Intelligent safety supervision method and system for construction site

Abstract

The application provides an intelligent safety supervision method and system for a construction site. The steps of the method include: obtaining a feature vector of a worker based on a preset monitoring device, comparing each feature vector in a feature vector set corresponding to each work group with the feature vector of the worker, determining the work group to which the worker belongs if the comparison is successful, and recording the work trajectory of the worker based on the preset monitoring device; if the comparison fails, determining the position of the worker at each time point based on the preset monitoring device, and constructing a to-be-judged space-time curve; comparing the to-be-judged space-time curve with each standard space-time curve to obtain a single comparison vector corresponding to each standard space-time curve; assembling each single comparison vector into a combined comparison vector, inputting the combined comparison vector into a pre-trained neural network model, and outputting the work group to which the worker belongs from the pre-trained neural network model.

Description

Technical Field

[0001] This invention relates to the field of construction site management technology, and in particular to an intelligent safety monitoring method and system for construction sites. Background Technology

[0002] Construction site personnel management plays a crucial role in building projects, and its importance is self-evident. Firstly, effective construction site personnel management directly impacts the safety and quality of the project. Through rigorous job training and qualification verification, ensuring that every worker possesses the necessary professional skills and safety awareness significantly reduces the risk of accidents during construction, guarantees personnel safety, improves project quality, and avoids rework and delays. Secondly, reasonable personnel allocation and management can improve construction efficiency. Scientifically allocating human resources according to project progress and job requirements ensures seamless coordination between different processes, reducing waiting time and resource waste. This not only shortens the construction period but also effectively controls costs, enhancing the project's economic benefits.

[0003] In conclusion, construction site personnel management not only concerns construction safety and quality, but also directly affects construction efficiency, cost control, and team stability and harmony. However, existing construction site personnel management methods often rely on management personnel, making effective management difficult during large-scale operations.

[0004] In view of this, the present invention is proposed. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent safety supervision method and system for construction sites. This solution can supervise workers in different work groups separately to ensure targeted supervision. On the other hand, for newly joined workers, this solution can automatically determine the work group to which the worker belongs through a series of judgment schemes.

[0006] This invention provides an intelligent safety monitoring method for construction sites, the method comprising the following steps:

[0007] The feature vectors of the staff are obtained based on the preset monitoring equipment. The feature vectors of the staff are compared with the feature vectors of each work group in the current work group. If the comparison is successful, the work group to which the staff belongs is determined. The work trajectory of the staff is recorded based on the preset monitoring equipment.

[0008] If the comparison fails, the location of the staff member at each time point is determined based on the preset monitoring equipment, and a spatiotemporal curve to be judged is constructed.

[0009] Obtain the standard spatiotemporal curves of each working group, compare the spatiotemporal curve to be judged with each standard spatiotemporal curve, and obtain a single comparison vector corresponding to each standard spatiotemporal curve.

[0010] The individual alignment vectors are combined into a combined alignment vector, which is then input into a pre-trained neural network model. The pre-trained neural network model outputs the work group to which the worker belongs.

[0011] The feature vector of the employee is entered into the feature vector set for that job.

[0012] By adopting the above scheme, this scheme can extract the feature vector of each staff member through monitoring equipment, and determine the work group to which the staff member belongs through the feature vector. Since the job types of each work group may be different, this scheme can monitor them separately to ensure the targeting of the supervision. On the other hand, for newly joined staff members, this scheme can determine the work group to which the staff member belongs through a series of judgment schemes and classify the staff member's feature vector. The grouping of staff members can be completed automatically without human intervention, thereby improving the degree of automation of supervision. This scheme can ensure the effectiveness and efficiency of supervision.

[0013] In some embodiments of the present invention, the spatiotemporal curve to be determined and the standard spatiotemporal curve are both established in the same Cartesian coordinate system. Each position of the horizontal axis of the Cartesian coordinate system represents a time point, and each position of the vertical axis represents a location. In the step of determining the location of the worker at each time point based on the preset monitoring equipment and constructing the spatiotemporal curve to be determined, based on the location of the worker at each time point, the corresponding points are marked in the Cartesian coordinate system, and the marked points are connected in chronological order to obtain the spatiotemporal curve to be determined.

[0014] By adopting the above scheme, the work trajectory of the staff can be represented in a Cartesian coordinate system, which can efficiently represent the work trajectory of the staff. Moreover, by representing it in the same Cartesian coordinate system, it is convenient to compare the work trajectory of the staff with the standards of each work group.

[0015] In some embodiments of the present invention, before the step of obtaining the standard spatiotemporal curves of each working group, the standard spatiotemporal curves of each working group are calculated. In the step of calculating the standard spatiotemporal curves of each working group, the positions of the staff in the same working group at each time point are obtained, and the corresponding points are marked in a Cartesian coordinate system. The marked points are connected based on the time sequence to construct the standard spatiotemporal curve of the working group.

[0016] In some embodiments of the present invention, in the step of obtaining the positions of staff members in the same work group at various time points and marking the corresponding points in a Cartesian coordinate system, if there are multiple staff members in different positions at the same time point, the marked point is the point corresponding to the position where the most staff members are located at that time point.

[0017] The above approach is problematic because the work of working groups is often adjusted according to actual needs, and each working group does not have fixed tasks. Therefore, it is difficult to use a standard spatiotemporal curve for working groups in a long-term manner. This approach generates a corresponding standard spatiotemporal curve based on the actual situation, ensuring the accurate representation of the standard spatiotemporal curve for each working group and further guaranteeing the accuracy of subsequent working group matching.

[0018] In some embodiments of the present invention, in the step of comparing the feature vector of the staff member with each feature vector in the feature vector set corresponding to each current work group, the similarity between the feature vector of the staff member and each feature vector in the feature vector set corresponding to each current work group is calculated. If any similarity value is greater than a preset comparison threshold, the comparison is successful; otherwise, the comparison fails.

[0019] In some embodiments of the present invention, in the step of calculating the similarity between the feature vector of a worker and each feature vector in the feature vector set corresponding to each current work group, the similarity is calculated based on the following formula:

[0020] similar=(1+α T β) / 2;

[0021] Where similar represents similarity, α T Let α represent the transpose of any eigenvector in the eigenvector set, β represent the eigenvector of the worker to be determined, and α represent the eigenvector of the worker to be determined. T β represents the inner product of the feature vector of the worker to be judged and the transpose of any feature vector in the feature vector set.

[0022] In some embodiments of the present invention, in the step of comparing the spatiotemporal curve to be judged with each standard spatiotemporal curve to obtain a single comparison vector corresponding to each standard spatiotemporal curve, points in the standard spatiotemporal curve at the same time point and points in the spatiotemporal curve to be judged are compared respectively. If they coincide, a first parameter is generated; if they are different, a second parameter is generated. The first parameter and the second parameter are combined based on the order of the time points to obtain a single comparison vector corresponding to each standard spatiotemporal curve.

[0023] By adopting the above scheme, this scheme can generate a single alignment vector for each working group, ensuring the accuracy of the comparison.

[0024] In some embodiments of the present invention, in the steps of assembling individual alignment vectors into a combined alignment vector, inputting the combined alignment vector into a pre-trained neural network model, and the pre-trained neural network model outputting the work group to which the worker belongs, multiple individual alignment vectors are sequentially connected to obtain a combined alignment vector, the combined alignment vector is input into the pre-trained neural network model, and the classification layer of the neural network model is used for classification to determine the work group to which the worker belongs.

[0025] By adopting the above scheme, this scheme can combine the comparison results with multiple working groups and make the final judgment through a pre-trained neural network model. Since each staff member has a certain degree of freedom in the actual work process, it can avoid the deviation of a single staff member from the standard working group and make a combined judgment.

[0026] In some embodiments of the present invention, in the step of obtaining the feature vector of the staff based on the preset monitoring device, the monitoring image of the staff is obtained based on the preset monitoring device, and the ReID feature vector of the staff is identified based on the monitoring image as the feature vector of the staff.

[0027] Another aspect of the present invention relates to an inspection robot inspection system, the system including a computer device, the computer device including a processor and a memory, the memory storing computer instructions, the processor being used to execute the computer instructions stored in the memory, and when the computer instructions are executed by the processor, the system implements the steps of the method.

[0028] In summary, the present invention has the following beneficial effects:

[0029] 1. This solution can extract the feature vector of each staff member through monitoring equipment, and determine the work group to which the staff member belongs through the feature vector. Since the job types of each work group may be different, this solution can monitor them separately to ensure the targeted nature of the monitoring.

[0030] 2. This solution can determine the work group to which a newly added staff member belongs through a series of judgment methods and classify the staff member's feature vector. The grouping of staff members can be completed automatically without human intervention, thereby improving the degree of automation of supervision. This solution can ensure the effectiveness and efficiency of supervision.

[0031] 3. Since the work of the working groups is often adjusted according to actual work needs, and each working group does not have fixed work, it is difficult to use the standard spatiotemporal curve of the working group in a long-term manner according to a single standard. This solution generates the corresponding standard spatiotemporal curve according to the actual situation, ensuring the accurate representation of the standard spatiotemporal curve of each working group, and further ensuring the accuracy when matching working groups in the future.

[0032] 4. This solution can combine the comparison results with multiple working groups and make the final judgment through a pre-trained neural network model. Since each staff member has a certain degree of freedom in the actual work process, it can avoid the deviation of a single staff member from the standard working group and make a combined judgment. Attached Figure Description

[0033] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0034] Figure 1 This is a schematic diagram of one embodiment of the intelligent safety monitoring method for construction sites according to the present invention;

[0035] Figure 2 This is a schematic diagram of another embodiment of the intelligent safety monitoring method for construction sites according to the present invention. Detailed Implementation

[0036] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of apparatuses and methods consistent with some aspects of the invention as detailed in the appended claims.

[0037] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The singular forms “a,” “the,” and “the” used in this invention and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0038] like Figure 1 As shown, the present invention provides an intelligent safety monitoring method for construction sites, the method comprising the following steps:

[0039] Step s100: Obtain the feature vector of the staff based on the preset monitoring equipment, and compare the feature vector of the staff with each feature vector in the feature vector set corresponding to each work group.

[0040] In practical implementation, the ReID feature vector of the staff member is identified based on the surveillance image and used as the staff member's feature vector. Specifically, ReID (Person Re-identification) is a computer vision technology aimed at identifying the same pedestrian from different camera perspectives. Essentially, it uses algorithms to find targets in an image database and extracts and matches features based on the pedestrian's clothing, accessories, and posture; it falls under the category of image retrieval problems. ReID technology is widely used in video surveillance, intelligent transportation, facial recognition, and other fields, effectively improving the intelligence level of surveillance systems and enhancing public safety and traffic efficiency.

[0041] Specifically, based on monitoring images from multiple monitoring devices, the ReID feature vector of each person in each monitoring image is determined, and the ReID feature vectors are clustered into a set as the feature vector of that person; the ReID feature vector can be calculated using a deep learning model.

[0042] In the specific implementation process, when clustering the ReID feature vectors into a set, traditional clustering algorithms can be used, or secondary feature optimization of the neural network model can be combined with traditional clustering algorithms for clustering.

[0043] Step S110: If the comparison is successful, the work group to which the staff member belongs is determined, and the staff member's work trajectory is recorded based on the preset monitoring equipment.

[0044] In the specific implementation process, the similarity between the feature vector of the staff member and each feature vector in the feature vector set corresponding to each work group is calculated. If any similarity value is greater than the preset comparison threshold, the comparison is successful; otherwise, the comparison fails.

[0045] In the specific implementation process, if the comparison is successful, it is determined that the staff member is not a newly joined staff member. The staff member's work trajectory is recorded according to the staff member's employee number. Specifically, the staff member's work trajectory includes the staff member's location at each point in time.

[0046] In practice, monitoring equipment is installed in various work areas, and image recognition technology is used to determine whether the staff member is in the work area, thereby determining the staff member's location at each point in time.

[0047] Step S120: If the comparison fails, the location of the staff member at each time point is determined based on the preset monitoring equipment, and a spatiotemporal curve to be judged is constructed.

[0048] In the specific implementation process, if the comparison fails, the employee is determined to be a newly joined employee. Specifically, this solution divides the employee's working time into multiple time points with the same time interval. For example, if the working time is 9:00-18:00, a total of 9 hours, and the selected time interval is 0.5 hours, then the time points to be judged include 9:00, 9:30, 10:00, 10:30, 11:00, 11:30, 12:00, 12:30, 13:00, 13:30, 14:00, 14:30, 15:00, 15:30, 16:00, 16:30, 17:00, 17:30, and 18:00.

[0049] Step S200: Obtain the standard spatiotemporal curves of each working group, compare the spatiotemporal curve to be judged with each standard spatiotemporal curve, and obtain a single comparison vector corresponding to each standard spatiotemporal curve.

[0050] In the specific implementation process, at the same time point, the two curves are compared to see if they are marked with the same vertical coordinate height position.

[0051] Step S300: Combine the individual alignment vectors into a combined alignment vector, input the combined alignment vector into a pre-trained neural network model, and output the work group to which the worker belongs.

[0052] Step S400: Input the feature vector of the worker into the feature vector set of the work.

[0053] By adopting the above scheme, this scheme can extract the feature vector of each staff member through monitoring equipment, and determine the work group to which the staff member belongs through the feature vector. Since the job types of each work group may be different, this scheme can monitor them separately to ensure the targeting of the supervision. On the other hand, for newly joined staff members, this scheme can determine the work group to which the staff member belongs through a series of judgment schemes and classify the staff member's feature vector. The grouping of staff members can be completed automatically without human intervention, thereby improving the degree of automation of supervision. This scheme can ensure the effectiveness and efficiency of supervision.

[0054] In some embodiments of the present invention, the spatiotemporal curve to be determined and the standard spatiotemporal curve are both established in the same Cartesian coordinate system. Each position of the horizontal axis of the Cartesian coordinate system represents a time point, and each position of the vertical axis represents a location. In the step of determining the location of the worker at each time point based on the preset monitoring equipment and constructing the spatiotemporal curve to be determined, based on the location of the worker at each time point, the corresponding points are marked in the Cartesian coordinate system, and the marked points are connected in chronological order to obtain the spatiotemporal curve to be determined.

[0055] By adopting the above scheme, the work trajectory of the staff can be represented in a Cartesian coordinate system, which can efficiently represent the work trajectory of the staff. Moreover, by representing it in the same Cartesian coordinate system, it is convenient to compare the work trajectory of the staff with the standards of each work group.

[0056] like Figure 2 As shown, in some embodiments of the present invention, before the step of obtaining the standard spatiotemporal curves of each working group, the standard spatiotemporal curves of each working group are calculated. In the step of calculating the standard spatiotemporal curves of each working group, step S130, the positions of the staff in the same working group at each time point are obtained, and the corresponding points are marked in the Cartesian coordinate system. The marked points are connected based on the time sequence to construct the standard spatiotemporal curve of the working group.

[0057] In some embodiments of the present invention, the horizontal axis of the Cartesian coordinate system is sequentially marked with time points of 9:00, 9:30, 10:00, 10:30, 11:00, 11:30, 12:00, 12:30, 13:00, 13:30, 14:00, 14:30, 15:00, 15:30, 16:00, 16:30, 17:00, 17:30, and 18:00, and each height on the vertical axis represents a working area.

[0058] In some embodiments of the present invention, in the step of obtaining the positions of staff members in the same work group at various time points and marking the corresponding points in a Cartesian coordinate system, if there are multiple staff members in different positions at the same time point, the marked point is the point corresponding to the position where the most staff members are located at that time point.

[0059] In some embodiments of the present invention, if there are 3 workers in the work group, 2 of whom are in work area b1 at time A and 1 is in work area b2 at time A, then the position with the most workers at time A is work area b1. The horizontal axis is determined to be time A and the vertical axis is the point corresponding to work area b1 as the calibrated point.

[0060] The above approach is problematic because the work of working groups is often adjusted according to actual needs, and each working group does not have fixed tasks. Therefore, it is difficult to use a standard spatiotemporal curve for working groups in a long-term manner. This approach generates a corresponding standard spatiotemporal curve based on the actual situation, ensuring the accurate representation of the standard spatiotemporal curve for each working group and further guaranteeing the accuracy of subsequent working group matching.

[0061] In some embodiments of the present invention, in the step of comparing the feature vector of the staff member with each feature vector in the feature vector set corresponding to each current work group, the similarity between the feature vector of the staff member and each feature vector in the feature vector set corresponding to each current work group is calculated. If any similarity value is greater than a preset comparison threshold, the comparison is successful; otherwise, the comparison fails.

[0062] In the specific implementation process, if any similarity value is greater than the preset comparison threshold, then the staff member is determined to be the staff member corresponding to the feature vector of the work group.

[0063] In some embodiments of the present invention, in the step of calculating the similarity between the feature vector of a worker and each feature vector in the feature vector set corresponding to each current work group, the similarity is calculated based on the following formula:

[0064] similar=(1+α T β) / 2;

[0065] Where similar represents similarity, α T Let α represent the transpose of any eigenvector in the eigenvector set, β represent the eigenvector of the worker to be determined, and α represent the eigenvector of the worker to be determined. T β represents the inner product of the feature vector of the worker to be judged and the transpose of any feature vector in the feature vector set.

[0066] In some embodiments of the present invention, in the step of comparing the spatiotemporal curve to be judged with each standard spatiotemporal curve to obtain a single comparison vector corresponding to each standard spatiotemporal curve, points in the standard spatiotemporal curve at the same time point and points in the spatiotemporal curve to be judged are compared respectively. If they coincide, a first parameter is generated; if they are different, a second parameter is generated. The first parameter and the second parameter are combined based on the order of the time points to obtain a single comparison vector corresponding to each standard spatiotemporal curve.

[0067] In some embodiments of the present invention, the first parameter is set to 1 and the second parameter is set to 0. If the working time is 9:00-18:00, a total of 9 hours, and the selected time interval is 0.5 hours, then each of the single comparison vectors is a vector with 19 dimensions.

[0068] By adopting the above scheme, this scheme can generate a single alignment vector for each working group, ensuring the accuracy of the comparison.

[0069] In some embodiments of the present invention, in the steps of assembling individual alignment vectors into a combined alignment vector, inputting the combined alignment vector into a pre-trained neural network model, and the pre-trained neural network model outputting the work group to which the worker belongs, multiple individual alignment vectors are sequentially connected to obtain a combined alignment vector, the combined alignment vector is input into the pre-trained neural network model, and the classification layer of the neural network model is used for classification to determine the work group to which the worker belongs.

[0070] In some embodiments of the present invention, the classification result of the classification layer of the neural network model includes a classification result belonging to any work group and a classification result not belonging to any work group. If the output result is a classification result not belonging to any work group, the manager is notified and the staff member's input information is reported to the manager's control terminal for further judgment. This can further ensure the safety of the construction site and prevent unauthorized personnel from sneaking into the construction site.

[0071] In some embodiments of the present invention, if there are a total of 5 working groups, and each of the single comparison vectors is a vector consisting of 19 values, then the combined comparison vector is a vector with 95 dimensions.

[0072] By adopting the above scheme, this scheme can combine the comparison results with multiple working groups and make the final judgment through a pre-trained neural network model. Since each staff member has a certain degree of freedom in the actual work process, it can avoid the deviation of a single staff member from the standard working group and make a combined judgment.

[0073] In some embodiments of the present invention, in the step of obtaining the feature vector of the staff based on the preset monitoring device, the monitoring image of the staff is obtained based on the preset monitoring device, and the ReID feature vector of the staff is identified based on the monitoring image as the feature vector of the staff.

[0074] In some embodiments of the present invention, after the step of inputting the feature vector of the worker into the feature vector set of the work, the method further includes the step of reconstructing the standard spatiotemporal curve of the work group based on the positions of other workers in the work group at various time points, in accordance with the method of constructing the standard spatiotemporal curve of the work group.

[0075] This invention also provides an inspection robot inspection system, which includes a computer device, a processor and a memory, wherein the memory stores computer instructions, and the processor executes the computer instructions stored in the memory. When the computer instructions are executed by the processor, the system implements the steps of the method.

[0076] This invention also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the aforementioned model construction method based on network architecture search. The computer-readable storage medium can be a tangible storage medium, such as random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, floppy disks, hard disks, removable storage disks, CD-ROMs, or any other form of storage medium known in the art.

[0077] Those skilled in the art will understand that the exemplary components, systems, and methods described in conjunction with the embodiments disclosed herein can be implemented in hardware, software, or a combination of both. Whether implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this invention. When implemented in hardware, it can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this invention are programs or code segments used to perform the desired tasks. The programs or code segments can be stored in a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried in a carrier wave.

[0078] It should be clarified that the present invention is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of the present invention.

[0079] In this invention, features described and / or illustrated for one embodiment may be used in the same or similar manner in one or more other embodiments, and / or combined with or in place of features of other embodiments.

[0080] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, various modifications and variations of the embodiments of the present invention are possible. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for intelligent safety supervision of construction sites, characterized in that, The steps of the method include: The feature vectors of the staff are obtained based on the preset monitoring equipment. The feature vectors of the staff are compared with the feature vectors of each work group in the current work group. If the comparison is successful, the work group to which the staff belongs is determined. The work trajectory of the staff is recorded based on the preset monitoring equipment. If the comparison fails, the location of the staff member at each time point is determined based on the preset monitoring equipment, and a spatiotemporal curve to be judged is constructed. Calculate the standard spatiotemporal curves of each work group. In the step of calculating the standard spatiotemporal curves of each work group, obtain the positions of the staff in the same work group at each time point. If there are multiple staff members in different positions at the same time point, the marked point is the point corresponding to the position with the most staff members at that time point. Mark the corresponding point in the Cartesian coordinate system. Connect the marked points according to the time order to construct the standard spatiotemporal curve of the work group. Obtain the standard spatiotemporal curves of each working group, compare the spatiotemporal curve to be judged with each standard spatiotemporal curve to obtain a single comparison vector for each standard spatiotemporal curve, compare the points in the standard spatiotemporal curve at the same time point with the points in the spatiotemporal curve to be judged, if they overlap, generate the first parameter, if they are different, generate the second parameter, and combine the corresponding first parameter and second parameter based on the order of time points to obtain a single comparison vector for each standard spatiotemporal curve. The individual alignment vectors are combined into a combined alignment vector, which is then input into a pre-trained neural network model. The pre-trained neural network model outputs the work group to which the worker belongs. Multiple individual alignment vectors are sequentially concatenated to obtain a combined alignment vector, which is then input into a pre-trained neural network model. The classification layer of the neural network model is used for classification to determine the work group to which the worker belongs. The feature vector of the employee is entered into the feature vector set for that job.

2. The intelligent safety monitoring method for construction sites according to claim 1, characterized in that: Both the spatiotemporal curve to be judged and the standard spatiotemporal curve are established in the same Cartesian coordinate system. Each position on the horizontal axis of the Cartesian coordinate system represents a time point, and each position on the vertical axis represents a location. In the step of determining the location of the staff member at each time point based on the preset monitoring equipment and constructing the spatiotemporal curve to be judged, the corresponding points are marked in the Cartesian coordinate system based on the location of the staff member at each time point. The marked points are connected in chronological order to obtain the spatiotemporal curve to be judged.

3. The intelligent safety monitoring method for construction sites according to claim 1 or 2, characterized in that: In the step of comparing the feature vectors of staff members with the feature vector sets corresponding to each work group, the similarity between the feature vectors of staff members and the feature vector sets corresponding to each work group is calculated. If any similarity value is greater than the preset comparison threshold, the comparison is successful; otherwise, the comparison fails.

4. The intelligent safety monitoring method for construction sites according to claim 1, characterized in that: In the step of calculating the similarity between the feature vectors of staff members and the feature vector sets corresponding to each work group, the similarity is calculated based on the following formula: ； in, Indicates similarity. This represents the transpose of any eigenvector in the eigenvector set. This represents the feature vector of the staff member to be judged. This represents the inner product of the feature vector of the worker to be judged and the transpose of any feature vector in the feature vector set.

5. The intelligent safety monitoring method for construction sites according to claim 1, characterized in that: In the step of obtaining the feature vector of the staff based on the preset monitoring equipment, the monitoring image of the staff is obtained based on the preset monitoring equipment, and the ReID feature vector of the staff is identified based on the monitoring image as the feature vector of the staff.

6. An inspection robot inspection system, characterized in that: The system includes a computer device, which includes a processor and a memory. The memory stores computer instructions, and the processor executes the computer instructions stored in the memory. When the computer instructions are executed by the processor, the system implements the steps of the method according to any one of claims 1-5.

Images