Construction site safety monitoring method and system based on image analysis

Abstract

The embodiment of the present application provides a construction site safety monitoring method and system based on image analysis, which relates to the field of artificial intelligence technology. By acquiring and analyzing the real-time image of the construction site and the risky construction path information in the construction drawings, efficient monitoring of the safety status of the construction site is achieved. The method can accurately analyze the characteristic deviation and deviation type between the on-site construction process image and the risky construction path, and then generate a construction safety confidence array and a correlation array, accurately identify the risky construction behavior corresponding to the target construction monitoring object in the construction drawings, effectively improve the intelligence level of construction site safety monitoring, and can timely discover and warn of potential safety hazards, providing strong technical support for the safety management of the construction site, significantly reducing the incidence of construction accidents, and ensuring the life safety of construction personnel.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and more specifically, to a construction site safety monitoring method and system based on image analysis. Background Art

[0002] In the safety management of construction sites, monitoring and warning of potential risky construction behaviors are crucial. Traditional solutions mainly rely on manual inspections and paper records, which are not only inefficient, but also prone to omissions and misjudgments due to human factors. Especially in large and complex construction projects, the monitoring scope of the construction site is wide and there are many risk factors. Traditional monitoring methods often find it difficult to achieve comprehensive and real-time monitoring.

[0003] In the field of construction site safety monitoring, due to the complexity and variability of the construction site environment, as well as the differences between construction drawings and actual on-site construction, how to accurately extract useful information from on-site construction process images and effectively compare and analyze it with the risky construction path information in the construction drawings is still a problem that needs to be solved urgently.

[0004] At present, although there are some construction site monitoring methods based on image analysis, most of these methods can only perform simple image recognition and analysis on the construction site, and cannot conduct in-depth analysis and early warning of the safety status of the construction site. In particular, when risky construction behaviors that do not conform to the construction drawings occur on the construction site, these methods often cannot detect and warn in time, thus posing hidden dangers to the safety of the construction site. Summary of the invention

[0005] In order to at least overcome the above-mentioned deficiencies in the prior art, the purpose of the present application is to provide a construction site safety monitoring method and system based on image analysis.

[0006] In a first aspect, the present application provides a construction site safety monitoring method based on image analysis, which is applied to a construction site safety monitoring system based on image analysis, and the method comprises:

[0007] Acquire the on-site construction process images acquired by the target construction monitoring object during the dynamic monitoring process, and acquire the risk construction path information within the target construction partition in the construction drawing; the target construction partition is determined after the monitoring range of the target construction monitoring object is aligned with the coordinate system of the construction drawing, and the risk construction path information is information constraining the path characteristics of the construction path of the risk construction behavior of the historical work ticket within the target construction partition;

[0008] Analyzing the characteristic deviation between the on-site construction process image and the risky construction path information to generate construction characteristic deviation information, and analyzing the deviation type between the on-site construction process image and the risky construction path information to generate construction deviation type information;

[0009] Based on the construction feature deviation information and the construction deviation type information, the construction safety confidence between the on-site construction process image and the risky construction path information is analyzed to generate a construction safety confidence array;

[0010] Analyzing the correlation between the on-site construction process image and the risky construction path information according to the construction safety confidence array to generate a correlation array;

[0011] The risky construction behavior corresponding to the target construction monitoring object in the construction drawing is determined based on the correlation array.

[0012] In a second aspect, an embodiment of the present application further provides a construction site safety monitoring system based on image analysis, the construction site safety monitoring system based on image analysis comprising a processor and a machine-readable storage medium, the machine-readable storage medium storing a computer program, the computer program being loaded and executed in conjunction with the processor to implement the construction site safety monitoring method based on image analysis of the above first aspect.

[0013] In combination with the above aspects, the embodiment of the present application achieves efficient monitoring of the safety status of the construction site by acquiring and analyzing the real-time image of the construction site and the risky construction path information in the construction drawings. The method can accurately analyze the characteristic deviation and deviation type between the on-site construction process image and the risky construction path, and then generate a construction safety confidence array and a correlation array, accurately identify the risky construction behavior corresponding to the target construction monitoring object in the construction drawings, effectively improve the intelligent level of construction site safety monitoring, and can timely discover and warn of potential safety hazards, providing strong technical support for the safety management of the construction site, significantly reducing the incidence of construction accidents, and ensuring the life safety of construction personnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required to be used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0015] Figure 1A schematic diagram of a construction site safety monitoring method based on image analysis provided in an embodiment of the present application;

[0016] Figure 2 A structural schematic block diagram of a construction site safety monitoring system based on image analysis for implementing the above-mentioned construction site safety monitoring method based on image analysis provided in an embodiment of the present application. DETAILED DESCRIPTION

[0017] The following description is intended to enable one of ordinary skill in the art to implement and utilize the present application, and the description is provided in the context of a specific application scenario and its requirements. It will be apparent to one of ordinary skill in the art that various changes may be made to the disclosed embodiments, and that the general principles defined in the present application may be applied to other embodiments and application scenarios without departing from the principles and scope of the present application. Therefore, the present application is not limited to the described embodiments, but should be given the broadest scope equal to the claims.

[0018] Step S110, obtaining the on-site construction process image obtained by the target construction monitoring object during the dynamic monitoring process, and obtaining the risk construction path information in the target construction partition in the construction drawing. The target construction partition is determined after the monitoring range of the target construction monitoring object is aligned with the coordinate system of the construction drawing, and the risk construction path information is information for constraining the path characteristics of the construction path of the risk construction behavior of the historical work ticket in the target construction partition.

[0019] In detail, the on-site construction process image is an image that reflects the actual construction activities at the construction site, and includes key information such as the operations of construction personnel, the use of construction equipment, and the status of construction materials.

[0020] The construction drawings are detailed plans of the construction project, including information such as the structure, layout, size, materials, etc. of the building, as well as various instructions and requirements during the construction process. The target construction zone is a specific construction area determined by aligning the monitoring range of the monitoring equipment with the coordinate system of the construction drawings. This area is the focus of monitoring and analysis.

[0021] The risky construction path information is information that imposes characteristic constraints on the construction paths of risky construction behaviors recorded in historical work tickets within the target construction zone. It describes the construction paths and behaviors that should be avoided or paid special attention to when constructing in these areas to prevent the occurrence of safety accidents.

[0022] In this embodiment, it is assumed that there is a large-scale construction project, and multiple surveillance cameras are deployed at the construction site as target construction monitoring objects. The server first obtains the on-site construction process images captured by these target construction monitoring objects during the dynamic monitoring process.

[0023] For example, one of the surveillance cameras captured a group of workers welding steel beams. The on-site construction process images clearly showed the workers' operating postures, the use of welding equipment, and the position and status of the steel beams.

[0024] At the same time, the server needs to obtain the risk construction path information within the target construction partition in the construction drawing. First, the server aligns the monitoring range of the target construction monitoring object with the coordinate system of the construction drawing. Assume that the coordinate area corresponding to the monitoring range is (X1, Y1) to (X2, Y2). In the construction drawing, determine the partitions located around this coordinate area, such as the area where concrete pouring and electrical wiring are being carried out, and set it as the target construction partition.

[0025] Then, the risky construction path information is obtained from the construction planning information of the historical work tickets in the target construction zone. For example, the historical work tickets record that when pouring concrete in the zone, if the pouring speed is too fast or the pouring sequence is incorrect, it may cause uneven concrete structure, which is regarded as a risky construction behavior. The corresponding construction path characteristics include the starting point, direction, speed and other information of pouring; in terms of electrical wiring, if the line direction does not meet the specifications or crosses and conflicts with other pipelines, it is also identified as a risky construction behavior. The relevant construction path characteristics cover the layout of the line, turning points, etc.

[0026] Step S120, analyzing the characteristic deviation between the on-site construction process image and the risky construction path information to generate construction characteristic deviation information, and analyzing the deviation type between the on-site construction process image and the risky construction path information to generate construction deviation type information.

[0027] In detail, the feature deviation refers to the degree of difference between the construction features presented in the on-site construction process image and the standard features specified in the risky construction path information.

[0028] The deviation type refers to different types of deviations between the on-site construction process image and the risky construction path information, such as the wrong sequence of construction steps, the selection deviation of the construction path, etc.

[0029] That is, in this embodiment, the server analyzes the characteristic deviation between the acquired on-site construction process image and the risky construction path information to generate construction characteristic deviation information.

[0030] Taking the on-site construction process image of steel beam welding as an example, the server extracts the actual operation steps of the welding workers, the location and sequence of welding points and other features in the image to form a construction representation vector corresponding to the on-site construction process image. At the same time, the server extracts the standardized steel beam welding construction steps, the standard location and sequence of welding points and other features from the risky construction path information to form a construction representation vector corresponding to the risky construction path information.

[0031] Then the cosine distance between the two construction characterization vectors is calculated. Assuming that there is a certain deviation between the welding point sequence of the workers in the on-site construction and the sequence in the specification, the cosine distance is larger through calculation.

[0032] Next, the server analyzes the construction path differences between the on-site construction process image and the risky construction path information based on the construction representation vector corresponding to the on-site construction process image, the construction representation vector corresponding to the risky construction path information, the construction accuracy confidence level of the on-site construction path corresponding to the on-site construction process image (for example, set according to the accuracy and reliability of the monitoring equipment), and the cosine distance calculated above, and generates the construction path deviation information.

[0033] For the generation of the construction step sequence deviation, it is assumed that the number of construction paths corresponding to the steel beam welding step sequence displayed in the on-site construction process image is 10, and the total number of on-site construction paths in the on-site construction process image is 20, and the first ratio information is 0.5. In the risky construction path information, the number of construction paths corresponding to the specified welding step sequence is 15, and the number of construction paths in the risky construction path information is 30, and the second ratio information is 0.5. By analyzing these two ratio information, it is found that there is a certain difference between the two, so the construction step sequence deviation is generated.

[0034] At the same time, the server analyzes the deviation type between the on-site construction process image and the risky construction path information to generate construction deviation type information.

[0035] Assume that the first ratio information of the construction step sequence of steel beam welding and the number of on-site construction paths in the on-site construction process image is 0.4, the second ratio information of the welding construction step sequence and the number of construction paths specified in the risk construction path information is 0.6, the number of on-site construction paths is 25, and the number of construction paths in the risk construction path information is 40. Through comprehensive analysis of these parameters, the deviation type is determined, such as whether it is an error in the order of construction steps or a deviation in the selection of construction paths, and the construction deviation type information is generated.

[0036] Step S130: Based on the construction feature deviation information and the construction deviation type information, the construction safety confidence between the on-site construction process image and the risky construction path information is analyzed to generate a construction safety confidence array.

[0037] In detail, the construction safety confidence refers to the trust in the degree of match between the on-site construction process image and the risky construction path information, reflecting the consistency between the on-site construction activities and the safety standards. The construction safety confidence array is an array or matrix containing multiple construction safety confidence values. Each value corresponds to the degree of match between the on-site construction process image and the risky construction path information in a specific aspect.

[0038] That is, in this embodiment, the server analyzes the construction safety confidence between the on-site construction process image and the risky construction path information based on the generated construction feature deviation information and construction deviation type information, and generates a construction safety confidence array.

[0039] First, the server obtains a first adjustment coefficient (e.g., 0.3) corresponding to the construction step sequence deviation and a second adjustment coefficient (e.g., 0.7) corresponding to the construction deviation type information. The first adjustment coefficient is used to adjust the importance coefficient corresponding to the construction step sequence deviation, and the second adjustment coefficient is used to adjust the importance coefficient corresponding to the construction deviation type information.

[0040] Then, the target construction step sequence deviation is generated based on the construction step sequence deviation and the first adjustment coefficient. Assuming that the construction step sequence deviation is 0.2, multiplying it by the first adjustment coefficient 0.3 results in a target construction step sequence deviation of 0.06. Similarly, the target construction deviation type information is generated based on the construction deviation type information and the second adjustment coefficient.

[0041] Finally, based on the construction path deviation information, the target construction step sequence deviation and the target construction deviation type information, the construction safety confidence of matching the on-site construction process image to the risky construction path information is analyzed to generate a construction safety confidence array.

[0042] For example, the construction path deviation information shows that the deviation degree is relatively high at 0.4, the target construction step sequence deviation is 0.06, and the target construction deviation type information shows that the deviation type is relatively serious at 0.3. Combining this information, a construction safety confidence array is generated through a specific algorithm and weight allocation, in which each element represents the safety confidence value of a different construction link.

[0043] Step S140: analyzing the correlation between the on-site construction process image and the risky construction path information according to the construction safety confidence array to generate a correlation array.

[0044] In detail, the correlation refers to the degree of association between the on-site construction process image and the risky construction path information, reflecting the similarity or consistency between the two in terms of construction characteristics, construction steps, construction paths, etc. The correlation array is an array or matrix used to represent the correlation between the on-site construction process image and the risky construction path information. By analyzing and processing the construction safety confidence array, a correlation array can be obtained to further evaluate the safety and compliance of construction activities.

[0045] In this embodiment, the server analyzes the correlation between the on-site construction process image and the risky construction path information according to the construction safety confidence array to generate a correlation array.

[0046] The server first obtains the preset risk adjustment weight, which is assumed to be 0.5. This weight is used to limit the weight of the construction safety confidence array to the correlation array.

[0047] Then, a basic correlation array is constructed based on the construction safety confidence array and preset risk adjustment weights.

[0048] Next, the rows in the basic correlation array are regularized so that the rows of the basic correlation array match the construction representation vectors corresponding to the risky construction path information. For example, the element values ​​in the rows are adjusted so that they can accurately reflect the correlation between the various features in the risky construction path information and the on-site construction process images. At the same time, the columns in the basic correlation array are regularized so that the columns of the basic correlation array match the construction representation vectors corresponding to the on-site construction process images until the set convergence requirements are met and the correlation array is generated.

[0049] Assume that during the processing, it is found that the element value related to a specific construction step in a row is too high or too low, and it is adjusted to make it more consistent with the actual correlation situation. The same is true for the processing of columns, ensuring that the correlation array can accurately reflect the relationship between the on-site construction process image and the risky construction path information.

[0050] Step S150: determining the risky construction behavior corresponding to the target construction monitoring object in the construction drawing based on the correlation array.

[0051] In detail, the risky construction behavior refers to the behavior that may cause safety accidents or quality problems during the construction process, which is usually contrary to the provisions in the construction drawings, or violates the construction safety standards and specifications.

[0052] That is, in this embodiment, the server determines the risky construction behavior corresponding to the target construction monitoring object in the construction drawing based on the generated correlation array.

[0053] First, based on the correlation array, a target construction path whose correlation with the on-site construction path corresponding to the on-site construction process image is greater than a threshold value (eg, 0.8) is determined among the construction paths corresponding to the risky construction path information.

[0054] Assume that the correlation array shows that in the construction path of concrete pouring, some paths have a correlation with the paths in the on-site construction image that exceeds the threshold value. Then, based on these target construction paths, the risky construction behaviors corresponding to the target construction monitoring objects in the construction drawings are determined. For example, it is determined whether the risky construction behaviors are too fast in concrete pouring speed, wrong pouring sequence, or inaccurate pouring point location.

[0055] It is worth noting that the above is just an example scenario built for detailed explanation. Actual construction monitoring and analysis may be more complex and diverse, and need to be adjusted and optimized according to specific construction projects and monitoring needs.

[0056] Based on the above steps, the embodiment of the present application achieves efficient monitoring of the safety status of the construction site by acquiring and analyzing the real-time image of the construction site and the risky construction path information in the construction drawings. This method can accurately analyze the characteristic deviation and deviation type between the on-site construction process image and the risky construction path, and then generate a construction safety confidence array and a correlation array, accurately identify the risky construction behavior corresponding to the target construction monitoring object in the construction drawings, effectively improve the intelligent level of construction site safety monitoring, and can timely discover and warn of potential safety hazards, providing strong technical support for the safety management of the construction site, significantly reducing the incidence of construction accidents, and ensuring the life safety of construction personnel.

[0057] In a possible implementation, step S120 includes:

[0058] Step S121, determining a cosine distance between a construction representation vector corresponding to the on-site construction process image and a construction representation vector corresponding to the risky construction path information.

[0059] Step S122, based on the construction representation vector corresponding to the on-site construction process image, the construction representation vector corresponding to the risky construction path information, the construction accuracy confidence of the on-site construction path corresponding to the on-site construction process image, and the cosine distance, the construction path difference between the on-site construction process image and the risky construction path information is analyzed to generate construction path deviation information.

[0060] Step S123, analyzing the deviation between the construction step sequence presented by the on-site construction path corresponding to the on-site construction process image and the construction step sequence presented by the construction path corresponding to the risky construction path information, and generating a construction step sequence deviation.

[0061] Step S124: determining the construction path deviation information and the construction step sequence deviation as the construction feature deviation information.

[0062] In this embodiment, it is further assumed that there is a large commercial complex construction project, and multiple surveillance cameras are deployed at the construction site as target construction monitoring objects. The server first obtains the on-site construction process images captured by these target construction monitoring objects during the dynamic monitoring process.

[0063] For example, one of the surveillance cameras captured a group of workers welding a steel beam. The images of the on-site construction process clearly showed the workers' operating postures, the use of welding equipment, and the position and status of the steel beam.

[0064] The server also obtains the risky construction path information within the target construction partition in the construction drawing. First, the server aligns the monitoring range of the target construction monitoring object with the coordinate system of the construction drawing. Assume that the coordinate area corresponding to the monitoring range is (X1, Y1) to (X2, Y2). In the construction drawing, determine the partitions located around this coordinate area, such as the area where concrete pouring and electrical wiring are being carried out, and set it as the target construction partition.

[0065] Then, the risky construction path information is obtained from the construction planning information of the historical work tickets in the target construction zone. For example, the historical work tickets record that when pouring concrete in the zone, if the pouring speed is too fast or the pouring sequence is incorrect, it may cause uneven concrete structure, which is regarded as a risky construction behavior. The corresponding construction path characteristics include the starting point, direction, speed and other information of pouring; in terms of electrical wiring, if the line direction does not meet the specifications or crosses and conflicts with other pipelines, it is also identified as a risky construction behavior. The relevant construction path characteristics cover the layout of the line, turning points, etc.

[0066] Next, the server analyzes the feature deviation between the on-site construction process image and the risky construction path information to generate construction feature deviation information.

[0067] First, the cosine distance between the construction representation vector corresponding to the on-site construction process image and the construction representation vector corresponding to the risky construction path information is determined. The server extracts features from the on-site construction process image of steel beam welding, such as the actual operation sequence of welding workers, the location distribution of welding points, and other features to form a construction representation vector corresponding to the on-site construction process image. At the same time, the standardized steel beam welding construction steps, the standard welding point location distribution, and other features are extracted from the risky construction path information to form a construction representation vector corresponding to the risky construction path information. Assuming that the feature vector of the on-site construction is [0.8, 0.6, 0.5, 0.3] and the feature vector of the standardized risky construction path is [0.9, 0.7, 0.4, 0.2], the cosine distance is calculated to be 0.7, which indicates that there is a certain deviation in the features between the on-site construction and the standardized construction path.

[0068] Then, based on the construction representation vector corresponding to the on-site construction process image, the construction representation vector corresponding to the risky construction path information, the construction accuracy confidence of the on-site construction path corresponding to the on-site construction process image, and the cosine distance calculated above, the construction path difference between the on-site construction process image and the risky construction path information is analyzed to generate the construction path deviation information. Assume that the construction accuracy confidence of the on-site construction path is 0.85 (derived from the accuracy and reliability evaluation of the monitoring equipment). The server comprehensively considers the differences in elements such as the spacing of welding points and welding sequence in the construction representation vector, as well as the cosine distance and construction accuracy confidence calculated above. For example, the spacing of some welding points in the on-site construction is slightly different from the specification requirements. Combining these factors and the corresponding weight distribution, the construction path deviation is calculated to be 0.3, which means that there is a relatively obvious difference between the on-site construction path and the risky construction path.

[0069] Next, the deviation between the construction step sequence presented by the on-site construction path corresponding to the on-site construction process image and the construction step sequence presented by the construction path corresponding to the risk construction path information is analyzed to generate the construction step sequence deviation. Assuming that the number of construction paths corresponding to the steel beam welding step sequence displayed in the on-site construction process image is 12, and the total number of on-site construction paths in the on-site construction process image is 20, the first ratio information is 0.6. In the risk construction path information, the number of construction paths corresponding to the prescribed welding step sequence is 15, while the number of construction paths in the risk construction path information is 30, and the second ratio information is 0.5. Through detailed comparison and analysis of the two ratio information, it is found that there are certain differences between the two. Further considering the importance of each step and the construction environment and other factors, after quantification and evaluation, it is calculated that the construction step sequence deviation is 0.2.

[0070] Finally, the construction path deviation information 0.3 and the construction step sequence deviation 0.2 are determined as the construction feature deviation information.

[0071] In another scene, surveillance cameras captured workers pouring concrete.

[0072] The server extracts features from the on-site construction process images of concrete pouring, including the pouring speed, the distribution of pouring points, etc., to form a construction characterization vector for on-site construction. The standardized concrete pouring speed range, the standard distribution of pouring points, etc. are extracted from the risky construction path information to form a construction characterization vector for the risky construction path.

[0073] Calculating the cosine distance between the two, assuming it is 0.65, shows that there is a certain deviation.

[0074] Based on the construction characterization vector of on-site construction, the construction characterization vector of the risky construction path, the construction accuracy confidence of the on-site construction path (assumed to be 0.78) and the cosine distance, the differences in construction paths are analyzed. For example, it is found that the pouring speed is slightly faster than the specification requirements, and the construction path deviation information is calculated to be 0.25.

[0075] Analyze the deviation of the construction step sequence. Assume that the number of construction paths corresponding to the concrete pouring step sequence in the on-site construction is 8, the total number of construction paths is 15, and the first ratio is 0.53. The number of construction paths corresponding to the pouring step sequence specified in the risk construction path is 10, the total number of construction paths is 20, and the second ratio is 0.5. The deviation of the construction step sequence is calculated to be 0.15.

[0076] The construction path deviation information of 0.25 and the construction step sequence deviation of 0.15 are determined as the construction feature deviation information.

[0077] In a possible implementation, the on-site construction process image includes the number of on-site construction paths, and the risky construction path information includes the number of construction paths. Step S123 includes:

[0078] Step S1231, determining first ratio information between the number of construction paths corresponding to the construction step sequence presented by the on-site construction path and the number of on-site construction paths. Determine second ratio information between the number of construction paths corresponding to the construction step sequence presented by the construction path and the number of construction paths.

[0079] Step S1232: performing construction step sequence deviation analysis based on the first proportion information and the second proportion information to generate the construction step sequence deviation degree.

[0080] In this embodiment, in this large commercial complex construction project, the server continuously analyzes on-site construction process images and risky construction path information.

[0081] Taking the construction link of steel beam welding as an example, there are 30 construction paths displayed in the on-site construction process image. Among them, the number of steel beam welding construction paths performed in a specific construction step sequence is 18. Then, the first ratio information between the number of construction paths corresponding to the construction step sequence presented by the on-site construction path and the number of on-site construction paths is 18÷30=0.6.

[0082] At the same time, it can be seen from the risk construction path information that the total number of construction paths is 50, of which the number of steel beam welding construction paths performed in accordance with the standard construction step sequence is 30. Then, the second ratio information between the number of construction paths corresponding to the construction step sequence presented by the construction path and the number of construction paths is 30÷50=0.6.

[0083] By comparing and analyzing the two proportion information 0.6 and 0.6, it can be found that the two are the same, which means that in terms of the construction step sequence of steel beam welding, there is no deviation between the on-site construction and the risk construction path information.

[0084] Let's look at the construction link of concrete pouring. There are 40 construction paths for concrete pouring in the on-site construction process image, of which 20 are construction paths that are carried out in a specific construction step sequence, and the first ratio information is 20÷40=0.5.

[0085] In the risky construction path information, the total number of construction paths for concrete pouring is 60, the number of construction paths carried out in accordance with the standard construction step sequence is 36, and the second proportion information is 36÷60=0.6.

[0086] After analyzing the two ratio information 0.5 and 0.6, it was found that there was a difference. The server further calculated that the difference between the two was 0.6-0.5=0.1, indicating that there was a certain degree of deviation between the sequence of concrete pouring construction steps in the on-site construction and the risk construction path information.

[0087] In the construction phase of electrical wiring, there are 50 on-site construction paths, of which 30 are construction paths in a specific sequence of construction steps, and the first ratio is 30÷50=0.6.

[0088] The total number of construction paths in the risky construction path information is 80, the number of construction paths with standardized construction step sequences is 48, and the second ratio is 48÷80=0.6.

[0089] Since the two ratios are the same, it means that the construction step sequence of the electrical wiring is consistent with the on-site construction and risk construction path information without deviation.

[0090] For the pipeline laying construction phase, the total number of on-site construction paths is 60, the number of paths in a specific construction step sequence is 35, and the first ratio is 35÷60≈0.58.

[0091] The total number of paths in the risky construction path information is 90, the number of paths with standardized construction step sequences is 54, and the second ratio is 54÷90=0.6.

[0092] There is a difference in the two proportions, and the calculated difference of 0.6-0.58=0.02 indicates that there is a small deviation in the sequence of construction steps for on-site pipeline laying.

[0093] The server integrates the proportional information analysis results of multiple construction links to comprehensively evaluate the deviation of the construction step sequence of different construction parts in the entire project, thereby providing a basis for generating accurate construction step sequence deviation, so as to timely discover and correct irregularities in construction and ensure construction quality and safety.

[0094] In a possible implementation, the on-site construction process image includes the number of on-site construction paths, and the risky construction path information includes the number of construction paths. Step S120 includes:

[0095] Step S121, determining first ratio information between the sequence of construction steps presented by the on-site construction path corresponding to the on-site construction process image and the number of the on-site construction paths. Determine second ratio information between the sequence of construction steps presented by the construction path corresponding to the risky construction path information and the number of the construction paths.

[0096] Step S122, based on the first proportion information, the second proportion information, the number of on-site construction paths and the number of construction paths, the deviation type between the on-site construction process image and the risky construction path information is analyzed to generate the construction deviation type information.

[0097] In this embodiment, in the aforementioned large commercial complex construction project, the server continues to perform analysis work.

[0098] Taking electrical wiring construction as an example, the number of on-site construction paths displayed in the on-site construction process image is 80. Among them, the number of paths performed in a specific construction step sequence is 40. Then, the first ratio information between the construction step sequence presented by the on-site construction path and the number of on-site construction paths is 40÷80=0.5.

[0099] In the risky construction path information, the total number of construction paths is 120, and the number of paths that follow the standard construction step sequence is 70. Therefore, the second ratio information between the construction step sequence and the number of construction paths presented by the construction path corresponding to the risky construction path information is 70÷120≈0.58.

[0100] The server analyzes the deviation type between the on-site construction process image and the risky construction path information based on the first ratio information 0.5, the second ratio information 0.58, the number of on-site construction paths 80, and the number of construction paths 120.

[0101] Since the first proportion information is smaller than the second proportion information, it means that the proportion of on-site construction in accordance with the standard construction step sequence is relatively low. Further analysis shows that the number of on-site construction paths is less than the number of construction paths in the risk construction path information, and there may be omissions or simplifications of construction steps. Combining these factors, the server determines that the deviation type of this construction link is the sequence deviation caused by the simplification of construction steps.

[0102] Looking at the pipeline installation construction, there are 60 on-site construction paths, of which 35 are in a specific construction step sequence, and the first ratio is 35÷60≈0.58. There are 90 construction paths in the risky construction path information, of which 50 are in a standard construction step sequence, and the second ratio is 50÷90≈0.56.

[0103] In this case, the first ratio is slightly higher than the second ratio, but the number of on-site construction paths is less than the number of risky construction path information. The server comprehensively judges that there may be deviations in the selection of construction paths in this construction link, resulting in some construction steps that should be carried out not being included. The deviation type is the sequence deviation caused by the deviation in the selection of construction paths.

[0104] For indoor decoration construction, there are 70 on-site construction paths, 40 of which follow a specific construction step sequence, and the first ratio is 40÷70≈0.57. There are 100 construction paths in the risky construction path information, 60 of which follow a standard construction step sequence, and the second ratio is 60÷100=0.6.

[0105] Through analysis, the server found that the proportion of on-site construction in accordance with the standard steps was lower than that in the risky construction path, and the total number of on-site construction paths was also small. This may be due to the avoidance of some complex construction steps during the construction process, resulting in the deviation type being the sequence deviation caused by the avoidance of construction steps.

[0106] In the installation of fire protection facilities, there are 50 on-site construction paths, 25 of which follow a specific construction step sequence, and the first ratio is 25÷50=0.5. There are 80 construction paths in the risky construction path information, 45 of which follow a standard construction step sequence, and the second ratio is 45÷80=0.56.

[0107] Based on this, the server judged that there were deficiencies in the completeness of the on-site construction step execution, and the deviation type was sequence deviation caused by missing construction steps.

[0108] In the door and window installation construction phase, there are 40 on-site construction paths, 20 of which follow a specific construction step sequence, and the first ratio is 20÷40=0.5. There are 60 construction paths in the risky construction path information, 35 of which follow a standard construction step sequence, and the second ratio is 35÷60≈0.58.

[0109] Analysis shows that the proportion of on-site construction that follows standard steps is not only low, but also the total number of construction paths is small. This may be because the construction personnel did not strictly follow the standard process during the operation, and the deviation type was sequence deviation caused by non-standard construction operations.

[0110] In the installation and construction of the air-conditioning system, there are 90 on-site construction paths, 50 of which follow a specific construction step sequence, and the first ratio is 50÷90≈0.56. There are 130 construction paths in the risky construction path information, 75 of which follow a standard construction step sequence, and the second ratio is 75÷130≈0.58.

[0111] The server comprehensively judged that some construction steps were incomplete during the on-site construction, and the deviation type was sequence deviation caused by partial missing construction steps.

[0112] In the elevator installation construction phase, there are 70 on-site construction paths, 38 of which follow a specific construction step sequence, and the first ratio is 38÷70≈0.54. There are 100 construction paths in the risk construction path information, 60 of which follow a standard construction step sequence, and the second ratio is 60÷100=0.6.

[0113] After analysis, it was found that there were gaps between the on-site construction and the risky construction path information in terms of step execution and the number of paths, and the deviation type was sequence deviation caused by comprehensive construction deviation.

[0114] In a possible implementation, step S130 includes:

[0115] Step S131, obtaining a first adjustment coefficient corresponding to the construction step sequence deviation and a second adjustment coefficient corresponding to the construction deviation type information. The first adjustment coefficient is used to adjust the importance coefficient corresponding to the construction step sequence deviation, and the second adjustment coefficient is used to adjust the importance coefficient corresponding to the construction deviation type information.

[0116] Step S132: generating a target construction step sequence deviation based on the construction step sequence deviation and the first adjustment coefficient, and generating target construction deviation type information based on the construction deviation type information and the second adjustment coefficient.

[0117] Step S133, based on the construction path deviation information, the target construction step sequence deviation and the target construction deviation type information, the construction safety confidence of matching the on-site construction process image to the risky construction path information is analyzed to generate the construction safety confidence array.

[0118] In this embodiment, in a large industrial park construction project, a large number of surveillance cameras are deployed at the construction site as target construction monitoring objects. The server continuously obtains the on-site construction process images taken by these monitoring objects, and determines the risk construction path information within the target construction area from the construction drawings.

[0119] First, the server obtains the first adjustment coefficient corresponding to the construction step sequence deviation and the second adjustment coefficient corresponding to the construction deviation type information. Assume that in this project, based on previous experience and data analysis, the first adjustment coefficient is determined to be 0.2, which is used to indicate the relative importance of the construction step sequence deviation; the second adjustment coefficient is 0.8, which is used to emphasize the importance of the construction deviation type information.

[0120] Next, the server generates a target construction step sequence deviation based on the obtained construction step sequence deviation and the first adjustment coefficient. For example, for the steel beam welding construction link, the previously calculated construction step sequence deviation is 0.3. Multiplying it by the first adjustment coefficient 0.2, the target construction step sequence deviation is 0.06.

[0121] Similarly, for the construction deviation type information, it is assumed that its value is 0.4, which is multiplied by the second adjustment coefficient 0.8 to generate the target construction deviation type information of 0.32.

[0122] Then, based on the construction path deviation information, the target construction step sequence deviation and the target construction deviation type information, the server parses the construction safety confidence of matching the on-site construction process image to the risky construction path information to generate a construction safety confidence array.

[0123] Taking electrical wiring construction as an example, the previously calculated construction path deviation information is 0.25. Combined with the previously generated target construction step sequence deviation of 0.06 and target construction deviation type information of 0.32, the server comprehensively considers these data.

[0124] The construction path deviation information indicates that there is a certain degree of deviation between the on-site construction path and the standard risk construction path in terms of route direction, turning points, etc. The target construction step sequence deviation is small, indicating that the sequence deviation of the electrical wiring construction steps is relatively insignificant. However, the target construction deviation type information is large, which may mean that there are more serious deviation types such as unreasonable construction path selection or non-standard construction operations.

[0125] The server generates a construction safety confidence array through a specific algorithm and weight distribution. Assume that the weight of the construction path deviation information in the algorithm is 0.3, the weight of the target construction step sequence deviation is 0.2, and the weight of the target construction deviation type information is 0.5.

[0126] The construction safety confidence of the electrical wiring construction link is calculated as: 0.25×0.3+0.06×0.2+0.32×0.5=0.075+0.012+0.16=0.247

[0127] For the concrete pouring construction phase, the construction path deviation information is 0.18, the target construction step sequence deviation is 0.08, and the target construction deviation type information is 0.25.

[0128] Calculated according to the same weight distribution, the construction safety confidence is: 0.18×0.3+0.08×0.2+0.25×0.5=0.054+0.016+0.125=0.195

[0129] In the pipeline installation construction, the construction path deviation information is 0.22, the target construction step sequence deviation is 0.05, and the target construction deviation type information is 0.3.

[0130] The calculated construction safety confidence is: 0.22×0.3+0.05×0.2+0.3×0.5=0.066+0.01+0.15=0.226

[0131] For the installation and construction of fire protection facilities, the construction path deviation information is 0.15, the target construction step sequence deviation is 0.07, and the target construction deviation type information is 0.28.

[0132] The construction safety confidence level is: 0.15×0.3+0.07×0.2+0.28×0.5=0.045+0.014+0.14=0.2

[0133] In the door and window installation construction phase, the construction path deviation information is 0.12, the target construction step sequence deviation is 0.04, and the target construction deviation type information is 0.2.

[0134] The construction safety confidence level is: 0.12×0.3+0.04×0.2+0.2×0.5=0.036+0.008+0.1=0.144

[0135] In the installation and construction of the air-conditioning system, the construction path deviation information is 0.19, the target construction step sequence deviation is 0.06, and the target construction deviation type information is 0.26.

[0136] The construction safety confidence level is: 0.19×0.3+0.06×0.2+0.26×0.5=0.057+0.012+0.13=0.2

[0137] In the elevator installation construction phase, the construction path deviation information is 0.2, the target construction step sequence deviation is 0.05, and the target construction deviation type information is 0.35.

[0138] The construction safety confidence level is: 0.2×0.3+0.05×0.2+0.35×0.5=0.06+0.01+0.175=0.245

[0139] Through the above calculations, the server generates an array containing the construction safety confidence of each construction link. Each value in this array reflects the matching degree between the corresponding construction link and the risky construction path information and the reliability of construction safety. The higher the value, the closer the construction link is to the standard risky construction path, and the construction safety risk is relatively low; the lower the value, the greater the deviation of the construction link, and the higher the construction safety risk.

[0140] The server can monitor and manage the construction process more specifically based on this construction safety confidence array. For links with low construction safety confidence, such as door and window installation, early warnings are issued in a timely manner, and special personnel are arranged to conduct on-site inspections and guidance to ensure that the construction is carried out in accordance with the specifications, reduce construction risks, and ensure project quality and safety. For links with high construction safety confidence, such as electrical wiring construction, a certain monitoring frequency is still maintained to prevent unexpected situations. At the same time, the entire construction safety confidence array is regularly updated and re-evaluated, and the construction management strategy is continuously optimized as the construction progresses and data accumulates.

[0141] In the subsequent construction process, suppose a new construction link appears, such as the installation of a ventilation system. The server obtains the on-site construction process image and risky construction path information in the same way, calculates the construction feature deviation information and construction deviation type information, and then generates the construction safety confidence and incorporates it into the array.

[0142] For the ventilation system installation construction, it is assumed that the construction path deviation information is 0.16, the target construction step sequence deviation is 0.03, and the target construction deviation type information is 0.22.

[0143] The construction safety confidence level calculated according to the established weights is: 0.16×0.3+0.03×0.2+0.22×0.5=0.048+0.006+0.11=0.164

[0144] Adding this new level of construction safety confidence to the array continuously improves safety assessment and monitoring of the entire construction site.

[0145] In addition, during the construction process, if the conditions of certain construction links change, such as changes in the construction environment, adjustments to construction personnel, or updates to construction equipment, the server will recalculate the relevant data of these links and update the construction safety confidence array.

[0146] For example, due to sudden changes in weather, the conditions for concrete pouring construction are affected. The server re-acquires the on-site construction process images and calculates that the construction path deviation information becomes 0.2, the target construction step sequence deviation becomes 0.09, and the target construction deviation type information becomes 0.3.

[0147] Recalculate the construction safety confidence level as follows: 0.2×0.3+0.09×0.2+0.3×0.5=0.06+0.018+0.15=0.228

[0148] The value of the concrete pouring construction link in the construction safety confidence array is updated to more accurately reflect the current construction safety status.

[0149] By continuously acquiring data, calculating and analyzing, and updating the construction safety confidence array, the server can provide real-time, accurate, and comprehensive construction safety assessments for the entire industrial park construction project, ensuring the smooth and safe progress of the project.

[0150] As the project progresses, some construction links may enter the final stage. Taking electrical wiring construction as an example, when this link is close to completion, the server obtains relevant data again. At this time, the construction path deviation information is reduced to 0.1, the target construction step sequence deviation is 0.02, and the target construction deviation type information is 0.15.

[0151] Recalculate the construction safety confidence level as follows: 0.1×0.3+0.02×0.2+0.15×0.5=0.03+0.004+0.075=0.109

[0152] This shows that the construction safety status of the electrical wiring construction in the final stage was good and in line with expectations.

[0153] For the pipeline installation construction that is still in progress, due to some difficulties encountered during the construction process, the construction path deviation information increased to 0.25, the target construction step sequence deviation was 0.06, and the target construction deviation type information was 0.35.

[0154] Recalculate the construction safety confidence level as follows: 0.25×0.3+0.06×0.2+0.35×0.5=0.075+0.012+0.175=0.262

[0155] Based on this change, the server will increase monitoring and management of pipeline installation and construction, promptly resolve problems that arise, and ensure construction safety and quality.

[0156] In the entire project, different construction links may have mutual influences. For example, the progress of fire protection facility installation construction may be affected by the progress and quality of previous electrical wiring construction. The server will fully consider this correlation factor when calculating the construction safety confidence of fire protection facility installation construction.

[0157] Assuming that the construction safety confidence of electrical wiring construction is low, it may lead to problems such as poor line connection during the installation and construction of fire protection facilities, thereby increasing the construction path deviation information of the installation and construction of fire protection facilities to 0.2, the target construction step sequence deviation to 0.08, and the target construction deviation type information to 0.3.

[0158] The construction safety confidence is calculated as: 0.2×0.3+0.08×0.2+0.3×0.5=0.06+0.016+0.15=0.226

[0159] That is to say, this embodiment can comprehensively analyze this mutual influence, take corresponding coordination and management measures, and ensure the smooth progress of the entire project. In addition, the construction safety confidence array is not only used for current construction monitoring and management, but also provides valuable experience data for subsequent similar projects. By analyzing and summarizing the construction safety confidence of each construction link in this project, it is possible to optimize the formulation standards of risky construction paths, adjust the value of the adjustment coefficient, improve the accuracy and reliability of construction safety assessment, and provide better reference and reference for future engineering projects.

[0160] Therefore, by obtaining the adjustment coefficient, generating target deviation information, and integrating various data to generate a construction safety confidence array, a comprehensive, dynamic, and accurate safety assessment and management of the construction site is achieved, effectively ensuring the quality and safety of the engineering project.

[0161] In a possible implementation, step S140 includes:

[0162] Step S141, obtaining a preset risk adjustment weight. The preset risk adjustment weight is used to limit the weight of the construction safety confidence array to the correlation array.

[0163] Step S142: constructing a basic correlation array based on the construction safety confidence array and the preset risk adjustment weights.

[0164] Step S143, regularizing the rows in the basic correlation array so that the rows of the basic correlation array match the construction representation vectors corresponding to the risky construction path information, and regularizing the columns in the basic correlation array so that the columns of the basic correlation array match the construction representation vectors corresponding to the on-site construction process images, until the set convergence requirements are met, thereby generating the correlation array.

[0165] Continuing to assume that in a large residential building construction project, the server has successfully generated a construction safety confidence array. The next step is to parse the correlation between the on-site construction process images and the risky construction path information based on this construction safety confidence array to generate a correlation array.

[0166] First, the server obtains the preset risk adjustment weight. In this residential project, the preset risk adjustment weight is set to 0.4 based on the complexity of the project, the experience of the construction team, and relevant regulatory requirements. This preset risk adjustment weight is used to limit the weight of the construction safety confidence array to the correlation array to ensure that various potential risk factors are fully considered when evaluating the correlation.

[0167] Then, the server constructs a basic correlation array based on the construction safety confidence array and the preset risk adjustment weight. Assume that the construction safety confidence array is as follows:

[0168] |Construction stage|Construction safety confidence level|

[0169] |Foundation Construction|0.85|

[0170] |Framework construction|0.78|

[0171] |Wall Masonry|0.65|

[0172] |Water and electricity installation|0.72|

[0173] The server multiplies each construction safety confidence by the preset risk adjustment weight 0.4 to obtain the initial value of the basic correlation array:

[0174] |Construction stage|Foundation correlation value|

[0175] |Foundation Construction|0.34|

[0176] |Framework construction|0.312|

[0177] |Wall masonry|0.26|

[0178] |Water and electricity installation|0.288|

[0179] Next, the server begins to regularize the rows in the basic correlation array. Taking the frame construction as an example, the server obtains the construction representation vector corresponding to the risky construction path information, which includes key features such as the specifications of the frame materials, the construction sequence, and the connection method. Then, the element values ​​of the frame construction row in the basic correlation array are matched and adjusted with these features. If a certain feature deviates greatly from the risky construction path information in actual construction, the correlation value of this row is reduced accordingly; conversely, if the deviation is small, the correlation value is appropriately increased. In this way, the rows of the basic correlation array can accurately reflect the correlation between each feature in the risky construction path information and the on-site construction process image.

[0180] At the same time, the server performs regularization on the columns in the basic correlation array. Taking the foundation construction column as an example, the server obtains the construction representation vector corresponding to the on-site construction process image, such as the depth of the foundation, the degree of soil compaction, the foundation type, etc. Then, the element values ​​of this column are adjusted according to these features. If certain features of the on-site construction process are significantly different from the expected risky construction path, the correlation value of the column is reduced accordingly; if the difference is small, the correlation value is increased. In this way, the columns of the basic correlation array can be matched with the construction representation vectors corresponding to the on-site construction process images.

[0181] During the processing, the server continuously checks and adjusts. For example, when processing the water and electricity installation row, it is found that the initially calculated correlation value is too high, and there are situations in which the line layout in actual construction does not meet the specifications. The server will lower the correlation value of this row to make it more accurately reflect the actual situation. For the wall masonry column, if the arrangement of bricks in the on-site construction deviates from the requirements in the risk construction path, the server will also adjust the correlation value of this column accordingly.

[0182] The server continues to perform such regular processing, constantly comparing and adjusting the element values ​​in the array until the set convergence requirements are met. This convergence requirement may mean that the change in the element values ​​in the array is less than a certain threshold after multiple adjustments, or that the error between the actual construction assessment results and the actual construction assessment results is within an acceptable range.

[0183] Assume that after multiple adjustments and optimizations, the resulting correlation array is as follows:

[0184] |Construction stage|Relevance value|

[0185] |Foundation Construction|0.32|

[0186] |Framework construction|0.28|

[0187] |Wall masonry|0.22|

[0188] |Water and electricity installation|0.25|

[0189] Through this correlation array, construction managers can clearly understand the degree of correlation between each construction link and the risky construction path information. For example, it can be seen from the array that the correlation between the wall masonry link and the risky construction path is relatively low, and there may be more construction deviations, which requires special attention and inspection. However, the correlation of the foundation construction link is relatively high, indicating that its construction process is more consistent with the expected risky construction path.

[0190] During the construction of the entire project, the server will continuously monitor and update the correlation array. As the construction progresses, if some construction links are rectified or improvement measures are taken, the server will re-acquire the relevant data and re-perform the above processing to ensure that the correlation array can always accurately reflect the relationship between on-site construction and risky construction paths.

[0191] For example, during the wall construction phase, rectification was carried out to improve the arrangement of bricks and the use of mortar. After re-evaluation, the server adjusted the correlation values ​​of the wall construction row and column to make it more consistent with the actual construction improvements.

[0192] For example, when fine decoration construction is carried out in the later stage of the project, the server also generates a new construction safety confidence array according to the above steps, and builds and processes the correlation array based on this, providing strong support for the monitoring and management of fine decoration construction.

[0193] In summary, by obtaining preset risk adjustment weights, constructing basic correlation arrays, and performing regular processing of rows and columns, the server can generate accurate and effective correlation arrays, providing important decision-making basis for quality control and risk management of construction projects.

[0194] In a possible implementation, step S150 includes:

[0195] Step S151, determining, based on the correlation array, a target construction path among the construction paths corresponding to the risky construction path information, whose correlation with the on-site construction path corresponding to the on-site construction process image is greater than a threshold value.

[0196] Step S152: determining the risky construction behavior corresponding to the target construction monitoring object in the construction drawing based on the target construction path.

[0197] In this embodiment, the server has generated a correlation array in a previous large-scale residential building construction project.

[0198] First, the server determines, based on the correlation array, a target construction path among the construction paths corresponding to the risky construction path information, whose correlation with the on-site construction path corresponding to the on-site construction process image is greater than a threshold value. Assume that the threshold value is set to 0.6.

[0199] The correlation array is as follows:

[0200] |Construction stage|Relevance value|

[0201] |Foundation Construction|0.7|

[0202] |Framework construction|0.5|

[0203] |Wall masonry|0.8|

[0204] |Water and electricity installation|0.4|

[0205] In this array, the correlation value of foundation construction is 0.7, which is greater than the threshold value of 0.6; the correlation value of wall masonry is 0.8, which is also greater than the threshold value of 0.6. Therefore, the construction paths corresponding to foundation construction and wall masonry are determined as target construction paths.

[0206] Next, the server determines the risky construction behaviors corresponding to the target construction monitoring objects in the construction drawings based on these target construction paths.

[0207] For foundation construction, the server refers to the risky construction behavior descriptions in the construction drawings and historical work tickets. The construction drawings stipulate that there are clear standards for the depth, width, and bearing capacity of the foundation. By comparing the on-site construction process images and the risky construction path information, it was found that the depth of one side of the foundation was shallower than the specified value, which was determined to be a risky construction behavior.

[0208] In terms of wall construction, the construction drawings have strict requirements on the arrangement of bricks, the thickness of mortar and the verticality of the wall. After analyzing the on-site construction process images and related data, it was found that the bricks of some walls were not arranged neatly, the mortar thickness was uneven, and the verticality of several walls exceeded the allowable range. These were all identified as risky construction behaviors.

[0209] For example, in the on-site construction process image of foundation construction, the server can accurately measure the specific location and degree of deviation of inconsistent foundation depth through measurement tools and image analysis technology. At the same time, it can also observe whether the soil compaction around the foundation meets the requirements.

[0210] For on-site construction images of wall masonry, the server can clearly see the size of the gaps between bricks and the arrangement pattern. By comparing with the standard arrangement in the construction drawings, the parts that do not meet the requirements are determined. For the thickness of the mortar, the grayscale analysis of the image or a specific measurement algorithm is used to determine its uniformity. The verticality of the wall is determined by laser measurement or line detection technology in image processing to determine the deviation.

[0211] The server records these identified risky construction behaviors in detail, including specific locations, degree of deviation, and relevant image evidence, which provides an important basis for subsequent construction rectification and quality assessment.

[0212] In the subsequent construction of the project, such as the roof construction phase, the server will perform the above analysis again. Assuming that the correlation value of the roof construction is greater than the threshold value, through further comparison, it is found that the laying of the roof waterproofing membrane has local wrinkles and weak bonding, which is determined to be a risky construction behavior.

[0213] During the door and window installation stage, the server found that the installation positions of some windows deviated from the construction drawings and the sealant was applied unevenly, which were also identified as risky construction behaviors that needed to be improved.

[0214] The server continuously monitors and analyzes each link in the entire construction process, promptly discovers and records risky construction behaviors, and ensures the quality and safety of residential building construction.

[0215] In a possible implementation, after step S110, this embodiment may also perform feature extraction on the on-site construction process image to generate a construction representation vector corresponding to the on-site construction process image, and perform feature extraction on the risky construction path information to generate a construction representation vector corresponding to the risky construction path information.

[0216] Step S120 includes:

[0217] The characteristic deviation between the construction representation vector corresponding to the on-site construction process image and the construction representation vector corresponding to the risky construction path information is analyzed to generate construction characteristic deviation information.

[0218] The deviation type between the construction representation vector corresponding to the on-site construction process image and the construction representation vector corresponding to the risky construction path information is analyzed to generate construction deviation type information.

[0219] In this embodiment, it is further assumed that in a large commercial complex construction project, multiple high-definition surveillance cameras are deployed at the construction site as target construction monitoring objects. The server begins to execute operations of acquiring on-site construction process images acquired by the target construction monitoring object during dynamic monitoring, and acquiring risk construction path information within the target construction partition in the construction drawing.

[0220] The server first obtains a set of on-site construction process images, one of which shows a group of workers pouring concrete on the building frame. From this on-site construction process image, the server extracts features such as the worker's pouring action flow, the concrete discharge position, and the pouring layer sequence, quantifies and encodes these features, and generates a construction representation vector corresponding to the on-site construction process image. For example, the worker's pouring action flow is encoded as a series of digitally represented action steps, the concrete discharge position is recorded in coordinate form, and the pouring layer sequence is represented by a specific sequence code.

[0221] At the same time, the server obtains the risky construction path information within the target construction zone in the construction drawings. For this concrete pouring zone, the server extracts the standard concrete pouring speed range, pouring starting position and direction, and pouring order of the layers from the construction planning information of the historical work ticket, and generates a construction representation vector corresponding to the risky construction path information. For example, the specified pouring speed range is encoded as a numerical interval, the pouring starting position and direction are represented by coordinates and angles, and the pouring order of the layers is represented by a specific coding sequence.

[0222] Next, the server analyzes the characteristic deviation between the construction representation vector corresponding to the on-site construction process image and the construction representation vector corresponding to the risky construction path information to generate construction characteristic deviation information.

[0223] The server compares the two construction characterization vectors in detail. Assume that in the on-site construction, the actual pouring speed of concrete is 18 cubic meters per hour, while the speed range specified in the risky construction path information is 10 to 15 cubic meters per hour. The server calculates the degree of deviation between the actual speed and the specified speed range to obtain the characteristic deviation of speed. For example, the starting position of pouring in the on-site construction deviates by 2 meters in the horizontal direction from the specified starting position. The server determines the characteristic deviation of position by calculating the value of the position deviation. Combining these different aspects of characteristic deviation, the server generates comprehensive construction characteristic deviation information, which clearly reflects the deviation of the current construction from the risky construction path in various key features.

[0224] Then, the server analyzes the deviation type between the construction representation vector corresponding to the on-site construction process image and the construction representation vector corresponding to the risky construction path information, and generates construction deviation type information.

[0225] The server deeply analyzes each dimension of the two construction representation vectors. For example, the pouring order of the concrete layer in the on-site construction is to pour it from the top layer downwards, while the order specified in the risk construction path information is to pour it from the bottom layer upwards. Through this detailed comparison, the server determines that the deviation type is an error in the pouring layer order. For another example, the discharge position of the concrete in the on-site construction is too close to one side of the building compared to the specified position, and the server determines that the deviation type is a discharge position deviation. Through accurate judgment of these specific deviation situations, the server generates detailed construction deviation type information, which provides an accurate basis for subsequent construction safety assessment and adjustment.

[0226] In another scenario, assume that the target construction monitoring object is the laying of electrical lines inside a building. The on-site construction process images obtained by the server show the path of the workers laying the lines, the fixing method of the wire trough, the connection method of the wire ends, etc. The server converts these features into construction representation vectors corresponding to the on-site construction process images. For example, the line path is represented by a series of coordinate points, the fixing method of the wire trough is represented by a specific code, and the connection method of the wire ends is represented by a digital sequence.

[0227] At the same time, features such as standard line laying path planning, standard fixing requirements of cable troughs, and correct connection process of wire ends are extracted from the risk construction path information of the construction drawings to generate a construction representation vector corresponding to the risk construction path information.

[0228] When analyzing the characteristic deviation, the server found that the line laying path in the on-site construction deviated from the standard plan at several turning points. By calculating the coordinate deviation values ​​of these turning points, the characteristic deviation of the path was obtained. For the fixing of the cable duct, the number of fixings used in the on-site construction was less than the specified number. The server determined the characteristic deviation of the fixing method by calculating the difference in the number of fixings.

[0229] When judging the deviation type, if the wire connection method in the on-site construction is different from the standard process, the server determines it as a deviation type of incorrect wire connection method. If the fixed position of the wire trough does not meet the requirements of the specification, it is determined as a deviation type of wire trough fixed position deviation.

[0230] In this way, it is possible to accurately and comprehensively extract features, analyze feature deviations, and judge deviation types of on-site construction process images and risky construction path information, providing strong support and guarantee for ensuring the quality and safety of construction.

[0231] In a possible implementation manner, the on-site construction process image includes an on-site construction behavior scale, an on-site construction path state, and an on-site construction path category label, and the feature extraction of the on-site construction process image to generate a construction representation vector corresponding to the on-site construction process image includes:

[0232] Feature extraction is performed on the on-site construction path state to generate an on-site construction path state representation vector.

[0233] Feature extraction is performed on the on-site construction path category label to generate an on-site construction path category label representation vector.

[0234] Feature extraction is performed on the on-site construction behavior scale to constrain the on-site construction behavior scale within a first set interval, and a on-site construction behavior scale representation vector is generated.

[0235] A construction representation vector corresponding to the on-site construction process image is generated based on the on-site construction path state representation vector, the on-site construction path category label representation vector and the on-site construction behavior scale representation vector.

[0236] Let's continue to assume that in a large residential construction project, the server is monitoring and analyzing the construction site.

[0237] The on-site construction process images obtained by the server contain rich information. For example, one of the images shows workers constructing a wall.

[0238] For the status of the on-site construction path, the server performs feature extraction. The server observes the placement of bricks, the uniformity of mortar application, and the progress of wall construction. Through quantitative analysis of these states, such as dividing the neatness of bricks into several levels, expressing the uniformity of mortar application with a specific numerical range, and measuring the progress of wall construction by the completion ratio, a vector representing the status of the on-site construction path is generated. For example, the neatness of bricks is rated 8 points (out of 10 points), the uniformity of mortar application is between 70% and 85%, and the progress of wall construction is 30%. These data are encoded into a specific vector to represent the status of the on-site construction path.

[0239] For the on-site construction path category label, the server extracts that this is a category label belonging to wall masonry construction. By encoding different construction category labels, such as wall masonry is encoded as "01", a on-site construction path category label representation vector is generated.

[0240] For the on-site construction behavior scale, the server notes the number of bricks carried by the worker each time, the amplitude of the bricklaying movement, and the duration of the work. In order to constrain these scales within the first set interval, for example, the reasonable number of bricks carried each time is between 5 and 10, the amplitude of the bricklaying movement is within a certain angle range, and the duration of the work does not exceed a specific duration. If the actual situation exceeds or falls below this interval, make corresponding adjustments and encoding. Assuming that in reality the worker carries 8 bricks each time, the amplitude of the bricklaying movement is within a reasonable range, and the duration of the work is not exceeded, this information is encoded into a vector representing the on-site construction behavior scale.

[0241] Then, the server generates a construction representation vector corresponding to the on-site construction process image based on the on-site construction path state representation vector, on-site construction path category label representation vector and on-site construction behavior scale representation vector generated above. This comprehensive construction representation vector comprehensively describes the characteristics of the on-site construction process of wall masonry.

[0242] In a possible implementation manner, the risky construction path information includes a risky construction behavior scale, a construction path state, and a construction path category label, and the feature extraction of the risky construction path information to generate a construction representation vector corresponding to the risky construction path information includes:

[0243] Feature extraction is performed on the construction path state to generate a construction path state construction representation vector.

[0244] Feature extraction is performed on the construction path category label to generate a construction path category label construction representation vector.

[0245] Feature extraction is performed on the risky construction behavior scale to constrain the risky construction behavior scale within a second set interval, and a risky construction behavior scale characterization vector is generated.

[0246] Based on the construction path status construction representation vector, the construction path category label construction representation vector and the risky construction behavior scale representation vector, a construction representation vector corresponding to the risky construction path information is generated.

[0247] In this embodiment, the risk construction path information obtained by the server also includes multiple aspects.

[0248] For the construction path status, the server analyzes the ideal state of wall construction, such as bricks should be completely aligned, mortar should be applied evenly, and the wall construction progress should be in line with the plan. Through a similar quantitative method, a construction path status construction representation vector is generated.

[0249] The construction path category labels are also clearly coded. For example, the risk construction category label of wall masonry is also coded as "01" to generate a construction path category label construction representation vector.

[0250] For the risk construction behavior scale, the server sets a standard scale range. For example, the ideal number of bricks to be moved each time is 6 to 8, the standard amplitude of bricklaying actions, and the reasonable duration of work. The actual risk construction behavior scale is compared and adjusted with this standard range to generate a risk construction behavior scale representation vector.

[0251] Finally, the server generates a construction representation vector corresponding to the risky construction path information based on the construction path status construction representation vector, the construction path category label construction representation vector and the risky construction behavior scale representation vector.

[0252] In another scenario, suppose the on-site construction is floor laying. In the on-site construction process image obtained by the server, for the on-site construction path status, the server observes the splicing tightness, flatness, and paving direction consistency of the floor material. Through precise measurement and evaluation, such as the width of the splicing gap between 1 and 2 mm is rated as good, the flatness deviation is within 2 mm and is considered qualified, and the angle deviation between the laying direction and the design direction does not exceed 5 degrees, etc., a vector representing the on-site construction path status is generated.

[0253] The on-site construction path category label is clearly defined as floor laying construction and encoded as a specific identifier to form a on-site construction path category label representation vector.

[0254] For the on-site construction behavior scale, the server focuses on the speed at which workers lay the floor, the amplitude of each movement when picking up the floor, and the frequency of tool changes. A reasonable range is set, such as a laying speed of 2 to 3 square meters per minute, an amplitude of movement within a certain angle, and tool changes not exceeding a certain number of times. Encoding is performed based on actual conditions to generate a vector representing the on-site construction behavior scale.

[0255] By integrating these three aspects, a construction representation vector corresponding to the on-site construction process image is generated to comprehensively describe the on-site construction characteristics of floor laying.

[0256] In terms of risk construction path information, for the construction path status, the ideal floor splicing should be seamless, the flatness deviation should be extremely small, and the laying direction should be completely accurate. Through strict standard quantification, the construction path status construction characterization vector is generated.

[0257] The construction path category label is also coded as an identifier consistent with the on-site construction.

[0258] The risk construction behavior scale is set to the optimal range, such as a laying speed of about 2.5 square meters per minute, precise control of the movement amplitude, and very few tool changes. A risk construction behavior scale characterization vector is generated.

[0259] Finally, a construction representation vector corresponding to the risky construction path information is generated for comparison and analysis with the representation vector of on-site construction.

[0260] In the scenario of roof waterproofing construction, the server observes the uniformity of the laying of waterproof materials, the overlap width of the rolls, and the sealing conditions from the on-site construction process images and the on-site construction path status. For example, a laying uniformity of more than 90% is excellent, a roll overlap width of 8 to 10 cm is acceptable, and the sealing is tight without loopholes, etc., and generates corresponding characterization vectors.

[0261] The category label of the on-site construction path is determined to be roof waterproofing construction, and a representation vector is formed after encoding processing.

[0262] In terms of on-site construction behavior scale, attention is paid to the speed of workers applying waterproof materials, the proficiency of laying rolls, and the continuity of work. Reasonable speed ranges, action standards, and working time limits are set, and encoding is performed to generate representation vectors.

[0263] The construction representation vector corresponding to the on-site construction process image is obtained comprehensively.

[0264] In the risk construction path information, for the construction path status, it is expected that the waterproof material is laid completely evenly, the roll material overlaps accurately, and the sealing is flawless. These standards are quantified to generate the construction path status construction representation vector.

[0265] The construction path category labels and risky construction behavior scales are processed similarly to generate a construction representation vector corresponding to the risky construction path information.

[0266] Through the above multiple detailed scenario examples, the server can comprehensively and accurately extract features and generate representation vectors for on-site construction process images and risky construction path information, providing a solid data foundation for subsequent construction monitoring and analysis, helping to promptly discover problems and risks in construction and ensure the quality and safety of construction.

[0267] In a possible implementation, step S110 includes:

[0268] Step S111, obtaining a positioning coordinate area determined after the monitoring range is aligned with the coordinate system of the construction drawing.

[0269] Step S112, determining the peripheral partitions located in the positioning coordinate area in the construction drawing, and generating the target construction partitions.

[0270] Step S113, obtaining the risky construction path information from the construction planning information of the historical work tickets in the target construction zone.

[0271] In this embodiment, the following is a detailed scenario description based on the above examples:

[0272] In this large-scale construction project, the server began to obtain risky construction path information within the target construction zone in the construction drawings.

[0273] The server first obtains the positioning coordinate area determined by aligning the monitoring range with the coordinate system of the construction drawing. Assume that the coordinate area corresponding to the monitoring range of one of the surveillance cameras is (X1=100, Y1=200) to (X2=150, Y2=250). The server seamlessly aligns the coordinates of this monitoring range with the coordinate system of the construction drawing through rigorous algorithms and precise calculations.

[0274] Then, the server determines the surrounding partitions of this positioning coordinate area in the construction drawings and generates the target construction partition. After careful comparison and analysis, the server finds that the surrounding area of ​​this coordinate area is an area where concrete pouring and electrical wiring are being carried out. Taking into account the relevance and potential risks of the construction, the server sets this area as the target construction partition.

[0275] Next, the server obtains risky construction path information from the construction planning information of historical work tickets within the target construction zone. In terms of concrete pouring, the server learned from historical work tickets that if the pouring speed exceeds 30 cubic meters per hour, or the pouring sequence is not gradually advanced from one end, it may cause uneven concrete structure. The corresponding construction path features include that the starting point of pouring is located at the coordinates (30,40), the pouring direction should be from left to right, and the allowable pouring speed range is 20 to 30 cubic meters per hour. The server accurately extracts these specific feature information as risky construction path information for concrete pouring.

[0276] In terms of electrical wiring, the historical work ticket pointed out that if the line direction deviates from the design drawing by more than 10 cm, or if it does not maintain a sufficient safety distance when crossing other pipelines, there will be safety hazards. The relevant construction path features include that the layout of the line should be along a specific position on the wall, and the turning point should be set 50 cm away from the corner of the wall. The server carefully extracts and organizes this key information to form risk construction path information for electrical wiring.

[0277] In this way, comprehensive and accurate risk construction path information within the target construction area was obtained, which made full preparations for subsequent construction monitoring and analysis, and was able to timely discover and prevent possible construction risks, ensuring the smooth progress and quality safety of construction.

[0278] Figure 2 In an embodiment of the present application, a construction site safety monitoring system 100 based on image analysis is provided, including a processor 1001 and a memory 1003 and a sensor node code stored in the memory 1003. The processor 1001 executes the above-mentioned sensor node code to implement the steps of the sensor data processing method applied to smart photovoltaics.

[0279] Figure 2 The construction site safety monitoring system 100 based on image analysis shown includes: a processor 1001 and a memory 1003. Among them, the processor 1001 and the memory 1003 are connected, such as through a bus 1002. Optionally, the construction site safety monitoring system 100 based on image analysis may also include a transceiver 1004, and the transceiver 1004 can be used for data interaction between the server and other servers, such as data sending and / or data receiving. It should be noted that in actual scheduling, the transceiver 1004 is not limited to one, and the structure of the construction site safety monitoring system 100 based on image analysis does not constitute a limitation on the embodiments of the present application.

[0280] Processor 1001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of this application. Processor 1001 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0281] The bus 1002 may include a path to transmit information between the above components. The bus 1002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus 1002 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 2 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0282] The memory 1003 may be a ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, a RAM (Random Access Memory) or other types of dynamic storage devices that can store information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, optical disk storage (including compressed optical disk, laser disk, optical disk, digital versatile disk, Blu-ray disk, etc.), magnetic disk storage medium, other magnetic storage devices, or any other medium that can be used to carry or store sensor node codes and can be read by a computer, without limitation herein.

[0283] The memory 1003 is used to store the sensor node code for executing the embodiment of the present application, and the execution is controlled by the processor 1001. The processor 1001 is used to execute the sensor node code stored in the memory 1003 to implement the steps shown in the above method embodiment.

[0284] The server includes but is not limited to: mobile terminals such as mobile phones, notebook computers, PADs, etc. and fixed terminals such as digital TVs, desktop computers, etc.

[0285] An embodiment of the present application provides a computer-readable storage medium, on which a sensor node code is stored. When the sensor node code is executed by a processor, the steps and corresponding contents of the aforementioned method embodiment can be implemented.

[0286] It should be understood that, although each operation step is indicated by arrows in the flowchart of the embodiment of the present application, the implementation order of these steps is not limited to the order indicated by the arrows. Unless there is clear explanation in this article, in some implementation scenarios of the embodiment of the present application, the implementation steps in each flowchart can be performed in other orders according to demand. In addition, some or all of the steps in each flowchart may include multiple sub-steps or multiple stages according to actual implementation scenarios, and some or all of these sub-steps or stages may be executed at the same time, and each sub-step or stage in these sub-steps or stages may also be executed at different times respectively. In different scenarios at the execution time, the execution order of these sub-steps or stages can be flexibly configured based on demand, and the embodiment of the present application does not limit this.

[0287] The above is only an optional implementation method for some implementation scenarios of the present application. It should be pointed out that for ordinary technicians in this technical field, without departing from the technical concept of the scheme of the present application, other similar implementation methods based on the technical ideas of the present application are also within the protection scope of the embodiments of the present application.

Claims

1. A construction site safety monitoring method based on image analysis, characterized in that: The method includes: Acquire the on-site construction process images acquired by the target construction monitoring object during the dynamic monitoring process, and acquire the risk construction path information within the target construction partition in the construction drawing; the target construction partition is determined after the monitoring range of the target construction monitoring object is aligned with the coordinate system of the construction drawing, and the risk construction path information is information constraining the path characteristics of the construction path of the risk construction behavior of the historical work ticket within the target construction partition; Analyzing the characteristic deviation between the on-site construction process image and the risky construction path information to generate construction characteristic deviation information, and analyzing the deviation type between the on-site construction process image and the risky construction path information to generate construction deviation type information; Based on the construction feature deviation information and the construction deviation type information, the construction safety confidence between the on-site construction process image and the risky construction path information is analyzed to generate a construction safety confidence array; Analyzing the correlation between the on-site construction process image and the risky construction path information according to the construction safety confidence array to generate a correlation array; Determine, based on the correlation array, the risky construction behavior corresponding to the target construction monitoring object in the construction drawing; The step of analyzing the characteristic deviation between the on-site construction process image and the risky construction path information to generate construction characteristic deviation information includes: Determine a cosine distance between a construction representation vector corresponding to the on-site construction process image and a construction representation vector corresponding to the risky construction path information; Based on the construction representation vector corresponding to the on-site construction process image, the construction representation vector corresponding to the risky construction path information, the construction accuracy confidence of the on-site construction path corresponding to the on-site construction process image, and the cosine distance, the construction path difference between the on-site construction process image and the risky construction path information is analyzed to generate construction path deviation information; Analyzing the deviation between the construction step sequence presented by the on-site construction path corresponding to the on-site construction process image and the construction step sequence presented by the construction path corresponding to the risky construction path information, and generating a construction step sequence deviation degree; The construction path deviation information and the construction step sequence deviation are determined as the construction feature deviation information.

2. The construction site safety monitoring method based on image analysis according to claim 1 is characterized in that: The on-site construction process image includes the number of on-site construction paths, and the risky construction path information includes the number of construction paths; the deviation between the construction step sequence presented by the on-site construction path corresponding to the on-site construction process image and the construction step sequence presented by the construction path corresponding to the risky construction path information is analyzed to generate the construction step sequence deviation, including: Determine first ratio information between the number of construction paths corresponding to the sequence of construction steps presented by the on-site construction path and the number of on-site construction paths; determine second ratio information between the number of construction paths corresponding to the sequence of construction steps presented by the construction path and the number of construction paths; A construction step sequence deviation analysis is performed based on the first proportion information and the second proportion information to generate the construction step sequence deviation degree.

3. The construction site safety monitoring method based on image analysis according to claim 1 is characterized in that: The on-site construction process image includes the number of on-site construction paths, and the risky construction path information includes the number of construction paths; the parsing of the deviation type between the on-site construction process image and the risky construction path information to generate the construction deviation type information includes: Determine first ratio information between the sequence of construction steps presented by the on-site construction path corresponding to the on-site construction process image and the number of on-site construction paths; determine second ratio information between the sequence of construction steps presented by the construction path corresponding to the risky construction path information and the number of construction paths; Based on the first proportion information, the second proportion information, the number of on-site construction paths and the number of construction paths, the deviation type between the on-site construction process image and the risky construction path information is analyzed to generate the construction deviation type information.

4. The construction site safety monitoring method based on image analysis according to claim 1 is characterized in that: The construction safety confidence between the on-site construction process image and the risky construction path information is parsed based on the construction feature deviation information and the construction deviation type information to generate a construction safety confidence array, including: Obtaining a first adjustment coefficient corresponding to the construction step sequence deviation and a second adjustment coefficient corresponding to the construction deviation type information; the first adjustment coefficient is used to adjust the importance coefficient corresponding to the construction step sequence deviation, and the second adjustment coefficient is used to adjust the importance coefficient corresponding to the construction deviation type information; generating a target construction step sequence deviation degree based on the construction step sequence deviation degree and the first adjustment coefficient, and generating a target construction deviation type information based on the construction deviation type information and the second adjustment coefficient; Based on the construction path deviation information, the target construction step sequence deviation and the target construction deviation type information, the construction safety confidence of matching the on-site construction process image to the risky construction path information is analyzed to generate the construction safety confidence array.

5. The construction site safety monitoring method based on image analysis according to claim 1 is characterized in that: The step of analyzing the correlation between the on-site construction process image and the risky construction path information based on the construction safety confidence array to generate a correlation array includes: Obtaining a preset risk adjustment weight; the preset risk adjustment weight is used to limit the weight of the construction safety confidence array to the correlation array; Constructing a basic correlation array based on the construction safety confidence array and the preset risk adjustment weights; The rows in the basic correlation array are regularized so that the rows of the basic correlation array match the construction characterization vectors corresponding to the risky construction path information, and the columns in the basic correlation array are regularized so that the columns of the basic correlation array match the construction characterization vectors corresponding to the on-site construction process images, until the set convergence requirements are met, thereby generating the correlation array.

6. The construction site safety monitoring method based on image analysis according to claim 1 is characterized in that: The determining, based on the correlation array, the risky construction behavior corresponding to the target construction monitoring object in the construction drawing includes: Determine, based on the correlation array, a target construction path among the construction paths corresponding to the risky construction path information, whose correlation with the on-site construction path corresponding to the on-site construction process image is greater than a threshold value; The risky construction behavior corresponding to the target construction monitoring object in the construction drawing is determined based on the target construction path.

7. The construction site safety monitoring method based on image analysis according to any one of claims 1 to 6, characterized in that: After acquiring the on-site construction process image of the target construction monitoring object during the dynamic monitoring process, and acquiring the risk construction path information within the target construction partition in the construction drawing, the method further includes: Performing feature extraction on the on-site construction process image to generate a construction representation vector corresponding to the on-site construction process image, and performing feature extraction on the risky construction path information to generate a construction representation vector corresponding to the risky construction path information; The step of analyzing the characteristic deviation between the on-site construction process image and the risky construction path information to generate construction characteristic deviation information includes: Analyzing the characteristic deviation between the construction representation vector corresponding to the on-site construction process image and the construction representation vector corresponding to the risky construction path information to generate construction characteristic deviation information; The step of analyzing the deviation type between the on-site construction process image and the risky construction path information to generate construction deviation type information includes: Analyzing the deviation type between the construction representation vector corresponding to the on-site construction process image and the construction representation vector corresponding to the risky construction path information to generate construction deviation type information; The on-site construction process image includes an on-site construction behavior scale, an on-site construction path state, and an on-site construction path category label. The feature extraction of the on-site construction process image to generate a construction representation vector corresponding to the on-site construction process image includes: Extracting features of the on-site construction path state to generate an on-site construction path state representation vector; Extracting features of the on-site construction path category label to generate a on-site construction path category label representation vector; Extracting features of the on-site construction behavior scale to constrain the on-site construction behavior scale within a first set interval and generate a on-site construction behavior scale representation vector; Generate a construction representation vector corresponding to the on-site construction process image based on the on-site construction path state representation vector, the on-site construction path category label representation vector and the on-site construction behavior scale representation vector; The risky construction path information includes a risky construction behavior scale, a construction path status, and a construction path category label. The feature extraction of the risky construction path information to generate a construction representation vector corresponding to the risky construction path information includes: Extracting features of the construction path state to generate a construction path state construction representation vector; Extracting features of the construction path category label to generate a construction representation vector of the construction path category label; Extracting features of the risky construction behavior scale to constrain the risky construction behavior scale within a second set interval and generate a risky construction behavior scale representation vector; Based on the construction path status construction representation vector, the construction path category label construction representation vector and the risky construction behavior scale representation vector, a construction representation vector corresponding to the risky construction path information is generated.

8. The construction site safety monitoring method based on image analysis according to any one of claims 1 to 6, characterized in that: The step of obtaining risky construction path information within a target construction zone in a construction drawing includes: Acquire a positioning coordinate area determined after the monitoring range is aligned with the coordinate system of the construction drawing; Determine the surrounding partitions located in the positioning coordinate area in the construction drawing to generate the target construction partitions; The risky construction path information is obtained from the construction planning information of the historical work tickets within the target construction zone.

9. A construction site safety monitoring system based on image analysis, characterized in that: The construction site safety monitoring system based on image analysis includes a processor and a machine-readable storage medium, in which machine-executable instructions are stored. The machine-executable instructions are loaded and executed by the processor to implement the construction site safety monitoring method based on image analysis of any one of claims 1-8.

Images