviors on the site.
The camera uses the setting: after the camera is started, the camera is set to be in a circulating shooting function, and the automatic shutdown is set to be manual. Each video time can be adjusted to 5 minutes, 10 minutes and 15 minutes according to the welding time of the weld craters, so that later-stage video editing is facilitated, and the monitoring code stream is set to be in a 380 × 480p30 mode. The camera is fixed on the movable support, and after the camera is started, the shooting position of the movable camera is adjusted according to different process requirements (for example, in the shooting welding process, the distance between the camera and a weld crater is approximately 1.6 meters, the height is 0.5 meter, a lower rod is supported by 400mm, and an upper rod is supported by 250mm) and a laser spot is required to be at the weld crater position. After construction is completed, a person in charge copies the video in a local area network mode and then puts the video into a designated folder, classifies the video, and well performs link and file filing. And merging the videos exceeding the time by using editing software, then renumbering, and linking the archived files to the intelligent judgment system for the non-conforming items.
The drone patrol system may include a multi-rotor drone and a ground signal transmission base station. Many rotor unmanned aerial vehicle have enough high pixel, sufficient battery capacity and anti-wind ability, can guarantee whole record high quality, stable construction video.
Specifically for unmanned aerial vehicle operating personnel accomplishes the setting-up with cell-phone APP connection unmanned aerial vehicle after, control unmanned aerial vehicle rises to the air and hovers at the construction point that needs other station, adjustment camera direction and focus, gather video data and transmit to project management platform through ground basic station real-time, unmanned aerial vehicle not only can gain with the same visual angle of other station of prison, can also gain the height and the angle that can't reach of prison, all-round construction video data of getting, and can keep the construction video except other station form as the process data, unmanned aerial vehicle patrols and examines and equally can load non-conformity item intelligent recognition system, the discovery is the action of violating the regulations, take a candid photograph the formation of image propelling movement to relevant managers.
And S130, during project acceptance, scanning the object two-dimensional codes of the inspection objects related to the construction project through the intelligent terminal, acquiring basic information of each inspection object, acquiring field acceptance image data, filling the basic information of the inspection objects and the field acceptance image data into corresponding positions of a preset table, and generating a related record table. The related record table may include a construction record table for sending to a construction unit and a side-station supervision record table for sending to a supervision unit, and the like.
It should be noted that the execution steps do not have a strict execution sequence, and the construction process data and the construction site image data are collected in real time.
In the above embodiment, on pasting the two-dimensional code including the basic information of the inspection object on the inspection object, when the project is accepted, the basic information of various inspection objects related to the construction project can be conveniently acquired and objectively input, such as personnel, equipment, materials, construction codes and the like, a record form can be conveniently and automatically generated, the record form is sent to a project management platform, online circulation approval is realized, the construction process data and the construction site image data are acquired in real time, the remote monitoring on the construction process and the non-conforming items can be realized, thereby assisting the side station supervision personnel to improve the side station supervision quality, improving the working efficiency, reducing the labor cost and saving the human resources.
The method comprises the steps that scanning keys or shooting keys are respectively arranged at different entry items of a preset table, the scanning function of the corresponding entry item is started through the scanning keys, basic information of an inspection object is obtained through scanning an object two-dimensional code of the inspection object, the obtained basic information of the inspection object is filled in the corresponding entry item, the shooting function of the corresponding entry item is started through the shooting keys, field acceptance image data is obtained, and the field acceptance image data is filled in the corresponding entry item.
As shown in fig. 2, taking a crater record form of a welding project as an example, the crater record form can be used for inputting information of a crater number, a front management number and a rear management number through a scanning key 210, and other contents displayed in the form can be manually input or can be made into a two-dimensional code form for scanning and inputting; the on-site acceptance image can be obtained through the shooting key 220 and filled in the corresponding position of the form, so that the recording form can be conveniently and quickly generated. In the embodiment, the scanning key or the shooting key is arranged on the record form, so that the basic information of the inspection object can be conveniently acquired, the image data can be accepted on site, and the record form can be conveniently and quickly generated.
Other conventional intelligent recognition systems currently adopt a Convolutional Neural Network (CNN) structure as a classifier of a main recognition algorithm. Because the CNN has weak translation invariance, the robustness is poor in a complex scene, for example, in a warning identification recognition scene of a construction unit, if the identification is partially damaged or is shielded by other objects, a model of a recognition algorithm using the CNN as a core may have false positive errors. Meanwhile, objects to be identified (such as safety helmets and work clothes) need to meet the unified standard, a large number of samples need to be manufactured for the classifier to learn, and the labor cost, the time cost and the material cost are extremely high.
In the embodiment of the invention, identifying the non-conforming item in the construction site image data by using a non-conforming item identification algorithm comprises the following steps: recognizing and judging the preset object by using ImageNet to assist an SVM algorithm; the PCA is used for assisting GNN and YOLO algorithm to realize the identification of personnel and the detection of behaviors; and judging whether the pipeline installation is standard or not by using an LSD straight line judgment algorithm.
In the embodiment of the invention, different algorithms are applied to respectively realize the identification of objects and people on the operation site and the judgment of the behaviors of the people, and the auxiliary technology is utilized to improve the efficiency and the identification accuracy and reduce the cost. The embodiment of the invention uses GNN, SVM and YOLO as recognition algorithms, mainly proceeds classification and discrimination of targets, and realizes separation of ROI (region of interest) and background. And a linear extraction algorithm LSD for judging whether the pipe trench is deformed. ImageNet and PCA technologies are utilized to provide support for the algorithm. The ImageNet provides a sample library, and the PCA mainly reduces the dimension of the data, so that the cost is saved, and the operation complexity is reduced.
Specifically, the method for recognizing and judging the preset object by using the ImageNet auxiliary SVM algorithm comprises the following steps: similar samples of preset objects needing to be detected in a construction scene are screened out in ImageNet to form a training set, and the training set is used for training to generate a deep learning model; sampling image data of a specific and standard-meeting preset object required by a construction scene during operation to generate a self-made data set; converting the solution space mapping of the full connection layer of the deep learning model into an SVM, reserving the trained parameter models of each layer, and performing transfer learning on the deep learning model by using the self-made data set to complete the integral training of a discrimination model; and identifying the image data of the actual construction scene by using the discrimination model finished by the integral training, and judging that a non-conforming item exists when the identification result is smaller than a preset threshold value.
The principle of the recognition algorithm applied by the embodiment of the invention is as follows:
1. and identifying and judging objects such as work clothes, safety helmets, operation tickets and the like by using ImageNet to assist the SVM algorithm. The non-compliance that can be identified is as follows: 1. the safety helmet is not worn; 2. no operation ticket is posted at dangerous construction operation places such as a limited space; 3. the field warning sign is not provided with a fence; 4. the labor insurance products of constructors are not regularly worn; 5. the constructors in the special operation site do not wear the red work clothes; 6. seat belt use is not compliant; 7. no traction rope is used in hoisting operation; 8. the fire fighting equipment is not placed in a standard way.
ImageNet is an open-source visual database, and comprises 1400 million pictures with labels, wherein the pictures comprise safety helmets, warning signs, constructor working clothes and the like which need to be detected by the invention. And the sample in ImageNet is standard, and the compatibility of the data format on the model is strong.
In the embodiment of the invention, firstly, required samples are screened in ImageNet, similar samples such as safety helmets and work clothes which need to be detected in an actual field are selected, and the selected training set is utilized for training to generate a deep network model; and then sampling image data of a preset object which is required by a construction scene and meets the standard during operation to generate a self-made data set, and performing migration training on the deep network model by using the self-made data set to ensure that the discrimination of the model meets the unified standard to obtain a discrimination model. After the model is initially trained, the problem of accurate recognition in a complex scene still cannot be well solved as the CNN technology is used. In order to improve the recognition accuracy in a complex scene, the solution space mapping of the full connection layer of the discrimination model is converted into the SVM, each trained layer parameter model is reserved, and migration learning is carried out by utilizing a self-made data set corresponding to a preset construction scene to complete the whole training of the discrimination model. And most of data sets contained in ImageNet are sampled abroad, and the standards and the details of the samples are different from those of the actual construction site, so that the requirements of the actual construction scene can be better met by using the SVM and a small amount of self-made data sets for transfer learning, the accuracy and the practicability of algorithm identification are ensured, the method is suitable for the actual application scene, and the accurate identification of cross-Domain (Domain) is realized. After the whole model is trained, the video monitoring is carried out on the site by using the unmanned aerial vehicle and the camera equipment for the detected application scene, the system judges preset articles such as tools, safety helmets and the like worn by constructors in the video, the judgment result is represented by the probability of 0-1, the probability is closer to 0, the more probable the probability is, the more standard the behavior of the identified object is, and the more probable the probability is, the 1 is, the more standard the behavior of the identified object is. The constructor can flexibly set a threshold value according to the requirement so as to improve the accuracy and the precision of identifying the application scene by the non-conforming item. The ImageNet auxiliary SVM algorithm can effectively utilize the existing data set to carry out deep learning, reduces the cost of data collection and solves the problems that the common algorithm CNN is poor in robustness and not accurate enough for complex scene recognition.
2. And the PCA is used for assisting GNN and YOLO algorithm to realize the identification of personnel and the detection of behaviors. The non-conforming term that GNN (graph neural network) can recognize is 1, the person and vehicle lingers in violation. 2. Gathering personnel on the line and illegal excavation. The YOLO algorithm is able to identify non-conformers as follows: 1. the personnel rest on the crane boom, the motor flywheel, the heavy object to be lifted and beside the electrical equipment; 2. lifting illegal operation-a crane person getting off the station; 3. standing on a hoisted object being hoisted; 4. dangerous construction operation places such as limited space and the like are not supervised by safety personnel.
First, an interaction relationship is established using an interacted object as a clue (person-object interaction), a model regarding a pose (position) is established, classification is performed by counting distribution of the pose (or more generally, a component), and a model regarding skeleton information of the person 2d is established using a Graph Neural Network (GNN). Firstly, obtaining 2d human skeleton information from a training sample image, namely the positions of all joints, and then connecting the positions according to a predefined sequence to form a complete skeleton of a person. And training the model by taking the extracted physical coordinates of the human skeleton in the scale space as training data. After training, the system can abstract the skeleton characteristics of constructors by using an unmanned aerial vehicle or monitoring equipment to generate a 2d abstract model, for detecting character aggregation, only the distance between the 2d character models needs to be compared, meanwhile, the number of samples with the model distance smaller than a specified range is counted, and when the number reaches a threshold value, the system judges that the characters are aggregated. And comparing the generated 2d character model with the 2d model of the gesture of the smoker in the scale space, and judging by utilizing the coordinates of the graph node where the gesture is located.
And detecting the illegal personnel station positions by using a YOLO algorithm, such as non-conforming items when constructors are illegally standing under large-scale equipment such as a crane and the like. The recognition algorithm comprises two detection parts: firstly, people are detected, namely whether constructors exist in a scene or not, and secondly, large equipment is detected. In the identification scene, the system distinguishes certain frames in the video through the unmanned aerial vehicle or the monitoring equipment, and the distinguishing process is as follows: the image is adjusted to a specific size after being modeled, the image is divided into a plurality of small lattices, people and large equipment are dispersed in the small lattices, each small lattice corresponds to N (N represents the number of targets to be detected) bounding boxes, the wide and high ranges of the bounding boxes are full graphs, and the positions of the bounding boxes of the target people or the large equipment are searched by taking the small lattices as centers. And after the boundary box where the target is located is obtained, comparing the distance between the centers of the boundary boxes, and if the distance is smaller than the safety distance, judging that the person violates the station.
3. And judging whether the pipeline landform is deformed or not by using an LSD straight line judgment algorithm. The method comprises the steps of obtaining a linear pixel point set through local image analysis, verifying and solving through hypothesis parameters, combining the pixel point set with an error control set, further adaptively controlling the number of error detections, and performing linear detection by utilizing gradient information and a level-line (level-line) to judge whether a pipeline is installed normally or not and whether the landform physical characteristics are suitable for pipeline installation or not.
The non-conforming item identification algorithm applies a big data deep learning technology to improve the accuracy of the AI intelligent identification non-conforming item function. After various non-conforming items such as a safety helmet is not worn on a construction site, illegal hoisting operation is carried out, a person who leaves a crane and a pipe orifice is not blocked are identified in the video, relevant audio and video fragments are automatically intercepted and reserved, the occurrence time of the non-conforming items, the category of the non-conforming items and the audio and video fragments of the non-conforming items are filled into the corresponding positions of a non-conforming item notification sheet through a data interface of a non-conforming item intelligent identification system and a project management platform, and relevant responsible persons are notified to correct and modify in time; the method replaces the mode of issuing the paper non-conformity item by the supervision in the original side station process by using an informatization and automation means, reduces the workload of supervision personnel and the space and cost for storing paper data, reduces the labor cost, and enables the supervision work to be more objective and transparent.
Optionally, in one embodiment, automatically generating the non-compliant item notification includes: and when the non-conforming item is identified by using a non-conforming item identification algorithm, automatically intercepting and retaining the non-conforming item audio/video clips in the construction site image data, and filling the occurrence time of the non-conforming item, the category of the non-conforming item and the non-conforming item audio/video clips into the corresponding positions of the non-conforming item notice. In the embodiment, the mode of issuing the paper non-conforming item by supervision in the original side station process is replaced by an informatization and automation means, so that the supervision workload and the space and cost for storing the paper form are reduced, the labor cost is reduced, and the supervision work is more objective and transparent.
Optionally, in an embodiment, the object two-dimensional code corresponding to the construction code is scanned, the construction code is obtained, a construction code index is generated, and the construction quantity is counted according to the construction code index. The construction quantity statistics is carried out through the construction code indexes, and the constructed quantity, the un-constructed quantity and the like can be conveniently known.
In addition, an embodiment of the present invention further provides an intelligent side station supervision system, as shown in fig. 3, the system includes: the system comprises an intelligent terminal 310, a data acquisition instrument 320, an image acquisition device 330 and a project management platform 340.
The intelligent terminal 310 is used for inputting basic information of the inspection objects related to side station supervision, and generating and printing object two-dimensional codes of the inspection objects according to the basic information of the inspection objects; pasting the object two-dimensional code on a corresponding inspection object; the method comprises the steps of obtaining basic information of each inspection object through an object two-dimensional code of the inspection object related to a construction project, obtaining field acceptance image data, filling the basic information of the inspection object and the field acceptance image data into corresponding positions of a preset table, generating a related record table, and sending the related record table to a project management platform.
The data acquisition instrument 320 is used for acquiring construction process data in real time; the image acquisition device 330 is used for acquiring image data of a construction site in real time; and sending the construction process data and the construction site image data to a project management platform 340. Wherein, the data acquisition instrument can adopt acquisition devices such as a current sensor, a voltage sensor, a gas flow sensor, a temperature sensor, a humidity sensor and the like. The image acquisition device can adopt fixed cameras, handheld mobile recorders, unmanned aerial vehicles and other acquisition devices.
The project management platform 340 is configured to monitor the construction process according to the construction process data, identify the non-conforming item in the construction site image data by using a non-conforming item identification algorithm, and automatically generate a non-conforming item notification sheet.
The method comprises the steps that scanning keys or shooting keys are respectively arranged at different entry items of a preset table, the scanning function of the corresponding entry item is started through the scanning keys, basic information of an inspection object is obtained through scanning an object two-dimensional code of the inspection object, the obtained basic information of the inspection object is filled in the corresponding entry item, the shooting function of the corresponding entry item is started through the shooting keys, field acceptance image data is obtained, and the field acceptance image data is filled in the corresponding entry item.
Optionally, in one embodiment, a front-end server 350 is further included; the intelligent terminal 310, the data acquisition instrument 320 and the image acquisition device 330 are communicated with the front-end server 350 through a local WIFI network established on a construction site; the front server 350 is configured to cache the relevant record table, the construction process data, and the construction site image data, and send the relevant record table, the construction process data, and the construction site image data to the project management platform 340. By arranging the front server, the information acquired by the intelligent terminal, the data acquisition instrument and the image acquisition device is cached in the front server, the data is stored under the condition of no network, and when the network is recovered, the front server sends the cached data to the project management platform.
In the embodiment of the invention, when project acceptance is carried out on a two-dimensional code containing basic information of an inspection object, the basic information of various inspection objects related to a construction project, such as personnel, equipment, materials, construction codes and the like, can be conveniently obtained and objectively input, a recording form can be conveniently and automatically generated and sent to a project management platform, the functions of on-line circulation approval and the like are realized, the data of the construction process and the image data of the construction site are obtained in real time, and the remote monitoring on the construction process and the non-conforming items can be realized, so that the side station supervision personnel is assisted to improve the side station supervision quality, improve the working efficiency, reduce the labor cost and save the human resources.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (10)
1. An intelligent side station supervision method is characterized by comprising the following steps:
inputting basic information of inspection objects related to side station supervision, and generating and printing object two-dimensional codes of the inspection objects according to the basic information of the inspection objects; pasting the object two-dimensional code on a corresponding inspection object;
acquiring construction process data in real time through a data acquisition instrument, and monitoring the construction process according to the construction process data; acquiring construction site image data in real time through an image acquisition device, identifying non-conforming items in the construction site image data by using a non-conforming item identification algorithm, and automatically generating a non-conforming item notice;
when the project is accepted, the intelligent terminal scans the object two-dimensional codes of the inspection objects related to the construction project, acquires the basic information of each inspection object, acquires the field acceptance image data, fills the basic information of the inspection objects and the field acceptance image data into the corresponding positions of the preset table, and generates the related record table.
2. The intelligent side station supervision method according to claim 1, wherein a scan key or a shooting key is respectively arranged at different entry items of the preset table, a scan function of the corresponding entry item is opened through the scan key, basic information of an inspection object is obtained by scanning an object two-dimensional code of the inspection object, the obtained basic information of the inspection object is filled in the corresponding entry item, a shooting function of the corresponding entry item is opened through the shooting key, field acceptance image data is obtained, and the field acceptance image data is filled in the corresponding entry item.
3. The intelligent side station supervision method according to claim 1, wherein the identifying non-compliant items in the construction site image data using a non-compliant item identification algorithm comprises:
recognizing and judging the preset object by using ImageNet to assist an SVM algorithm;
the PCA is used for assisting GNN and YOLO algorithm to realize the identification of personnel and the detection of behaviors;
and judging whether the pipeline installation is standard or not by using an LSD straight line judgment algorithm.
4. The intelligent side station supervision method according to claim 3, wherein the recognizing and judging of the preset object by using ImageNet auxiliary SVM algorithm comprises:
similar samples of preset objects needing to be detected in a construction scene are screened out in ImageNet to form a training set, and the training set is used for training to generate a deep learning model;
sampling image data of a specific and standard-meeting preset object required by a construction scene during operation to generate a self-made data set;
converting the solution space mapping of the depth model full-connected layer into an SVM, reserving the trained parameter models of each layer, and performing transfer learning on the deep learning model by using the self-made data set to complete the integral training of the discrimination model;
and identifying the image data of the actual construction scene by using the discrimination model finished by the integral training, and judging that a non-conforming item exists when the identification result is smaller than a preset threshold value.
5. The intelligent wayside proctoring method as defined in claim 1, wherein the automatically generating a notice of non-compliant items comprises: and when the non-conforming item is identified by using a non-conforming item identification algorithm, automatically intercepting and retaining the non-conforming item audio/video clips in the construction site image data, and filling the occurrence time of the non-conforming item, the category of the non-conforming item and the non-conforming item audio/video clips into the corresponding positions of the non-conforming item notice.
6. The intelligent side station supervision method according to any one of claims 1 to 5, wherein the inspection objects include persons, equipment, materials and construction codes, and the construction codes refer to numbers representing construction sites.
7. The intelligent side station supervision method according to claim 6, wherein the object two-dimensional code corresponding to the construction code is scanned, the construction code is obtained, a construction code index is generated, and the construction quantity is counted according to the construction code index.
8. The utility model provides a side station of intelligence supervision system which characterized in that includes: the system comprises an intelligent terminal, a data acquisition instrument, an image acquisition device and a project management platform;
the intelligent terminal is used for inputting basic information of the inspection object related to side station supervision; acquiring basic information of each inspection object through an object two-dimensional code of the inspection object related to a construction project, acquiring field acceptance image data, filling the basic information of the inspection object and the field acceptance image data into corresponding positions of a preset table, generating a related record table, and sending the related record table to a project management platform;
the data acquisition instrument is used for acquiring construction process data in real time; the image acquisition device is used for acquiring image data of a construction site in real time; sending the construction process data and the construction site image data to a project management platform;
the project management platform is used for executing the intelligent side station supervision method of any one of claims 1 to 7.
9. The intelligent wayside proctoring system of claim 8, further comprising a front-end server; the intelligent terminal, the data acquisition instrument and the image acquisition device are communicated with the front-end server through a local WIFI network established on a construction site; the front-end server is used for caching the related record list, the construction process data and the construction site image data and sending the construction process data and the construction site image data to a project management platform.
10. The intelligent wayside station supervision system according to any one of claims 8 to 9, wherein the data acquisition instrument comprises at least one of a current sensor, a voltage sensor, a gas flow sensor, a temperature sensor, a humidity sensor; the image acquisition device comprises at least one of a fixed camera, a handheld mobile recorder and an unmanned aerial vehicle.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111031812.9A CN113850562A (en) | 2021-09-03 | 2021-09-03 | Intelligent side station supervision method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111031812.9A CN113850562A (en) | 2021-09-03 | 2021-09-03 | Intelligent side station supervision method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113850562A true CN113850562A (en) | 2021-12-28 |
Family
ID=78973103
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111031812.9A Pending CN113850562A (en) | 2021-09-03 | 2021-09-03 | Intelligent side station supervision method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113850562A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114677123A (en) * | 2022-04-19 | 2022-06-28 | 赵德群 | Management system for low-carbon green building supervision engineering project |
| CN115484437A (en) * | 2022-09-09 | 2022-12-16 | 广东重工建设监理有限公司 | Wearable side station recorder, wearable side station recording system and side station supervision method |
| CN116228166A (en) * | 2023-04-21 | 2023-06-06 | 广州智算信息技术有限公司 | Water heater installation acceptance system based on artificial intelligence |
| CN117112768A (en) * | 2023-10-23 | 2023-11-24 | 公诚管理咨询有限公司 | Intelligent question-answering method and system based on unsupervised semantic extraction |
-
2021
- 2021-09-03 CN CN202111031812.9A patent/CN113850562A/en active Pending
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114677123A (en) * | 2022-04-19 | 2022-06-28 | 赵德群 | Management system for low-carbon green building supervision engineering project |
| CN115484437A (en) * | 2022-09-09 | 2022-12-16 | 广东重工建设监理有限公司 | Wearable side station recorder, wearable side station recording system and side station supervision method |
| CN116228166A (en) * | 2023-04-21 | 2023-06-06 | 广州智算信息技术有限公司 | Water heater installation acceptance system based on artificial intelligence |
| CN116228166B (en) * | 2023-04-21 | 2023-08-04 | 广州智算信息技术有限公司 | Water heater installation acceptance system based on artificial intelligence |
| CN117112768A (en) * | 2023-10-23 | 2023-11-24 | 公诚管理咨询有限公司 | Intelligent question-answering method and system based on unsupervised semantic extraction |
| CN117112768B (en) * | 2023-10-23 | 2024-02-02 | 公诚管理咨询有限公司 | Intelligent question-answering method and system based on unsupervised semantic extraction |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113850562A (en) | Intelligent side station supervision method and system | |
| CN110826538B (en) | Abnormal off-duty identification system for electric power business hall | |
| CN106331636A (en) | Intelligent video monitoring system and method of oil pipelines based on behavioral event triggering | |
| CN110619277A (en) | Multi-community intelligent deployment and control method and system | |
| CN111563557A (en) | Method for detecting target in power cable tunnel | |
| CN104240349A (en) | Method for quickly confirming real name identity in important place and human image and identity comparison safety inspection system | |
| CN205644670U (en) | Metropolitan environment air quality monitoring's intelligence system of patrolling and examining | |
| CN112070135A (en) | Power equipment image detection method and device, power equipment and storage medium | |
| CN110458794B (en) | Quality detection method and device for accessories of rail train | |
| CN112434827A (en) | Safety protection identification unit in 5T fortune dimension | |
| CN112434828A (en) | Intelligent identification method for safety protection in 5T operation and maintenance | |
| CN115220479A (en) | Dynamic and static cooperative power transmission line refined inspection method and system | |
| CN112949457A (en) | Maintenance method, device and system based on augmented reality technology | |
| CN113128393A (en) | Method and system for detecting whether person is under large-scale mechanical boom | |
| CN111404874A (en) | Taxi suspect vehicle discrimination analysis system architecture | |
| CN116846059A (en) | Edge detection system for power grid inspection and monitoring | |
| CN113505704B (en) | Personnel safety detection method, system, equipment and storage medium for image recognition | |
| CN110197341A (en) | Social contradications dispute information management system and method | |
| CN202815869U (en) | Vehicle microcomputer image and video data extraction apparatus | |
| CN111179458A (en) | Method and device for routing inspection of rail transit equipment | |
| CN113392706A (en) | Device and method for detecting smoking and using mobile phone behaviors | |
| CN111047731A (en) | AR technology-based telecommunication room inspection method and system | |
| CN106204343A (en) | A kind of construction quality control system and method based on mobile terminal | |
| CN110826473A (en) | Neural network-based automatic insulator image identification method | |
| CN115457411A (en) | Unmanned inspection method and device for aviation oil pipeline and aviation oil safe operation and maintenance system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |