Video monitoring system for ensuring safety during construction

Abstract

The application relates to the technical field of video monitoring, and particularly discloses a video monitoring system for guaranteeing construction process safety, which comprises the following modules: a video recognition module: surface mark points and edge contours are acquired based on multi-angle video streams; a deviation acquisition module: local geometric deviations are acquired according to the surface mark points and the edge contours; a load acquisition module: a load spectrum is extracted based on environmental event records; a simulation module: time-domain dynamic response simulation is carried out based on the local geometric deviations and the load spectrum, and a contact stress time sequence curve is obtained; an analysis module: evolution characteristics of the contact stress time sequence curve are analyzed, and a contact stress peak value space distribution thermal diagram and a peak value reappearance interval sequence are generated; and an output module: a contact stress potential distribution diagram is acquired, and the contact stress potential distribution diagram identifies component connection parts with repeatedly appearing high stress. The application can reduce the possibility of safety problems caused by abnormal stress of components in the construction process.

Description

Technical Field

[0001] This invention relates to the field of video surveillance technology, and more specifically to a video surveillance system for ensuring safety during construction processes. Background Technology

[0002] During the construction of large block structures such as rock breakwaters, components are typically installed and formed gradually under complex marine environments and dynamic loads. The construction phase is significantly affected by external environmental factors such as waves and currents, resulting in marked time-varying and spatial variations in the contact state and stress conditions between components. However, existing construction safety monitoring methods primarily focus on visually recording the surface condition and operational process at the construction site, making it difficult to effectively reflect the component installation quality, changes in contact stress between components, and the potential safety risks these factors may pose under wave action from video monitoring data.

[0003] Especially during construction, due to unavoidable geometric differences such as positional and orientation deviations in component installation, the connection points of components may experience localized stress concentrations under repeated external environmental influences, exhibiting periodic or irregular high-stress states over time. However, current technologies struggle to effectively correlate video monitoring information during construction with the spatial distribution characteristics and temporal evolution of stress on component contacts. This makes it impossible to identify which component connection points may repeatedly experience high stress during construction, hindering the timely detection of potential construction safety hazards and impacting the safety and reliability of the construction process. Summary of the Invention

[0004] The purpose of this invention is to provide a video monitoring system for ensuring the safety of the construction process, thereby solving the aforementioned technical problems.

[0005] The objective of this invention can be achieved through the following technical solutions: The video surveillance system used to ensure the safety of the construction process includes: a video recognition module: acquiring multi-angle video streams of the rock breakwater components and identifying surface markers and edge contours of the components in the multi-angle video streams; Deviation acquisition module: Based on surface markers and edge contours, it acquires the spatial position and attitude angle data of each component, and thereby obtains the local geometric deviations generated during component installation. Load acquisition module: Based on environmental event records of the sea area where the rock breakwater is located, extract the joint probability distribution of wave action direction, wave height and period to form a load spectrum; Simulation module: Local geometric deviation is used as the initial defect input into the parametric structural model of the stone breakwater. The load spectrum is applied as a dynamic boundary condition to the surface of the parametric structural model, and time-domain dynamic response simulation is performed to obtain the contact stress time series curve. Analysis module: Analyzes the evolution characteristics of the contact stress time series curve, extracts the spatial location and time interval patterns of the contact stress peak occurrence, and generates a heat map of the spatial distribution of contact stress peak and a peak recurrence interval sequence; Output module: The thermal map of the spatial distribution of contact stress peak is superimposed with the peak recurrence interval sequence to obtain the potential distribution map of contact stress. The potential distribution map of contact stress identifies the component connection parts where high stress occurs repeatedly.

[0006] As a further aspect of the present invention, the process of determining local geometric deviations is as follows: The theoretical contact surface equation between adjacent components is determined based on the design drawings. The theoretical contact surface equation is defined by a mathematical expression for a plane or curved surface. Using the spatial position and attitude angle data of the component, the three-dimensional design model of the component is transformed into a rigid body. The point set on the surface of the transformed model represents the actual installation form of the component. Dense sampling was performed in the contact area of ​​the actual installation configuration. The signed perpendicular distance from each sampling point to the theoretical contact surface was calculated. The mean and variance of the signed perpendicular distance were statistically analyzed. The mean and variance reflect the overall non-parallelism and local undulation of the contact area, respectively. Calculate the maximum misalignment distance between the actual contour lines of the two components on the theoretical contact surface projection, and use it as the misalignment distance; The overall non-parallelism, the degree of local undulation, and the distance of misalignment together constitute the local geometric deviation.

[0007] As a further aspect of the present invention: the process of obtaining the load spectrum is as follows: Environmental event logs contain historical data sequences of wave action direction, wave height, and period; A joint statistical analysis was performed on the historical data sequences of wave action direction, wave height, and period to establish a joint probability distribution of wave action direction, wave height, and period. In the joint probability distribution, all wave parameter combinations with an occurrence probability higher than a preset probability threshold are extracted. Each wave parameter combination includes a wave action direction, a wave height, and a period. Based on the probability of occurrence of each wave parameter combination, a weight value is assigned to each wave parameter combination. Multiple wave parameter combinations with weight values ​​are weighted and superimposed to form a load spectrum.

[0008] As a further aspect of the present invention, the specific process for obtaining the contact stress time curve is as follows: A parametric structural model of the stone breakwater is established. Each unit in the model corresponds to a component. Units are connected by contact pairs. The mechanical properties of the contact pairs are set according to local geometric deviations. The load spectrum is converted into a time-series wave input, which includes the variation of wave pressure over time and space. Wave input is applied as an external load to the wet surface of the parametric structural model; Numerical calculation methods were used to solve the contact stress time series curve of the parameterized structural model under wave input.

[0009] As a further aspect of the present invention, the specific process for obtaining the contact stress time series curve also includes: Set the simulation time step and total duration, with the total duration covering the complete cycle of the wave input in the load spectrum; At each time step, the distribution of wave pressure on the wetted surface of the model is calculated, and the interaction forces between the contact pairs are also calculated. Based on the geometric area and interaction force of the contact pair, calculate the contact stress of each contact pair at each time step; After the entire simulation period is over, the data on the change of contact stress over time for each contact pair is output, forming a contact stress time series curve.

[0010] As a further aspect of the present invention, the specific process for generating a thermal map of the spatial distribution of peak contact stress is as follows: Traverse all contact stress time series curves and identify the stress peak points on each curve that exceed the preset threshold. Record the spatial location and stress amplitude of each stress peak point; All stress peak points are classified according to their spatial location, and the number or cumulative amplitude of stress peak points in each spatial location area is counted. Based on the statistical results, a corresponding color intensity is assigned to each spatial location region to generate a heat map reflecting the spatial distribution characteristics of the peak contact stress.

[0011] As a further aspect of the present invention, the specific process for generating the peak recurrence interval sequence is as follows: For the same spatial region, all identified stress peak points are sorted according to the order of their occurrence. Calculate the time difference between adjacent stress peak points after sorting to obtain a set of peak recurrence intervals at that location; For each spatial location region, a corresponding set of peak recurrence intervals is generated, and the peak recurrence interval sets of all regions are organized to form a data sequence containing location identifiers and interval values, which serves as the peak recurrence interval sequence.

[0012] As a further aspect of the present invention, the specific process for obtaining the potential distribution map of contact stress is as follows: Use the thermal map of the spatial distribution of peak contact stress as the base layer; The peak recurrence interval sequence is overlaid on the base layer as a graphical feature, and the visual features of the graphical feature are used to express the length of the interval. The base layer and the graphical feature layer are merged and rendered to generate a composite image. In the composite image, regions with high color intensity and short intervals between peak recurrence intervals in the thermal map of contact stress peak spatial distribution are identified by combining visual features. These regions are then marked as component connection points where high stress occurs repeatedly in the potential distribution map of contact stress.

[0013] The beneficial effects of this invention compared to the prior art are as follows: This invention enables more targeted analysis and assessment of the safety status of connection points in rock breakwater components during the construction phase. Through comprehensive analysis of the component installation status and the influence of the external environment during construction, this invention can identify potential areas where high stress may repeatedly occur at component connection points during construction, thereby transforming construction safety monitoring from simply recording on-site conditions to proactive identification of risk characteristics.

[0014] This invention reflects the spatial distribution characteristics of the stress state at component connection points and its temporal variation, enabling construction managers to proactively identify critical areas prone to safety hazards during construction, thus improving the safety and controllability of the construction process. Furthermore, the analytical results obtained by this invention are holistic and continuous, avoiding judgment biases arising from relying solely on single-moment or localized observations, thereby enhancing the reliability of construction safety assessments. This invention facilitates the timely identification of potential risk areas during construction, providing more valuable reference for construction safety management and reducing the likelihood of safety issues arising from abnormal component stress during construction. Attached Figure Description

[0015] The invention will now be further described with reference to the accompanying drawings.

[0016] Figure 1 This is a flowchart illustrating the video surveillance system of the present invention used to ensure safety during construction. Figure 2 This is a schematic diagram of the process for obtaining the contact stress time series curve according to the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Please see Figures 1-2 As shown, the present invention is a video surveillance system for ensuring safety during construction, comprising: Video recognition module: acquires multi-angle video streams of the stone breakwater components and identifies surface markers and edge contours of the components in the multi-angle video streams; Specifically, multiple imaging devices are deployed in the construction area to simultaneously acquire multi-angle video streams of the components from different spatial locations as input. The video streams are then time-aligned and preprocessed, including denoising, brightness equalization, and distortion correction. The component regions are located from each preprocessed frame, and surface markers are identified within these regions. The two-dimensional coordinates of the markers are recorded, and cross-viewpoint correspondences are established. Simultaneously, the component edge contours are extracted as shape boundary data and correlated with multiple viewpoints at the same time point. The set of marker coordinates and the set of edge contour data are output for subsequent pose calculation.

[0019] It should be noted that during the construction phase, multi-angle video streams are used to identify surface markers and edge contours of components, enabling image data from different observation angles to form a set of associative geometric features, and outputting basic data for subsequent pose calculation and deviation analysis.

[0020] Deviation acquisition module: Based on surface markers and edge contours, it acquires the spatial position and attitude angle data of each component, and thereby obtains the local geometric deviations generated during component installation. The process of acquiring spatial position and attitude angle data is as follows: After identifying the surface markers and edge contours of the component in the multi-angle video stream, the two-dimensional image coordinates of the markers and the pixel set of the edge contours corresponding to the same time point of each video stream are input into the pose calculation process. The reference coordinates required for pose calculation adopt the spatial coordinate reference established in advance in the construction area. The spatial coordinate reference is used to describe the three-dimensional positional relationship of the component in the construction site and serves as a unified coordinate framework for the fusion of data from different observation angles. Spatial position and attitude angle data refer to the three-dimensional translational position parameters of the component under the spatial coordinate reference and the angle parameters describing the rotational state of the component around three orthogonal axes. The spatial position is used to characterize the coordinate value of the component reference point relative to the origin of the spatial coordinate reference, and the attitude angle is used to characterize the directional relationship of the component's local coordinate axes relative to the coordinate axes of the spatial coordinate reference.

[0021] To obtain spatial position and attitude angle data, surface markers of the same component under different observation angles are matched across perspectives. The matching is based on the trajectory continuity of the markers in consecutive frames and their geometric consistency under different perspectives, so that the same physical point forms a correspondence in different video streams. After establishing the correspondence, the markers are reconstructed in three dimensions according to the imaging geometry between different observation angles to obtain the set of three-dimensional coordinates of the markers under the spatial coordinate reference. At the same time, the coordinates of the corresponding markers of the component in the design three-dimensional model are used as the input of the model feature point set. By performing rigid body registration on the three-dimensional reconstructed point set and the model feature point set, the rigid body transformation relationship that minimizes the error between the two sets of point sets is obtained. The rigid body transformation relationship includes translation and rotation. The translation is used as the spatial position data of the component, and the rotation is expressed as the attitude angle data of the component through angularization.

[0022] For cases where insufficient marker points or local occlusion leads to decreased point set stability, edge contours are used as auxiliary constraint inputs. The reconstructed edge contours in spatial coordinate reference are registered with the outer boundary of the design 3D model. Contour consistency is used to correct the rigid body registration results, ensuring that the spatial position and attitude angle data remain smooth and continuous over time and consistent with the component's shape. After pose calculation, the spatial position and attitude angle data of each component at each time point are output and associated with the component identifier. The component identifier distinguishes different components and matches the component number in the design drawings, thus providing input for the geometric relationship analysis between adjacent components.

[0023] In a preferred embodiment of the present invention, the process of determining local geometric deviations is as follows: The connection relationship between adjacent components in the design drawings is used as input to read the theoretical contact surface information of adjacent components in the design state. The theoretical contact surface equation is used to describe the geometric surface that the two components should contact in the design state. It is defined by the mathematical expression of the plane or curved surface and established in the design coordinate system. The design coordinate system is a coordinate frame consistent with the design three-dimensional model and is used to express the position of points, lines and surfaces of the component design form.

[0024] After determining the theoretical contact surface equation based on the design drawings, the three-dimensional design model of the component is input into the geometric transformation processing flow. The three-dimensional design model includes the outer surface mesh or point cloud representation of the component and the geometric feature information related to the component. Using the spatial position and attitude angle data obtained above, a rigid body transformation is performed on the three-dimensional design model, so that the three-dimensional design model is mapped from the design coordinate system to the spatial coordinate reference. The point set of the model surface after rigid body transformation is used to represent the actual installation form of the component at the current time point. The actual installation form is an engineering equivalent expression in this invention, used to characterize the macroscopic geometric state of the component during the installation process and to provide a unified geometric basis for contact deviation calculation.

[0025] It is understood that in this invention, the three-dimensional design model of the component is used as the reference model for the component's geometric shape, and its position and orientation are adjusted through rigid body transformation to approximate the actual installation shape of the component. This processing method is mainly used to describe the changes in the macroscopic geometric relationships of the component during the construction and installation process.

[0026] For local shape errors, edge damage, or surface wear that may occur during the manufacturing, transportation, and construction of components, this invention does not distinguish their sources but rather includes them equivalently in the local geometric deviation parameters for comprehensive reflection. Through this equivalent processing method, multiple geometric deviation factors can be uniformly incorporated into the subsequent structural dynamic response analysis and risk identification process without increasing the complexity of additional detection.

[0027] After obtaining the actual installation form of two adjacent components, the sampling area is determined according to the theoretical contact area of ​​the two components in the design drawings. The sampling area is geometrically located near the theoretical contact surface and is constrained by the component's outer shape boundary. Dense sampling is performed on the transformed model surface within the sampling area to obtain a set of sampling points. Dense sampling means obtaining a sufficient number of surface points within the sampling area with a sampling interval smaller than the typical geometric feature scale of the component, so as to ensure that the sampling point set can reflect the overall shape change of the contact area. The sampling point set serves as the input for calculating the vertical distance.

[0028] For each sampling point, the vertical distance from it to the theoretical contact surface equation is calculated and a sign is assigned. The sign direction of the signed vertical distance is defined by the normal of the theoretical contact surface. When the sampling point is located on the side pointed to by the normal, the vertical distance is positive, and when it is located on the opposite side, the vertical distance is negative. The signed vertical distance is used to distinguish the deviation direction of the component relative to the theoretical contact surface and to avoid the loss of deviation direction information caused by using only the absolute value. All the signed vertical distances obtained constitute the vertical distance distribution data.

[0029] The vertical distance distribution data is statistically processed to output the mean and variance. The mean is used to characterize the overall deviation trend within the sampling area and serves as an equivalent quantity of overall non-parallelism. The variance is used to characterize the dispersion of the vertical distance within the sampling area and serves as an equivalent quantity of local undulation. The overall non-parallelism and local undulation together describe the macroscopic tilt and local fluctuation characteristics of the contact area relative to the theoretical contact surface.

[0030] The calculation of the misalignment distance takes the actual contour lines of two adjacent components in their actual installation form as input. The actual contour lines are determined by the intersection line or projection boundary of the outer boundary of the transformed model in the neighborhood of the theoretical contact surface. The actual contour lines of the two components are projected onto the geometric plane or surface parameter domain where the theoretical contact surface is located. The projected contours are used to eliminate the influence of different height directions and highlight the relative misalignment relationship in the plane. The maximum misalignment distance between the two contour lines is calculated in the projection space and output as the misalignment distance. The maximum misalignment distance represents the most unfavorable degree of misalignment of the two components in the plane at the edge of the theoretical contact area.

[0031] After completing the above processing, the overall non-parallelism, local undulation degree and misalignment distance are combined to form local geometric deviation data and associated with the corresponding components for storage. The local geometric deviation data serves as input to describe the degree to which the contact state of adjacent components deviates from the design state during component installation, providing basic data for subsequent analysis of the force evolution of component connection parts.

[0032] It should be noted that during the construction of the breakwater components, the actual contact area between adjacent components may not be directly and completely observed by video equipment due to component fitting, obstruction, or the influence of the construction environment. This invention does not require point-by-point visualization measurement of the entire surface of the contact area. Instead, it indirectly inverts the contact geometry between adjacent components based on the overall spatial position and attitude angle data of the components, combined with the component design model and its geometric constraints.

[0033] In this process, local geometric deviation serves as an equivalent geometric descriptor, reflecting the overall and local deviation trends of the component's installation state relative to its design state, rather than a precise reconstruction of the actual microscopic morphology of the contact area. Even if some contact areas are not directly observed, the installation quality of adjacent components can still be effectively characterized through visible surface markers, edge contours, and the overall rigid body transformation relationships of the components.

[0034] The overall non-parallelism, local undulation, and misalignment distance are engineering characteristic parameters used to describe the contact state of adjacent components. Their physical meaning lies in characterizing the degree of deviation of the components from the designed contact state after installation. These parameters are not used to accurately describe the microscopic stress distribution in the contact area, but rather serve as initial geometric defect inputs in subsequent structural analysis to reflect the influence trend of component installation quality on contact mechanical behavior.

[0035] It should be noted that obtaining spatial position and attitude angle data through surface markers and edge contours, and further forming local geometric deviations, essentially transforms observable external geometric features into calculable spatial geometric relationships on the construction site. This allows the deviations between adjacent components relative to the designed contact state to be expressed in a unified parametric form. The spatial position and attitude angles of the components reflect the overall installation state of the components, while local geometric deviations reflect the degree of geometric mismatch between adjacent components in the contact area. This type of geometric mismatch is an important cause of abnormal contact and stress concentration at the component connection points during the construction phase. Therefore, parametric characterization of these components can provide a prerequisite for identifying the stress change patterns of component connection points under the influence of the external environment, and provide a geometric basis directly related to installation quality for the safety assessment of the construction process.

[0036] Load acquisition module: Based on environmental event records of the sea area where the rock breakwater is located, extract the joint probability distribution of wave action direction, wave height and period to form a load spectrum; In a preferred embodiment of the present invention, the process of obtaining the load spectrum is as follows: Environmental event records are used as input, including historical data sequences of wave action direction, wave height, and period. The data sequences are then cleaned and standardized in terms of units.

[0037] After processing, the historical data sequences of wave direction, wave height, and period are input into the joint statistical analysis process. The joint statistical analysis is used to characterize the probability distribution relationship of the three types of wave parameters occurring at the same time. When establishing the joint probability distribution, the wave direction, wave height, and period are divided into several parameter intervals. The parameter intervals are used to map the continuously valued wave parameters into a finite number of discrete states so as to count the frequency of occurrence. The intervals of wave direction are represented by angle ranges, and the intervals of wave height and period are represented by numerical ranges. After the intervals are divided, each record in the historical data sequence is classified, and its direction value, wave height value, and period value are mapped to the corresponding interval number, thereby converting each record into a discrete ternary combination state.

[0038] The frequency of each ternary combination state in the historical data sequence is obtained by counting all recorded ternary combination states. The frequency is then normalized to the total number of records. The normalization process converts the frequency of occurrence into a probability value, so that the sum of the probability values ​​of each ternary combination state is 1. This yields the joint probability distribution of wave action direction, wave height, and period. The output of the joint probability distribution is a data set containing ternary combination identifiers and corresponding probability values. The ternary combination identifier represents a specific combination of direction interval, wave height interval, and period interval, and the probability value represents the relative probability of this combination occurring in the historical data sequence.

[0039] After the joint probability distribution is established, it is input into the wave parameter combination screening process. A preset probability threshold is used as the screening condition input. The preset probability threshold is used to limit the minimum occurrence probability level of wave parameter combinations that need to be included in the load spectrum. The selection of the threshold is related to the data volume and analysis requirements and is preset during implementation. The screening process traverses the probability values ​​of all ternary combination states in the joint probability distribution and extracts ternary combination states with an occurrence probability higher than the preset probability threshold as valid wave parameter combinations. Each wave parameter combination includes a wave action direction, a wave height, and a period. The wave action direction is taken from the representative value or the center value of the corresponding direction interval, the wave height is taken from the representative value or the center value of the corresponding wave height interval, and the period is taken from the representative value or the center value of the corresponding period interval. This allows each wave parameter combination to be output in a specific numerical form and can be directly used for subsequent simulation input construction.

[0040] For each valid combination of wave parameters, its probability of occurrence in the joint probability distribution is read as the basis for weight allocation. The weight value is used to represent the relative importance of the combination in the environmental event record. The weight value and the probability of occurrence of the combination are monotonically consistent and recorded in numerical form. After the weight allocation is completed, a set of multiple wave parameter combination data with weight values ​​is obtained.

[0041] Multiple sets of wave parameter combinations with weighted values ​​are input into a weighted superposition processing flow. Weighted superposition is used to organize discrete wave parameter combinations into a unified load spectrum expression. In this invention, the load spectrum refers to a collective descriptive data composed of multiple sets of weighted wave parameter combinations. The load spectrum includes the direction value, wave height value, period value, and corresponding weight value of each set of wave parameter combinations, and maintains a structured data format that can be converted into time series wave input in subsequent steps. During the weighted superposition process, multiple sets of wave parameter combinations are normalized or scaled according to the weight values, so that the contribution of different combinations in the load spectrum can be compared, while retaining the independent parameter information of each combination for subsequent organization or retrieval according to weight.

[0042] It is understood that the load spectrum described in this invention is used to characterize the wave action characteristics that the sea area where the rock breakwater is located may experience within a certain time scale. The environmental event records can be selected from historical data sequences covering the construction period or statistical data from a longer time scale; the specific time range can be determined according to the safety assessment requirements of the construction phase. Therefore, the load spectrum is not limited to a single time scale; its function is to provide statistically representative wave input conditions for subsequent time-domain dynamic response simulation, thereby identifying potential areas in the component connection parts that are prone to repeated high contact stress under various possible wave scenarios.

[0043] By establishing a joint probability distribution based on environmental event records and filtering weighted wave parameter combinations, the statistical characteristics of historical wave conditions in the construction sea area are essentially extracted in a structured form. This allows the external environmental effects to be referenced in subsequent processes in the form of representative parameter combinations. The joint probability distribution reflects the correlation between direction, wave height, and period occurring simultaneously. The preset probability threshold focuses the load spectrum on wave conditions that have a stable basis in historical data. The weight values ​​retain the differences in the probability of different wave conditions, so that the external environmental input faced by subsequent analysis has both statistical sources and organization. This enables the stress evolution analysis of component connection parts during the construction stage to be based on wave conditions consistent with the environmental event records of the construction sea area.

[0044] Simulation module: Local geometric deviation is used as the initial defect input into the parametric structural model of the stone breakwater. The load spectrum is applied as a dynamic boundary condition to the surface of the parametric structural model, and time-domain dynamic response simulation is performed to obtain the contact stress time series curve. In a preferred embodiment of the present invention, the specific process for obtaining the contact stress time series curve is as follows: The parametric structural model is used to express the structural state of a stone breakwater composed of multiple components in a spatial coordinate reference. The elements in the model are engineering representations of individual components. The element geometry originates from the component's design 3D model or a simplified representation, maintaining consistency with the component's external dimensions. The spatial position and attitude angle of the elements are taken from the component's spatial position and attitude angle data obtained in the preceding steps, ensuring the model reflects the actual installation state of the components at the initial moment. Adjacent components establish interaction relationships through contact pairs. A contact pair is a mechanical connection element in the parametric structural model used to describe the potential contact area between two components. The geometric definition of a contact pair is determined by the theoretical contact area of ​​the two components or the adjacent area between adjacent surfaces in the actual installation state. Each contact pair is assigned a unique identifier to record its force and stress information in subsequent calculations.

[0045] The mechanical properties of the contact pair are used to describe the normal compression response and tangential relative slip response of the two components when they come into contact. The mechanical properties include normal contact stiffness, tangential contact stiffness, friction-related parameters, and contact damping parameters for numerical stabilization. Normal contact stiffness is used to characterize the relationship between the relative displacement and normal contact force when the contact surface is compressed in the normal direction. Tangential contact stiffness is used to characterize the growth trend of tangential force when the contact surface undergoes a small relative displacement in the tangential direction. Friction-related parameters are used to limit the relationship between tangential force and normal force and characterize the conditions under which slip occurs. Contact damping parameters are used to characterize the energy dissipation and vibration attenuation characteristics during the contact process.

[0046] When the mechanical properties of the contact pair are set based on local geometric deviations, the local geometric deviations are used as input data. These local geometric deviations consist of overall non-parallelism, local undulation, and misalignment distance. Overall non-parallelism is the mean of the signed vertical distance distribution in the contact area; local undulation is the degree of dispersion characterized by the variance of the vertical distance distribution; and misalignment distance is the maximum distance between the actual contour lines of the two components projected onto the theoretical contact surface. The mechanical properties of the contact pair in this invention include at least normal contact stiffness, tangential contact stiffness, friction-related parameters, and contact damping parameters. Normal contact stiffness describes the response of the normal contact force in the normal compression direction to changes in the normal relative displacement; tangential contact stiffness describes the response of the tangential contact force in the tangential direction to changes in the tangential relative displacement; friction-related parameters limit the upper limit of the tangential interaction force and the slip condition; and contact damping parameters describe the influence of energy dissipation on the interaction force during the contact process.

[0047] To apply geometric deviations to the assignment of the aforementioned mechanical properties, local geometric deviations are first transformed into three types of contact state descriptors for the contact pair. These contact state descriptors include the degree of initial contact non-uniformity, the degree of contact interval, and the effective contact ratio. The degree of initial contact non-uniformity is determined by the overall non-parallelism. The greater the overall non-parallelism, the greater the difference in the initial gap along the normal of the theoretical contact surface, causing the contact to gradually expand from local initial contact to surface contact. The degree of contact interval is determined by the degree of local undulation. The greater the degree of local undulation, the more significant the difference between local high points and low points in the contact area, making it easier for the contact to repeatedly contact and separate during the stress process. The effective contact ratio is determined by the misalignment distance. The greater the misalignment distance, the more obvious the misalignment of the contact area edges in the plane, reducing the overlapping area that can form a stable pressure-bearing structure.

[0048] Based on the above three types of contact state description quantities, when setting the normal contact stiffness of the contact pair, the overall non-parallelism and the degree of local undulation are used together as correction criteria. Under the premise of the same material and the same contact surface type, a normal contact stiffness reference value is preset. When the overall non-parallelism is small and the degree of local undulation is small, it is considered that the contact is closer to a uniform surface contact, and the normal contact stiffness is taken as the reference value or a value near the reference value. When the overall non-parallelism increases and the initial contact is obviously uneven, the normal contact stiffness is taken as a value lower than the reference value to reflect the equivalent softening brought about by the gradual expansion of local contact. When the degree of local undulation increases and the intermittency of contact increases, the normal contact stiffness is further reduced to avoid incorrectly equating intermittent contact with continuous rigid contact.

[0049] When setting the contact damping parameter, the degree of local undulation is taken as the main basis and combined with the overall non-parallelism. When the degree of local undulation is small and the overall non-parallelism is small, the contact damping parameter is taken as a lower value to reflect the smoothness of the contact process. When the degree of local undulation increases and the contact separation and re-contact are more frequent, the contact damping parameter is taken as a higher value than the benchmark to reflect the increased energy consumption in the contact process and suppress the high-frequency oscillation caused by intermittent contact in the numerical calculation.

[0050] When setting the tangential contact stiffness, the effective contact ratio is used as the primary basis, combined with the degree of local undulation. Based on the preset tangential contact stiffness benchmark value, when the misalignment distance is small and the effective contact ratio is close to the complete contact area, the tangential contact stiffness is taken as the benchmark value or a value near the benchmark value. When the misalignment distance increases and the effective contact ratio decreases significantly, the tangential contact stiffness is reduced accordingly to reflect the reduction in the effective contact area that can transmit tangential force. When the degree of local undulation increases and the tangential contact changes from in-plane overall deformation to local micro-slippage, the tangential contact stiffness is further reduced to reflect the decrease in tangential constraint capability.

[0051] When setting friction-related parameters, the misalignment distance and the degree of local undulation are used together as equivalent correction basis. Based on the preset friction coefficient benchmark value, when the misalignment distance increases and the contact is concentrated at the edge or local strip area, the friction-related parameters adopt a value lower than the benchmark value to equivalently reflect the reduction in the actual usable friction bearing area. When the degree of local undulation increases and the local contact point is more prone to micro-slippage, the friction-related parameters also adopt a value lower than the benchmark value to equivalently reflect the contact state that is more prone to slippage.

[0052] To ensure the traceability of the setup process, the local geometric deviation input value of each contact pair, along with the corresponding output normal contact stiffness, tangential contact stiffness, friction-related parameters, and contact damping parameters, are recorded and stored together with the contact pair identifier. This allows the mechanical properties of each contact pair in subsequent numerical calculations to be traced back to the source of the corresponding local geometric deviation.

[0053] After obtaining the load spectrum and completing the definition of the parametric structural model and contact pairs, the load spectrum is used as the statistical input of the wave environment and transformed into wave input data that can be directly applied to the wet surface of the model. The load spectrum consists of multiple sets of wave parameter combinations and their occurrence probability weights. The wave parameter combinations include the wave action direction, wave height, and period. The occurrence probability weights are used to characterize the relative probability of the combination in the environmental event record and serve as the basis for subsequent sampling and time organization. Wave input refers to the distributed pressure time series data set indexed by time and the spatial position of the wet surface of the parametric structural model. The wet surface of the parametric structural model refers to the set of surface units on the outer surface of the model that are in contact with the water and bear the wave pressure. The spatial position of the wet surface unit is described by three-dimensional coordinates under the spatial coordinate reference and stored in association with the normal direction of the unit, so that the pressure can act on the corresponding unit in the form of surface force. In one embodiment of the present invention, a conversion method may be: Multiple sets of wave parameter combinations and their occurrence probability weights are extracted from the load spectrum; for each set of wave parameter combinations, the free surface wave profile equation corresponding to the set of wave parameters is calculated based on linear wave theory or Stokes wave theory, and the velocity and acceleration of wave water particles are obtained at each spatial position on the wet surface of the parameterized structural model accordingly. The motion of water particles is converted into wave action applied to the wet surface of the parameterized structural model using any of the following methods: (1) The instantaneous dynamic pressure caused by the waves (e.g., Froude-Krylov pressure) is obtained based on the wave potential function and superimposed with the hydrostatic pressure to obtain the distributed pressure time history on each unit of the wet surface; (2) When the wet surface unit can be equivalent to a slender member or a local member that satisfies the applicable conditions of the Morrison equation, the hydrodynamic force of the inertial term and resistance term of the unit is calculated according to the Morrison equation, and the hydrodynamic force is converted into the equivalent distributed pressure time history by combining the characteristic length of the unit and the equivalent force area; Wave parameter combinations are sampled based on their probability weights, and each sampled wave parameter combination is assigned a corresponding duration to generate multiple distributed pressure time histories. Adjacent pressure time histories are connected using a smooth transition function to obtain a long-term, continuous wave pressure time series, which serves as the wave input.

[0054] Specifically, at the start of the load spectrum transformation, multiple sets of wave parameter combinations and their occurrence probability weights are read from the load spectrum as calculation inputs. For each set of wave parameter combinations, a free surface wave profile expression corresponding to that set of parameters is established based on linear wave theory or Stokes wave theory. Wave height and period determine the undulation characteristics of the free surface over time, and the wave action direction determines the propagation direction of the free surface phase along the spatial coordinate reference. Based on this, each spatial position on the wet surface of the model is taken as a solution point. According to the selected wave theory, the velocity and acceleration time series of wave water particles at that spatial position are obtained. The velocity and acceleration of water particles are vector data that change with time and are related to the wave propagation direction and water depth conditions. The obtained data is used as the intermediate output of hydrodynamic calculation.

[0055] When converting the motion of water particles into wave action applied to a wet surface, a pressure calculation method matching the geometric characteristics of the wet surface unit is selected to generate a distributed pressure time history output. The wave potential function is used to superimpose the instantaneous dynamic pressure caused by the wave with the hydrostatic pressure to obtain the total pressure time series on the wet surface unit. The instantaneous dynamic pressure is determined by the characteristics of the wave potential function changing with time and changes with the phase of the free surface. The hydrostatic pressure is determined by the immersion depth of the wet surface unit and serves as a reference pressure that changes slowly or is approximately constant with time. The total pressure obtained after superposition is indexed by the wet surface unit identifier and timestamp to form a pressure time history data table. When the wet surface element can be equivalent to a slender member or a local member that satisfies the applicable conditions of the Morrison equation, the inertial hydrodynamic and drag hydrodynamic terms of the element are calculated based on the velocity and acceleration of the water particles. The inertial hydrodynamic term changes with acceleration, and the drag hydrodynamic term changes with velocity and is related to the velocity direction. After combining the two hydrodynamic terms into a time series of element hydrodynamics, the hydrodynamics are converted into an equivalent distributed pressure time series by combining the element characteristic length and the equivalent force-bearing area, so that the pressure is still borne by the wet surface element and is consistent with the interface of subsequent load application.

[0056] After calculating the pressure time history for each set of wave parameter combinations, the probability weights are used as the sampling basis to sample the wave parameter combinations. The sampling process follows the principle that the higher the weight, the higher the probability of selection, forming a wave parameter combination sequence. A corresponding duration is assigned to each sampled wave parameter combination in the sequence. The duration determines the length of the pressure time history set on the wave input time axis. The duration can be a multiple of the set's period to ensure complete periodic variation within the pressure time history segment. For example, when a set's period is 6, the duration is set to an integer multiple of its period. To ensure continuous and repeatable pressure changes within a segment, multiple distributed pressure time histories are generated sequentially and spliced ​​into a long-term series. To avoid discontinuous pressure jumps caused by parameter abrupt changes between adjacent time histories, a smooth transition function is introduced at the boundary of adjacent pressure time histories to weight and mix the two pressure segments. This ensures that the pressure in the transition interval smoothly transitions from the previous segment to the next segment while maintaining the continuity of the time series. After splicing, a wave pressure time series data set covering the total duration is output. This data set uses timestamp sequences and wetted surface unit identifiers as dual indices and stores the corresponding pressure values, serving as the final output of the wave input.

[0057] When wave input is applied as an external load to the wet surface of the parametric structural model, the wet surface pressure distribution corresponding to each time step of the numerical calculation is read. The pressure is multiplied by the wet surface element area and applied as an equivalent surface force along the element normal direction, so that the external load enters the structural dynamic solution process. At the same time, the spatial distribution and temporal variation of the pressure are kept consistent with the wave input.

[0058] When setting the simulation time step and total duration, the time step is defined as the time increment between two adjacent calculation moments and is consistent with the resolution of the wave input timestamp or its integer division. This ensures that each time step can obtain the corresponding pressure value from the wave input or obtain the pressure value through time interpolation. The total duration covers the complete cycle of the wave input in the load spectrum and is consistent with the time axis after the wave input is spliced, thereby ensuring that the output contact stress time series curve covers the entire wave input process.

[0059] When using numerical calculation methods to solve the model response, load updates and contact solutions are completed at each time step. Load updates convert the wet surface pressure of the current time step into an external load and apply it to the outer surface of the model. Contact solutions determine whether the relative gap between the two sides of each contact pair in the normal direction meets the contact conditions for each pair and calculate the interaction force between the contact pairs accordingly. The interaction force includes normal interaction force and tangential interaction force. The normal interaction force is determined by the normal contact stiffness of the contact pair and the current normal relative displacement, and can be combined with contact damping to reflect the energy dissipation during the contact process. The tangential interaction force is determined by the tangential contact stiffness and tangential relative displacement and is constrained by friction-related parameters to reflect the slip conditions and the force limitations after slip.

[0060] Specifically, in the contact detection stage, representative point sets or representative surface patches are selected in the candidate contact area of ​​the contact pair to determine contact. The relative gap between the two surfaces in the normal direction of the theoretical contact surface is calculated as the gap value and compared with the preset contact tolerance. When the gap value is less than or equal to the contact tolerance, the contact state is marked as contact and enters the subsequent calculation. When the gap value is greater than the contact tolerance, the contact state is marked as separation and the normal interaction force and tangential interaction force are set to 0. At the same time, the cumulative amount of tangential relative slip is cleared to zero or kept at 0.

[0061] During the contact geometry update stage, when contact is determined, the contact normal direction is determined. The contact normal direction is taken as the theoretical contact surface normal or as the local average normal of the two surfaces in the contact area and uniformly pointed to one side element. Then, the normal relative displacement is calculated. The normal relative displacement represents the compression of the two surfaces in the normal direction. Its value can be obtained by subtracting the current gap value from the contact tolerance and is limited to non-negative to avoid generating tension. At the same time, the relative velocity vector is calculated and decomposed into normal relative velocity and tangential relative velocity. The normal relative velocity is used for damping calculation, and the tangential relative velocity is used for friction and tangential elasticity calculation.

[0062] In the normal interaction force calculation stage, the normal relative displacement is used as the elastic compression input. The elastic normal force is calculated based on the normal contact stiffness, and the damping normal force is calculated based on the contact damping parameters and the normal relative velocity. The elastic normal force and the damping normal force are synthesized along the contact normal direction to obtain the normal interaction force vector. If the synthesis result points in the pulling direction due to numerical reasons, the normal interaction force is truncated to 0 to ensure that only the compressive force is generated.

[0063] In the tangential interaction force calculation stage, a tangential base direction is established in the contact plane. The tangential base direction can be determined by the tangential relative velocity direction and continues along the tangential direction of the previous time step when the tangential relative velocity is close to 0. Then, the tangential relative slip accumulation is updated. The tangential relative slip accumulation is obtained by adding the accumulation of the previous time step to the tangential relative velocity of the current time step and multiplying by the time step size, and is used as the equivalent quantity of tangential elastic deformation. The tangential elastic force is calculated based on the tangential contact stiffness and the candidate tangential force vector is obtained by oriented along the tangential base direction. In the friction constraint stage, the maximum allowable tangential force amplitude is calculated. The maximum tangential force amplitude is determined by the friction-related parameters and the amplitude of the normal interaction force. When the candidate tangential force amplitude is less than the maximum tangential force amplitude, it is determined to be an adhesive state and the candidate tangential force is used as the tangential interaction force. At the same time, the updated tangential relative slip accumulation is retained. When the candidate tangential force amplitude is greater than the maximum tangential force amplitude, it is determined to be a slip state and the tangential interaction force amplitude is truncated to the maximum tangential force amplitude with the direction taken from the candidate tangential force. At the same time, the tangential relative slip accumulation is back-calculated to make it consistent with the truncated tangential interaction force, ensuring the continuity of the tangential force evolution in the next time step.

[0064] The effective contact area serves as the fundamental quantity for calculating contact stress in the interaction force calculation and is updated synchronously with the contact state. When the contact state is "contact," the effective contact area is taken from the effective area defined by the geometry of the contact pair, combined with area correction results based on factors such as misalignment distance. When the contact state is "separation," the effective contact area is set to 0. After completing the above calculations, the normal interaction force vector, tangential interaction force vector, contact state flag, and effective contact area of ​​the contact pair at the current time step are output, and the accumulated tangential relative slip is written back as a state variable for continued use in the next time step.

[0065] Contact stress is calculated at each time step based on the geometric area and interaction force of the contact pair. The geometric area is the effective contact area of ​​the contact pair and is consistent with the geometric definition of the contact pair. Contact stress is an engineering equivalent quantity. Normal contact stress is calculated from normal interaction force and geometric area, and tangential contact stress is calculated from tangential interaction force and geometric area. When a unified characterization is required, the normal and tangential stresses can be combined into a single contact stress amplitude as the stress value of the contact pair at that time step. The contact stress calculated at each time step, along with the timestamp and contact pair identifier, is written into the time-series data record table. The time-series data record table forms a stress sequence arranged by time for each contact pair. After the total simulation duration is completed, the time-series data record tables for all contact pairs are output. The output data set includes contact pair identifiers, timestamp sequences, and corresponding contact stress sequences, thus forming a contact stress time-series curve showing the change of contact stress over time.

[0066] During the establishment of the parametric structural model, the mechanical property parameters encountered can be set based on engineering experience values, existing experimental data, or parameter ranges from similar projects. Due to the complexity of component contact behavior and wave action, this invention does not use the absolute value of contact stress under a single working condition as the sole criterion for judgment. Instead, it focuses more on the spatial relative distribution characteristics of peak contact stress under different parameter combinations and wave input conditions.

[0067] By performing numerical calculations and analyses on various possible parameter combinations, the spatial locations where contact stress peaks repeatedly occur within the parameter variation range can be identified, thereby improving the robustness of the identification results for high-risk component connection parts.

[0068] Analysis module: Analyzes the evolution characteristics of the contact stress time series curve, extracts the spatial location and time interval patterns of the contact stress peak occurrence, and generates a heat map of the spatial distribution of contact stress peak and a peak recurrence interval sequence; In a preferred embodiment of the present invention, the specific process for generating a thermal map of the spatial distribution of contact stress peaks and a peak recurrence interval sequence is as follows: After obtaining the contact stress time-series curves for all contact pairs, each time-series curve, along with its contact pair identifier and spatial location identifier, is used as input data into the peak identification and spatial statistics process. The spatial location identifier is used to characterize the position of the contact pair in the parametric structural model of the breakwater. It can be represented by the geometric center coordinates of the contact area corresponding to the contact pair or the center coordinates of the contact boundary between the two components, and is associated with the contact pair identifier one by one, so that subsequent statistics can map peak events to specific spatial locations. When traversing the time-series curves, a preset threshold is used as the peak filtering condition input. The preset threshold is the criterion for judging the contact stress amplitude and is preset during implementation. Each curve is read point by point in chronological order. When the stress value at a certain time point is higher than the preset threshold and greater than the stress value at its adjacent time points, that time point is determined as the stress peak point, and its occurrence time and stress amplitude are recorded. At the same time, the peak point is associated with the spatial location identifier of the corresponding contact pair to form a peak event record table. The peak event record table includes the peak event number, contact pair identifier, spatial location identifier, occurrence timestamp, and stress amplitude. After identifying the peak values ​​of all curves, the peak event record table is categorized according to spatial location identifiers. Peak events with the same spatial location identifier are grouped into the same spatial location region. This spatial location region represents a statistical unit in the heatmap and can be directly indexed by the spatial location identifier of the contact pair, or adjacent contact pairs can be aggregated into the same region index according to a preset spatial resolution to avoid overly fragmented regions. Within each spatial location region, the number of peak events or the cumulative value of peak amplitudes are counted. The number represents the frequency of peak events exceeding a threshold in that region, and the cumulative amplitude represents the overall level of peak intensity in that region. The statistical results are written into a region statistics table, which contains the spatial location region index and the corresponding statistical value. Based on the region statistics table, a color intensity is assigned to each spatial location region. When mapping statistical values ​​to color intensity, a monotonic mapping relationship is used, so that regions with larger statistical values ​​exhibit higher color intensities. This results in a heatmap data layer, which is presented in color intensity form on the corresponding contact areas of the 3D model of the breakwater and output as a heatmap of the spatial distribution of contact stress peak values. The generation of the peak recurrence interval sequence takes the peak event record table as input. For the same spatial location area, all peak events in that area are read and sorted in ascending order by their occurrence timestamps. The difference between the timestamps of two adjacent peak events in the sorted sequence is calculated and used as a peak recurrence interval value for that area. This process is repeated for all spatial location areas. Finally, the spatial location area indices and their corresponding peak recurrence interval sets are organized into a unified data sequence and output. The output sequence contains location identifiers and interval values ​​and maintains a structured format that can be used for subsequent overlay analysis, thus forming the peak recurrence interval sequence.

[0069] Output module: The thermal map of the spatial distribution of contact stress peak is superimposed with the peak recurrence interval sequence to obtain the potential distribution map of contact stress. The potential distribution map of contact stress identifies the component connection parts where high stress occurs repeatedly.

[0070] In a preferred embodiment of the present invention, the specific process for obtaining the potential distribution map of contact stress is as follows: After obtaining the thermal map of the spatial distribution of peak contact stress and the peak recurrence interval sequence, the thermal map is used as a base layer and input into the image overlay processing flow. The base layer uses the spatial location region of the 3D model of the breakwater as an index and the color intensity to represent the statistical level of the peak event in that region. At the same time, the peak recurrence interval sequence is used as the overlay data input. The peak recurrence interval sequence contains spatial location region identifiers and corresponding interval value sets. Based on the interval value set of each spatial location region, a graphical element layer is generated. The graphical elements are a set of visual symbols attached to the same spatial coordinate system and correspond one-to-one with the spatial location regions. The visual features of the graphical elements are used to express the length of the interval time. The visual features are encoded using symbol density or symbol size, so that regions with shorter interval times correspond to higher density or larger size symbol representations, and regions with longer interval times correspond to lower density or smaller size symbol representations, thereby forming a distinguishable temporal feature expression under the same view. After generating the graphical elements, the graphical element layers are overlaid onto the base layer according to their spatial location regions, ensuring that their coordinates are consistent. A blending rendering process is then performed on the two overlaid layers to generate a composite image. During the blending rendering process, the transparency of the graphical elements is set while preserving the color information of the base layer, ensuring that both color intensity and the graphical elements are discernible. After the composite image is generated, the component connection points that repeatedly exhibit high stress are identified in the composite image according to preset discrimination rules. The discrimination rules define areas with high color intensity as areas with significant peak distribution and areas with short intervals between graphical element representations as areas with frequent peak occurrences. When the same spatial location region simultaneously meets both of these conditions, that spatial location region is marked as a component connection point that repeatedly exhibits high stress, and the marking result and corresponding spatial location region identifier are output in the contact stress potential distribution map.

[0071] In the process of identifying and statistically analyzing peak contact stress, the preset threshold is used to distinguish stress peak events of engineering significance in the contact stress time series curve. This threshold can be set according to the component material properties, engineering experience, or analytical accuracy requirements; its specific numerical form does not constitute a limitation of this invention. By using threshold filtering, significant peak characteristics in the contact stress variation process can be highlighted while ensuring computational efficiency, thereby providing a stable data foundation for subsequent spatial distribution analysis and time interval statistics.

[0072] In the spatial statistical analysis of contact stress peak values, the spatial location region can be divided according to component size, structural division method, or analysis accuracy requirements. The purpose is to classify and statistically analyze the spatial distribution characteristics of contact stress peak values. The specific division method of the spatial location region does not affect the technical effect of this invention in identifying recurring high-stress component connection points through peak distribution characteristics.

[0073] The potential contact stress distribution map is not merely for graphical display, but serves as an output of analytical results comprehensively reflecting the amplitude and temporal evolution characteristics of contact stress. By fusing the spatial distribution characteristics and peak recurrence interval characteristics of contact stress peaks, component connection points exhibiting both high stress levels and high recurrence frequencies can be identified within a unified coordinate system. This potential contact stress distribution map provides a basis for risk assessment and construction process adjustments, enabling construction managers to proactively identify potentially high-risk areas during the construction phase, thereby improving the safety assurance capabilities of the construction process.

[0074] In this invention, the "component connection location exhibiting repeated high stress" refers to a location where, under various wave input scenarios and numerical analysis conditions, the peak contact stress has a high spatial concentration and exhibits a short recurrence interval in the time series. This determination reflects the risk that the component connection location may be subjected to relatively frequent or significant contact effects during construction and operation, and is not limited to situations where material failure or structural damage has already occurred.

[0075] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation. The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the present invention should still fall within the scope of the present invention.

Claims

1. A video surveillance system for ensuring safety during construction, characterized in that, include: Video recognition module: acquires multi-angle video streams of the stone breakwater components and identifies surface markers and edge contours of the components in the multi-angle video streams; Deviation acquisition module: Based on surface markers and edge contours, it acquires the spatial position and attitude angle data of each component, and thereby obtains the local geometric deviations generated during component installation. Load acquisition module: Based on environmental event records of the sea area where the rock breakwater is located, extract the joint probability distribution of wave action direction, wave height and period to form a load spectrum; Simulation module: Local geometric deviation is used as the initial defect input into the parametric structural model of the stone breakwater. The load spectrum is applied as a dynamic boundary condition to the surface of the parametric structural model, and time-domain dynamic response simulation is performed to obtain the contact stress time series curve. Analysis module: Analyzes the evolution characteristics of the contact stress time series curve, extracts the spatial location and time interval patterns of the contact stress peak occurrence, and generates a heat map of the spatial distribution of contact stress peak and a peak recurrence interval sequence; Output module: The thermal map of the spatial distribution of contact stress peak is superimposed with the peak recurrence interval sequence to obtain the potential distribution map of contact stress. The potential distribution map of contact stress identifies the component connection parts where high stress occurs repeatedly.

2. The video surveillance system for ensuring safety during construction as described in claim 1, characterized in that, The process of determining local geometric deviations is as follows: The theoretical contact surface equation between adjacent components is determined based on the design drawings. The theoretical contact surface equation is defined by a mathematical expression for a plane or curved surface. Using the spatial position and attitude angle data of the component, the three-dimensional design model of the component is transformed into a rigid body. The point set on the surface of the transformed model represents the actual installation form of the component. Dense sampling was performed in the contact area of ​​the actual installation configuration. The signed perpendicular distance from each sampling point to the theoretical contact surface was calculated. The mean and variance of the signed perpendicular distance were statistically analyzed. The mean and variance reflect the overall non-parallelism and local undulation of the contact area, respectively. Calculate the maximum misalignment distance between the actual contour lines of the two components on the theoretical contact surface projection, and use it as the misalignment distance; The overall non-parallelism, the degree of local undulation, and the distance of misalignment together constitute the local geometric deviation.

3. The video surveillance system for ensuring safety during construction as described in claim 1, characterized in that, The process of obtaining the load spectrum is as follows: Environmental event logs contain historical data sequences of wave action direction, wave height, and period; A joint statistical analysis was performed on the historical data sequences of wave action direction, wave height, and period to establish a joint probability distribution of wave action direction, wave height, and period. In the joint probability distribution, all wave parameter combinations with an occurrence probability higher than a preset probability threshold are extracted. Each wave parameter combination includes a wave action direction, a wave height, and a period. Based on the probability of occurrence of each wave parameter combination, a weight value is assigned to each wave parameter combination. Multiple wave parameter combinations with weight values ​​are weighted and superimposed to form a load spectrum.

4. The video surveillance system for ensuring safety during construction as described in claim 1, characterized in that, The specific process for obtaining the contact stress time series curve is as follows: A parametric structural model of the stone breakwater is established. Each unit in the model corresponds to a component. Units are connected by contact pairs. The mechanical properties of the contact pairs are set according to local geometric deviations. The load spectrum is converted into a time-series wave input, which includes the variation of wave pressure over time and space. Wave input is applied as an external load to the wet surface of the parametric structural model; Numerical calculation methods were used to solve the contact stress time series curve of the parameterized structural model under wave input.

5. The video surveillance system for ensuring safety during construction as described in claim 4, characterized in that, The specific process of obtaining the contact stress time series curve also includes: Set the simulation time step and total duration, with the total duration covering the complete cycle of the wave input in the load spectrum; At each time step, the distribution of wave pressure on the wetted surface of the model is calculated, and the interaction forces between the contact pairs are also calculated. Based on the geometric area and interaction force of the contact pair, calculate the contact stress of each contact pair at each time step; After the entire simulation period is over, the data on the change of contact stress over time for each contact pair is output, forming a contact stress time series curve.

6. The video surveillance system for ensuring safety during construction as described in claim 1, characterized in that, The specific process for generating a thermal map of the spatial distribution of peak contact stress is as follows: Traverse all contact stress time series curves and identify the stress peak points on each curve that exceed the preset threshold. Record the spatial location and stress amplitude of each stress peak point; All stress peak points are classified according to their spatial location, and the number or cumulative amplitude of stress peak points in each spatial location area is counted. Based on the statistical results, a corresponding color intensity is assigned to each spatial location region to generate a heat map reflecting the spatial distribution characteristics of the peak contact stress.

7. The video surveillance system for ensuring safety during construction as described in claim 6, characterized in that, The specific process for generating the peak recurrence interval sequence is as follows: For the same spatial region, all identified stress peak points are sorted according to the order of their occurrence. Calculate the time difference between adjacent stress peak points after sorting to obtain a set of peak recurrence intervals at that location; For each spatial location region, a corresponding set of peak recurrence intervals is generated, and the peak recurrence interval sets of all regions are organized to form a data sequence containing location identifiers and interval values, which serves as the peak recurrence interval sequence.

8. The video surveillance system for ensuring safety during construction as described in claim 7, characterized in that, The specific process for obtaining the potential distribution map of contact stress is as follows: Use the thermal map of the spatial distribution of peak contact stress as the base layer; The peak recurrence interval sequence is overlaid on the base layer as a graphical feature, and the visual features of the graphical feature are used to express the length of the interval. The base layer and the graphical feature layer are merged and rendered to generate a composite image. In the composite image, regions with high color intensity and short intervals between peak recurrence intervals in the thermal map of contact stress peak spatial distribution are identified by combining visual features. These regions are then marked as component connection points where high stress occurs repeatedly in the potential distribution map of contact stress.

Images