Intelligent construction site management system and method

Abstract

The invention relates to the technical field of intelligent construction site management, in particular to an intelligent construction site management system and method. Comprising an image transmission connection unit, a worker information extraction unit, a worker entry recording unit, a team priority improvement unit, a safety helmet identification unit and a worker priority improvement unit. The image transmission connection unit is used for establishing image transmission connection with a face recognition machine of a real-name system channel of a construction site and a peripheral camera for monitoring the real-name system channel, and is connected to a construction site personnel management end; combination distribution is performed according to team information in the worker information, so that each worker has a clear team attribution, team priority ranking is performed according to the distance between a pre-recognized worker and a face recognition machine when the workers of a plurality of teams appear at the same time, the team with the nearest distance has the highest priority, and meanwhile, the team with the nearest distance has the highest priority. In combination with quantity analysis of safety helmet colors, the priority of workers who wear the maximum number of safety helmet colors and have appropriate distances is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of smart construction site management, and in particular to a smart construction site management system and method. Background Art

[0002] In terms of construction site management, existing technologies are designed to improve construction efficiency, protect worker safety and ensure project quality. Through manual management combined with some information technology, such as attendance clock-in systems and safety inspection forms, a certain degree of monitoring of basic management of construction site personnel can be achieved.

[0003] At present, when workers enter and exit a construction site, they need to undergo facial recognition through a real-name channel. However, after a day's work at the construction site, due to the high dust on the construction site, the workers' faces may be obscured by impurities, resulting in the facial recognition device being unable to quickly authenticate the workers' identities when performing facial recognition exit detection. Workers are required to keep trying, and when the recognition efficiency is low, the real-name channel will be crowded, worker management will be chaotic, and workers will not queue up to scramble to pass through the facial recognition machine, resulting in incomplete records of workers leaving the site and inefficient management of workers. Therefore, a smart construction site management system and method are proposed. Summary of the invention

[0004] The purpose of the present invention is to provide a smart construction site management system and method to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, one of the purposes of the present invention is to provide a smart construction site management system, including an image transmission connection unit, a worker information extraction unit, a worker entry recording unit, a team priority improvement unit, a helmet recognition unit and a worker priority improvement unit;

[0006] The image transmission connection unit is used to establish an image transmission connection with a face recognition machine of a real-name channel of a construction site and a peripheral camera that monitors the real-name channel, and is connected to a construction site personnel management terminal at the same time;

[0007] The worker information extraction unit is used to extract worker information at the construction site personnel management end, group and distribute the worker information according to the team, and then establish a worker database, and at the same time establish a real-time record library of construction site workers;

[0008] The worker entry recording unit is used to retain the image data of workers who enter the construction site through the face recognition machine, and input the worker's team and helmet color into the construction site worker real-time record library for recording according to the worker image data;

[0009] The team priority improvement unit is used to pre-recognize the faces of workers leaving the site based on the image data obtained by the surrounding cameras, extract the teams from the real-time record library of workers on the construction site based on the pre-recognized faces, and input the extracted teams into the face recognition machine for priority improvement;

[0010] The helmet recognition unit is used to extract the helmet color from the image data acquired by the surrounding cameras, and compare the extracted helmet color with the helmet color of the team extracted by the team priority improvement unit. When the helmet color recorded by the team contains other helmet colors, the priority of the workers with other helmet colors is restored;

[0011] The worker priority enhancement unit is used to perform quantitative analysis on different helmet colors, extract the most numerous helmet color, and perform dynamic position recognition in image data acquired by surrounding cameras, and dynamically adjust the recognition priority of the helmet color in the team according to the recognition result.

[0012] As a further improvement of the present technical solution, the image transmission connection unit is connected to the construction site personnel management terminal, and then establishes an image transmission connection between the construction site personnel management terminal and the face recognition machine of the construction site real-name channel and the surrounding cameras that monitor the real-name channel through the Internet of Things, thereby being able to collect image data from the face recognition machine and image data from the surrounding cameras in real time.

[0013] As a further improvement of the technical solution, the worker information extraction unit includes a team allocation module and a database establishment module;

[0014] The team allocation module is used to extract worker information at the construction site personnel management end, and allocate workers to team combinations according to the team information in the worker information, so that each worker has a corresponding team;

[0015] The database establishment module is used to establish a worker database based on the worker information assigned in combination, and at the same time establish a real-time record library for construction site workers. The real-time record library for construction site workers is applied to the face recognition machine, and a corresponding number of labels are established in the real-time record library for construction site workers according to the number of teams.

[0016] As a further improvement of the technical solution, the worker entry recording unit includes an image retention module and a worker recording module;

[0017] The image retention module is used to retain the image data of the worker entering the construction site through the face recognition machine, and when the worker leaves the construction site through the face recognition machine, the data retained for the worker is deleted from the real-time record library of the construction site workers;

[0018] The worker recording module is used to group the retained worker image data, save the worker in the construction site worker real-time record library with a corresponding label according to the identified team, and then perform helmet color analysis on the worker image data to obtain the color of the helmet worn by the worker.

[0019] As a further improvement of the technical solution, the team priority enhancement unit includes a pre-identification module and a priority identification module;

[0020] The pre-identification module is used to extract image data acquired by surrounding cameras, perform face pre-identification on workers leaving the site based on the image data, obtain recognizable face information, and use it as the pre-identified worker;

[0021] The priority recognition module is used to extract the team according to the pre-recognized workers in the real-time record library of the construction site workers, feed the extracted team back to the real-time record library of the construction site workers, and then send it to the face recognition machine through the real-time record library of the construction site workers to improve the priority of recognition;

[0022] When multiple teams are extracted, they are sorted according to the distance between the pre-identified workers and the face recognition machine. When the pre-identified workers of a team are closest to the face recognition machine, the recognition priority of this team is the highest.

[0023] As a further improvement of the technical solution, the helmet recognition unit includes a helmet extraction module and a helmet comparison module;

[0024] The helmet extraction module is used to analyze and extract the color of the helmets worn by the workers according to the image data obtained by the surrounding cameras, so as to obtain the color of the helmets worn by each worker;

[0025] The hard hat comparison module is used to compare the acquired hard hat colors with those of the work group with increased recognition priority. When the hard hat colors recorded by the work group in the real-time record library of construction site workers contain other hard hat colors that are not obtained by the hard hat extraction module, the workers with other hard hat colors in the work group with increased recognition priority will be prioritized, and only workers with the same hard hat colors as those extracted from the analysis will be retained in the work group with increased recognition priority. Secondly, when the hard hat colors recorded by the work group in the real-time record library of construction site workers are the same as the number of hard hat colors obtained by the hard hat extraction module, if there is only one color, continue to monitor; if there are multiple colors, send them to the worker priority improvement unit for the next step.

[0026] As a further improvement of the technical solution, the worker priority enhancement unit includes a quantity analysis module and a safety helmet priority adjustment module;

[0027] The quantity analysis module is used to perform quantity analysis on different helmet colors, obtain the quantity of each helmet color, and then extract the helmet color with the largest quantity;

[0028] The safety helmet priority adjustment module is used to dynamically identify the positions of workers wearing the most numerous safety helmet colors in the image data acquired by the surrounding cameras. When the safety helmet color worn by the worker closest to the face recognition machine is inconsistent with the safety helmet color worn by the most numerous workers, the priority of the workers wearing the most numerous safety helmet colors will be lowered than the workers wearing other safety helmet colors in the team with increased recognition priority. Secondly, when the safety helmet color worn by the worker closest to the face recognition machine is consistent with the safety helmet color worn by the most numerous workers, the priority of the workers wearing the most numerous safety helmet colors will be increased compared with the workers wearing other safety helmet colors in the team with increased recognition priority.

[0029] A second object of the present invention is to provide a smart construction site management method, including any one of the smart construction site management systems described above, comprising the following steps:

[0030] S1. Establish an image transmission connection with the face recognition machine of the real-name channel of the construction site and the surrounding cameras that monitor the real-name channel, and connect to the construction site personnel management terminal at the same time. At the same time, the worker information is combined and allocated according to the team, and then a worker database is established, and a real-time record library of the construction site workers is established;

[0031] S2. The image data of workers who enter the construction site through the face recognition machine is retained, and the team and helmet color of the worker are input into the real-time record library of the construction site workers according to the image data of the worker for recording;

[0032] S3. Pre-recognize the faces of the workers leaving the site based on the image data obtained by the surrounding cameras, extract the teams from the real-time record library of the workers on the construction site based on the pre-recognized faces, input the extracted teams into the face recognition machine for priority improvement, compare the extracted helmet colors with the extracted teams, and when the helmet colors recorded by the teams contain other helmet colors, restore the workers with other helmet colors with priority;

[0033] S4. Quantitatively analyze the different colors of helmets, extract the color of the helmet with the largest number, and dynamically identify its position in the image data obtained by the surrounding cameras. Dynamically adjust the recognition priority of the helmet color in the team according to the recognition result.

[0034] Compared with the prior art, the present invention has the following beneficial effects:

[0035] The smart construction site management system and method combines and allocates workers according to the team information in the worker information, so that each worker has a clear team affiliation. When workers from multiple teams appear at the same time, the team priorities are sorted according to the distance between the pre-identified workers and the face recognition machine. The team closest to the worker has the highest priority. At the same time, combined with the quantitative analysis of the colors of safety helmets, the priority of workers wearing the largest number of safety helmets of the same color and at the right distance is increased. This dynamic adjustment method can better adapt to the actual situation of the construction site, improve the pertinence and effectiveness of management, and at the same time increase the speed of workers leaving and entering the site to avoid large-scale congestion. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is the overall structural principle diagram of the present invention.

[0037] The meaning of each number in the figure is:

[0038] 10. Image transmission connection unit; 20. Worker information extraction unit; 30. Worker entry recording unit; 40. Team priority improvement unit; 50. Safety helmet recognition unit; 60. Worker priority improvement unit. DETAILED DESCRIPTION

[0039] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] like Figure 1 As shown, one of the purposes of the present invention is to provide a smart construction site management system, including an image transmission connection unit 10, a worker information extraction unit 20, a worker entry recording unit 30, a team priority improvement unit 40, a helmet recognition unit 50 and a worker priority improvement unit 60;

[0041] The image transmission connection unit 10 is used to establish an image transmission connection with the face recognition machine of the real-name channel of the construction site and the surrounding cameras that monitor the real-name channel, and at the same time connect to the construction site personnel management terminal;

[0042] The image transmission connection unit 10 is connected to the construction site personnel management terminal, and then establishes an image transmission connection between the construction site personnel management terminal and the face recognition machine of the construction site real-name channel and the surrounding cameras monitored by the real-name channel through the Internet of Things, so that the image data of the face recognition machine and the image data of the surrounding cameras can be collected in real time. The specific steps are as follows:

[0043] Access to the construction site personnel management terminal: determine the interface type and communication protocol of the construction site personnel management terminal so that it can be connected to other devices later;

[0044] Connect the site personnel management terminal to the equipment through the Internet of Things: select appropriate Internet of Things communication technology, such as Wi-Fi, Bluetooth, ZigBee or cellular network, etc., consider the actual environment and needs of the construction site, ensure the stability and reliability of communication, and then configure the network parameters of the Internet of Things device, including IP address, subnet mask, gateway, etc., so that a data transmission channel can be established between the site personnel management terminal and the face recognition machine and surrounding cameras

[0045] Real-time image data collection: The face recognition machine and surrounding cameras continuously collect image data and convert it into digital signals. Through the established image transmission connection, the image data is transmitted in real time to the construction site personnel management end.

[0046] The worker information extraction unit 20 is used to extract worker information at the construction site personnel management end, group and distribute the worker information according to the team, and then establish a worker database, and at the same time establish a real-time record library of construction site workers;

[0047] The worker information extraction unit 20 includes a team allocation module and a database establishment module;

[0048] The team allocation module is used to extract worker information on the construction site personnel management end, and allocate workers to team combinations according to the team information in the worker information, so that each worker has a corresponding team;

[0049] The database establishment module is used to establish a worker database based on the information of the workers assigned by the combination, and to establish a real-time record library for construction site workers. The real-time record library for construction site workers is applied to the face recognition machine. The corresponding number of labels is established in the real-time record library for construction site workers according to the number of teams. The specific steps are as follows:

[0050] Extract worker information: connect to the database or data source of the construction site personnel management end, obtain the worker information stored therein, perform data cleaning and preprocessing on the obtained worker information, remove invalid data and erroneous information, and ensure the accuracy and completeness of the data;

[0051] Combine and allocate workers according to team information: Extract team information from worker information, group workers according to team information, and ensure that each worker is assigned to a specific team;

[0052] Establish a worker database: Select a suitable database management system, such as MySQL, Oracle or MongoDB, etc., and design the table structure of the database according to the needs of construction site management, including the worker information table, team information table, etc., and store the combined and allocated worker information in the worker database;

[0053] Create tags in the real-time record library: According to the number of teams, create corresponding tags for each team in the real-time record library of construction site workers, and associate the worker information with the corresponding team tag to ensure that the worker information of a specific team can be quickly queried in the real-time record library.

[0054] The worker entry recording unit 30 is used to retain the image data of the worker who enters the construction site through the face recognition machine, and input the worker's team and helmet color into the construction site worker real-time record library for recording according to the worker image data;

[0055] The worker entry recording unit 30 includes an image retention module and a worker recording module;

[0056] The image retention module is used to retain the image data of workers entering the construction site through the face recognition machine. At the same time, when the worker leaves the construction site through the face recognition machine, the data retained for the worker will be deleted from the real-time record library of the construction site workers. The specific steps are as follows:

[0057] Data reception and recognition: When the face recognition machine detects a worker performing a recognition operation, it will collect the worker's facial image data in real time and convert it into a digital format. This data is sent to the server side where the real-time record library of the construction site workers is located through the previously established transmission channel;

[0058] The facial recognition machine will process the collected images using a facial recognition algorithm to determine whether the worker matches the registered personnel information. If the match is successful, it will be confirmed that the worker is a legal person who can enter the construction site and prepare for subsequent data retention operations;

[0059] Storage preparation and integration of related information: At the construction site worker real-time record library, determine the storage location and storage method of the image data according to the storage rules set by the system, extract other related information corresponding to the worker, such as name, team, entry time, etc., and associate this information with the image data for subsequent query and management. For example, a data structure can be established, which contains image data fields and other text information fields, and the corresponding content can be filled in.

[0060] Departure detection and confirmation: When the worker comes to the face recognition machine again to perform the recognition operation and prepare to leave the construction site, the face recognition machine also collects his facial image and processes it through the recognition algorithm to confirm the worker's identity and his departure behavior. By comparing the entry time and the current time, it can be comprehensively judged that the worker is indeed leaving the construction site, thereby triggering the subsequent data deletion process;

[0061] Data location and deletion operations: Based on the association information and index mechanism established during previous storage, the image data record retained when the worker previously entered is quickly located in the real-time record library of the construction site workers, and the deletion operation is performed.

[0062] The worker record module is used to group the retained worker image data, save the worker in the construction site worker real-time record library with the corresponding label according to the identified team, and then perform helmet color analysis on the worker image data to obtain the color of the helmet worn by the worker. The specific steps are as follows:

[0063] Image feature extraction: Select images that need to be identified by the team from the worker image data stored in the real-time record library of workers on the construction site, use computer vision-related technologies to extract key features from the images, and normalize the extracted features to make them within a suitable numerical range for subsequent comparison and analysis;

[0064] Compare with known team member features: compare the currently extracted worker image features with the features in the feature library of each team one by one, and determine which team the worker belongs to based on the set similarity threshold. Once the team the worker belongs to is determined, obtain the label corresponding to the team;

[0065] Image color space conversion: Select the image part of the helmet color to be analyzed from the retained worker image data, and convert the image color space from the common RGB (red, blue, yellow) color space to a color space that is more convenient for color analysis, such as the HSV (hue, saturation, value) color space;

[0066] Color range definition and recognition: Set the corresponding threshold interval in the HSV color space according to the common color range of hard hats, traverse the pixels in the hard hat area in the image after the color space conversion, and determine whether the H, S, and V values ​​of each pixel fall within the set threshold interval of a certain color. If the number of pixels falling within a certain color interval reaches a certain proportion (for example, more than 50% of the hard hat recognition units of the total number of pixels in the entire hard hat area), the hard hat is identified as this color, and then the identified hard hat color information is saved in the corresponding record of the worker in the real-time record library of construction site workers.

[0067] The team priority improvement unit 40 is used to pre-recognize the faces of the workers leaving the site according to the image data obtained by the surrounding cameras, extract the teams from the real-time record library of the workers on the construction site according to the pre-recognized faces, and input the extracted teams into the face recognition machine for priority improvement;

[0068] The team priority enhancement unit 40 includes a pre-identification module and a priority identification module;

[0069] The pre-identification module is used to extract image data obtained by surrounding cameras, perform face pre-identification on workers leaving the site based on the image data, obtain recognizable face information, and use it as the pre-identified worker. The specific steps are as follows:

[0070] Data acquisition: extracting image data from the storage location or real-time transmission channel established by surrounding cameras;

[0071] Face detection: Using Haar feature-based cascade classifiers and deep learning-based face detection algorithms (such as SSD, MTCNN, etc.), the pre-processed image data is input into the selected face detection algorithm. The algorithm scans various parts of the image and determines which areas may be faces based on the pre-learned facial feature patterns. It then outputs the coordinate information of the area where the face is located and the possible face confidence score.

[0072] Facial feature extraction: For the detected face area, the corresponding facial feature extraction technology is used. Commonly used methods include convolutional neural network (CNN) based methods, such as using network models that have been trained on large-scale face datasets (such as VGG-Face, ArcFace, etc.). These models can convert face images into feature vectors of fixed dimensions. This vector can effectively represent the unique features of the face, such as the relative position of the facial features, facial contour shape, texture and other information;

[0073] Face matching and pre-recognition: Each face feature vector currently extracted from the surrounding camera images is compared with the vectors in the worker database one by one. The formula is as follows:

[0074]

[0075] in, is the current face feature vector, is the face feature vector in the comparison library, n is the feature vector dimension, a i and b i is a vector and The elements in the i-th dimension, For vector and The Euclidean distance between

[0076] The distance threshold is set to 0.8. When the calculated distance is less than the threshold, the two are considered to match, that is, the corresponding worker identity is identified, the relevant facial information of the worker is obtained, and the worker is marked as a pre-identified worker.

[0077] The priority recognition module is used to extract the team according to the pre-recognized workers in the real-time record library of the workers on the construction site, feed the extracted team back to the real-time record library of the workers on the construction site, and then send it to the face recognition machine through the real-time record library of the workers on the construction site to improve the recognition priority;

[0078] When multiple teams are extracted, they are sorted according to the distance between the pre-identified workers and the face recognition machine. When the pre-identified workers of a team are closest to the face recognition machine, the recognition priority of this team is the highest. The specific steps are as follows:

[0079] Locate the pre-identified worker record: obtain the pre-identified worker's unique identifier from his / her facial information, and then search the real-time record library of construction workers based on this unique identifier to locate the record entry corresponding to the pre-identified worker, and extract the team information to which the pre-identified worker belongs from the located record entry;

[0080] Send information to the face recognition machine to increase recognition priority: Encapsulate the extracted and updated team information, and attach the instruction information for increasing the recognition priority of the team. Send it to the face recognition machine through the established communication connection in accordance with the established data transmission format (such as JSON format), so that the face recognition machine knows that the recognition of the team needs to be given a higher priority;

[0081] Obtaining the distance information of pre-identified workers: When multiple teams are extracted, it is necessary to obtain the distance data between the pre-identified workers of each team and the face recognition machine. The formula is as follows:

[0082]

[0083] Among them, Z represents the actual distance, that is, the actual spatial distance between the worker and the face recognition machine, f is the focal length of the camera, which is an inherent parameter of the camera. It is usually marked when the camera leaves the factory or can be obtained through calibration experiments, D is the size of the object in the image, and d is the actual size of the object in the real world. For example, if the average height of the workers is known to be 1.7 meters, the distance can be estimated by substituting it into this formula after converting the pixel size in the image;

[0084] Distance sorting operation: Sort the distance data of the pre-identified workers of each team. After the sorting is completed, the team where the pre-identified worker corresponding to the team with the closest distance is located is the team with the highest recognition priority. The team information is sent to the face recognition machine again through the above-mentioned method to clearly inform the face recognition machine that this team should be given the highest recognition priority.

[0085] The helmet recognition unit 50 is used to extract the helmet color from the image data acquired by the surrounding cameras, and compare the extracted helmet color with the helmet color of the team extracted by the team priority improvement unit 40. When the helmet color recorded by the team contains other helmet colors, the priority of the workers with other helmet colors is restored;

[0086] The helmet recognition unit 50 includes a helmet extraction module and a helmet comparison module;

[0087] The helmet extraction module is used to analyze and extract the color of the helmets worn by workers based on the image data obtained by the surrounding cameras, so as to obtain the color of the helmets worn by each worker;

[0088] The image area containing the worker and the helmet is selected from the image data obtained by the surrounding cameras, and the color characteristics of the pixels in the helmet area are analyzed.

[0089] The hard hat comparison module is used to compare the acquired hard hat colors with those of the work team with increased recognition priority. When the hard hat colors recorded by the work team in the real-time record library of construction site workers contain other hard hat colors not obtained by the hard hat extraction module, the workers with other hard hat colors in the work team with increased recognition priority will be prioritized, and only workers with the same hard hat colors as those extracted from the analysis will be retained in the work team with increased recognition priority. Secondly, when the hard hat colors recorded by the work team in the real-time record library of construction site workers are the same as the number of hard hat colors obtained by the hard hat extraction module, if there is only one color, continue monitoring; if there are multiple colors, send them to the worker priority improvement unit 60 for the next step.

[0090] The worker priority enhancement unit 60 is used to perform quantitative analysis on different helmet colors, extract the helmet color with the largest number, and perform dynamic position recognition in the image data obtained by the surrounding cameras, and dynamically adjust the recognition priority of the helmet color in the team according to the recognition result.

[0091] The worker priority enhancement unit 60 includes a quantity analysis module and a helmet priority adjustment module;

[0092] The quantity analysis module is used to perform quantity analysis on different helmet colors, obtain the quantity of each helmet color, and then extract the helmet color with the largest quantity;

[0093] The safety helmet priority adjustment module is used to dynamically identify the positions of workers wearing the most numerous safety helmet colors in the image data acquired by the surrounding cameras. When the safety helmet color of the worker closest to the face recognition machine is inconsistent with the safety helmet color with the largest number, the priority of the workers wearing the most numerous safety helmet colors will be lowered than the workers wearing other safety helmet colors in the team with increased recognition priority. Secondly, when the safety helmet color of the worker closest to the face recognition machine is consistent with the safety helmet color with the largest number, the priority of the workers wearing the most numerous safety helmet colors will be increased compared with the workers wearing other safety helmet colors in the team with increased recognition priority.

[0094] In the image data obtained from the surrounding cameras, the image area where the workers wearing the most helmet colors are located is located again through target detection and other algorithms. Then, based on the camera calibration parameters and some geometric relationship calculations, the position information in the image is converted into the position coordinates in the actual space relative to the face recognition machine. The formula is as follows:

[0095]

[0096] Among them, (X, Y, Z) are world coordinates, that is, the position coordinates of the worker in the actual space, (u, v) are the pixel coordinates corresponding to the object on the image plane, K is the intrinsic parameter matrix of the camera, which contains parameters related to the imaging characteristics of the camera itself, such as focal length, and is generally obtained through calibration. R is the rotation matrix, and T is the translation vector, which together constitute the external parameters of the camera, describing the position and posture of the camera in the world coordinate system, and also need to be obtained through calibration. Z is the depth information, which is often determined with the help of some prior knowledge or other auxiliary measurement methods to complete the conversion from image coordinates to world coordinates and realize the determination of the worker's position in the actual space.

[0097] A second object of the present invention is to provide a smart construction site management method, including any one of the smart construction site management systems described above, comprising the following steps:

[0098] S1. Establish an image transmission connection with the face recognition machine of the real-name channel of the construction site and the surrounding cameras that monitor the real-name channel, and connect to the construction site personnel management terminal at the same time. At the same time, the worker information is combined and allocated according to the team, and then a worker database is established, and a real-time record library of the construction site workers is established;

[0099] S2. The image data of workers who enter the construction site through the face recognition machine is retained, and the team and helmet color of the worker are input into the real-time record library of the construction site workers according to the image data of the worker for recording;

[0100] S3. Pre-recognize the faces of the workers leaving the site based on the image data obtained by the surrounding cameras, extract the teams from the real-time record library of the workers on the construction site based on the pre-recognized faces, input the extracted teams into the face recognition machine for priority improvement, compare the extracted helmet colors with the extracted teams, and when the helmet colors recorded by the teams contain other helmet colors, restore the workers with other helmet colors with priority;

[0101] S4. Quantitatively analyze the different colors of helmets, extract the color of the helmet with the largest number, and dynamically identify its position in the image data obtained by the surrounding cameras. Dynamically adjust the recognition priority of the helmet color in the team according to the recognition result.

[0102] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. Smart construction site management system, characterized by: It comprises an image transmission connection unit (10), a worker information extraction unit (20), a worker entry recording unit (30), a team priority improvement unit (40), a safety helmet recognition unit (50) and a worker priority improvement unit (60); The image transmission connection unit (10) is used to establish an image transmission connection with a face recognition machine of a real-name channel of a construction site and a peripheral camera that monitors the real-name channel, and is connected to a construction site personnel management terminal at the same time; The worker information extraction unit (20) is used to extract worker information at the construction site personnel management end, group and distribute the worker information according to the work groups, and then establish a worker database, and at the same time establish a real-time record library of construction site workers; The worker entry recording unit (30) is used to retain the image data of workers who enter the construction site through the face recognition machine, and input the worker's team and helmet color into the construction site worker real-time record library for recording according to the worker image data; The team priority improvement unit (40) is used to pre-recognize the faces of workers leaving the site based on image data acquired by surrounding cameras, extract teams from the real-time record library of workers on the construction site based on the pre-recognized faces, and input the extracted teams into the face recognition machine for priority improvement; The helmet recognition unit (50) is used to extract the helmet color from the image data acquired by the peripheral cameras, and compare the extracted helmet color with the helmet color of the team extracted by the team priority improvement unit (40); when the helmet color recorded by the team contains other helmet colors, the priority of the workers with other helmet colors is restored; The worker priority enhancement unit (60) is used to perform quantitative analysis on different helmet colors, extract the helmet color with the largest number, and perform dynamic position recognition in image data acquired by surrounding cameras, and dynamically adjust the recognition priority of the helmet color in the team according to the recognition result.

2. The intelligent construction site management system according to claim 1, characterized in that: The image transmission connection unit (10) is connected to the construction site personnel management terminal, and then establishes an image transmission connection between the construction site personnel management terminal and the face recognition machine of the construction site real-name channel and the surrounding cameras that monitor the real-name channel through the Internet of Things, so that image data of the face recognition machine and image data of the surrounding cameras can be collected in real time.

3. The intelligent construction site management system according to claim 1, characterized in that: The worker information extraction unit (20) includes a team allocation module and a database establishment module; The team allocation module is used to extract worker information at the construction site personnel management end, and allocate workers to team combinations according to the team information in the worker information, so that each worker has a corresponding team; The database establishment module is used to establish a worker database based on the worker information assigned in combination, and at the same time establish a real-time record library for construction site workers. The real-time record library for construction site workers is applied to the face recognition machine, and a corresponding number of labels are established in the real-time record library for construction site workers according to the number of teams.

4. The intelligent construction site management system according to claim 1, characterized in that: The worker entry recording unit (30) comprises an image retention module and a worker recording module; The image retention module is used to retain the image data of the worker entering the construction site through the face recognition machine, and when the worker leaves the construction site through the face recognition machine, the data retained for the worker is deleted from the real-time record library of the construction site workers; The worker record module is used to classify the retained worker image data into work groups, save the worker in the construction site worker real-time record library with corresponding labels according to the identified work groups, and then perform helmet color analysis on the worker image data to obtain the color of the helmet worn by the worker.

5. The intelligent construction site management system according to claim 1, characterized in that: The team priority enhancement unit (40) includes a pre-identification module and a priority identification module; The pre-identification module is used to extract image data acquired by surrounding cameras, perform face pre-identification on workers leaving the site based on the image data, obtain recognizable face information, and use it as the pre-identified worker; The priority recognition module is used to extract the team according to the pre-recognized workers in the real-time record library of the construction site workers, feed the extracted team back to the real-time record library of the construction site workers, and then send it to the face recognition machine through the real-time record library of the construction site workers to improve the priority of recognition; When multiple teams are extracted, they are sorted according to the distance between the pre-identified workers and the face recognition machine. When the pre-identified workers of a team are closest to the face recognition machine, the recognition priority of this team is the highest.

6. The intelligent construction site management system according to claim 1, characterized in that: The helmet recognition unit (50) comprises a helmet extraction module and a helmet comparison module; The helmet extraction module is used to analyze and extract the color of the helmets worn by the workers according to the image data obtained by the surrounding cameras, so as to obtain the color of the helmets worn by each worker; The hard hat comparison module is used to compare the hard hat colors obtained with those of the work group with increased recognition priority. When the hard hat colors recorded by the work group in the real-time record library of construction site workers contain other hard hat colors that are not obtained by the hard hat extraction module, the workers with other hard hat colors in the work group with increased recognition priority are prioritized, and only workers with the same hard hat colors extracted from the analysis are retained in the work group with increased recognition priority. Secondly, when the hard hat colors recorded by the work group in the real-time record library of construction site workers are the same as the number of hard hat colors obtained by the hard hat extraction module, if there is only one color, then continue to monitor; if there are multiple colors, send them to the worker priority improvement unit (60) for the next step.

7. The intelligent construction site management system according to claim 1, characterized in that: The worker priority enhancement unit (60) includes a quantity analysis module and a safety helmet priority adjustment module; The quantity analysis module is used to perform quantity analysis on different helmet colors, obtain the quantity of each helmet color, and then extract the helmet color with the largest quantity; The safety helmet priority adjustment module is used to dynamically identify the positions of workers wearing the most numerous safety helmet colors in the image data acquired by the surrounding cameras. When the safety helmet color worn by the worker closest to the face recognition machine is inconsistent with the safety helmet color worn by the most numerous workers, the priority of the workers wearing the most numerous safety helmet colors will be lowered than the workers wearing other safety helmet colors in the team with increased recognition priority. Secondly, when the safety helmet color worn by the worker closest to the face recognition machine is consistent with the safety helmet color worn by the most numerous workers, the priority of the workers wearing the most numerous safety helmet colors will be increased compared with the workers wearing other safety helmet colors in the team with increased recognition priority.

8. A method for implementing a smart construction site management method, comprising the smart construction site management system according to any one of claims 1 to 7, characterized in that: The steps include: S1. Establish an image transmission connection with the face recognition machine of the real-name channel of the construction site and the surrounding cameras that monitor the real-name channel, and connect to the construction site personnel management terminal at the same time. At the same time, the worker information is combined and allocated according to the team, and then a worker database is established, and a real-time record library of the construction site workers is established; S2. The image data of workers who enter the construction site through the face recognition machine is retained, and the team and helmet color of the worker are input into the real-time record library of the construction site workers according to the image data of the worker for recording; S3. Pre-recognize the faces of the workers leaving the site based on the image data obtained by the surrounding cameras, extract the teams from the real-time record library of the workers on the construction site based on the pre-recognized faces, input the extracted teams into the face recognition machine for priority improvement, compare the extracted helmet colors with the extracted teams, and when the helmet colors recorded by the teams contain other helmet colors, restore the workers with other helmet colors with priority; S4. Quantitatively analyze the different colors of helmets, extract the color of the helmet with the largest number, and dynamically identify its position in the image data obtained by the surrounding cameras. Dynamically adjust the recognition priority of the helmet color in the team according to the recognition result.

Images