Road construction safety monitoring method based on unmanned aerial vehicle image shooting

Abstract

The application discloses a road construction safety monitoring method based on unmanned aerial vehicle image shooting, and belongs to the technical field of construction safety monitoring. The road construction scene is first divided into three risk partitions of a roadbed, a bridge and a culvert, and a foundation pit operation area, a differentiated unmanned aerial vehicle collection scheme is formulated, and a standardized image set with a time-space label of each partition is obtained; then, for each partition core danger source, corresponding safety monitoring parameters are extracted, and exclusive safety data sets of each risk partition are constructed; coupling analysis is carried out on the safety data set through the construction of parameter coupling correlation rules, and a danger hidden danger index of each partition is output; finally, risk grading alarm is completed in combination with a preset threshold rule, a safety monitoring report is integrated and generated and is pushed to corresponding management personnel. The application realizes global precise monitoring, risk quantitative evaluation and intelligent early warning of road construction safety, and greatly improves the efficiency and intelligent level of construction safety control.

Description

Technical Field

[0001] This invention relates to the field of construction safety monitoring technology, and more specifically, to a method for monitoring road construction safety based on drone imagery. Background Technology

[0002] In recent years, drone aerial photography technology has been gradually promoted and applied in engineering construction inspection scenarios, but existing road construction monitoring technologies based on drone imagery still have the following technical shortcomings: First, a standardized safety monitoring parameter system for core risk zones such as roadbeds, bridges, culverts, and foundation pits has not been established to address the characteristics of hazards and the monitoring accuracy requirements of different risk operation areas in road construction. As a result, it is impossible to automatically and standardize the extraction of multi-dimensional safety parameters from UAV images and derived models. The monitoring results are mainly qualitative descriptions of hidden dangers, lacking quantitative safety status evaluation and risk judgment basis. Without a multi-parameter coupled quantitative assessment model for construction risks, it is impossible to output a hazard index that can be directly used for risk classification based on multi-dimensional monitoring data, making it difficult to achieve automated risk alarms and accurate push of hazard information.

[0003] In summary, existing road construction safety monitoring technologies cannot simultaneously meet the multiple requirements of full coverage, high-frequency dynamic data acquisition, accurate parameter quantification, comprehensive risk assessment, and intelligent alarms. This has become a key technical bottleneck restricting the improvement of the digitalization and intelligence level of road construction safety management. Therefore, a road construction safety monitoring method based on UAV imagery is proposed. Summary of the Invention

[0004] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a method for monitoring road construction safety based on drone imagery.

[0005] To achieve the above objectives, the present invention provides the following technical solution: Road construction safety monitoring methods based on drone imagery include: Step 1: Based on the road construction scenario, risk zones are pre-divided, and differentiated data acquisition plans are developed for different zones. Image data of each construction zone is collected by drones to obtain standardized image sets with spatiotemporal labels for each zone. The risk zones include roadbed operation zone, bridge and culvert operation zone, and foundation pit operation zone; Step 2: For the core hazards associated with each risk zone, extract the monitoring parameters related to the safety level of the corresponding zone from the corresponding standardized image set, and construct the safety dataset for each risk zone; Step 3: Extract the safety dataset corresponding to each risk zone, construct parameter coupling association rules within different zones, and use the parameter coupling association rules to perform coupling analysis on the safety datasets of different risk zones, outputting the hazard index of different risk zones corresponding to the road construction site; Step 4: Based on the hazard index output by each risk zone and combined with the preset threshold judgment rules, complete the risk alarm for each risk zone, and integrate and generate a safety monitoring report to be sent to the management personnel of the respective risk zone.

[0006] Specifically, the security datasets for each risk partition include: The safety dataset for the roadbed work area includes the roadbed slope deviation rate, the time-series uneven settlement of the roadbed top surface, the integrity of the slope protection structure, and the compliance rate of illegal surcharges on the roadbed. The safety dataset for bridge and culvert work areas includes verticality deviation values, maximum relative deformation, and alignment deviation rate. The safety dataset for the foundation pit operation area includes the foundation pit displacement, cumulative soil settlement, and degree of crack development.

[0007] Specifically, the output shows the hazard index for different risk zones corresponding to the road construction site: The hazard index of the roadbed operation area is obtained by comprehensively processing the roadbed slope deviation rate, the time-series uneven settlement of the roadbed top surface, the integrity of the slope protection structure, and the compliance rate of illegal surcharges on the roadbed, combined with a preset standard set for the roadbed operation area. The hazard index of bridge and culvert work areas is obtained by comprehensively processing the verticality deviation value, the maximum relative deformation, and the alignment deviation rate in combination with a preset standard set for bridge and culvert work areas. The hazard index of the foundation pit operation area is obtained by comprehensively processing the foundation pit displacement, cumulative soil settlement, and crack development degree in combination with the preset standard set of foundation pit operation areas.

[0008] Specifically, the calculation logic for the safety dataset of the roadbed work area includes: From the DSM digital surface model of the UAV, a standard cross section is extracted along the direction of the roadbed slope; the actual slope height and actual horizontal projection width of the slope are extracted from the cross section, and the actual slope ratio is calculated. The preset design slope is retrieved as the benchmark value, and the deviation rate is calculated by combining it with the actual slope, and the slope deviation rate is output. The arithmetic mean of the slopes of the entire roadbed operation area is then taken to obtain the roadbed slope deviation rate. Extract the DSM data of the roadbed top surface generated by the UAV at the previous historical acquisition time point before the current acquisition time point; pre-deploy equally spaced monitoring points on the roadbed top surface and obtain the elevation values ​​of each point at adjacent acquisition time points; for the same monitoring point, subtract the elevation value of the previous historical acquisition time point from the elevation value of the current acquisition time point to obtain the cumulative settlement of the target monitoring point; calculate the absolute value of the difference between the cumulative settlement of two adjacent monitoring points and divide it by the horizontal distance between the two points to obtain the uneven settlement gradient; identify the maximum uneven settlement gradient in the entire roadbed operation area to obtain the temporal uneven settlement of the roadbed top surface.

[0009] Specifically, the calculation logic for the safety dataset of the roadbed work area also includes: Based on images collected by drones, the slope protection structure is identified through AI semantic segmentation. Through threshold segmentation, the areas of damaged, void, eroded, and cracked areas in the slope protection structure are identified and summed to obtain the damaged and missing area. Combining the damaged and missing area with the preset total design protection area, the integrity of the slope protection structure is output. In the images collected by the drone, no-stack zones are pre-defined, and the presence of stacking is determined segment by segment along the roadbed slope. The length of continuous segments without illegal stacking is counted and divided by the total monitored length of the slope to obtain the compliance rate of illegal stacking on the roadbed.

[0010] Specifically, the calculation logic for the safety dataset in the bridge and culvert work area includes: From the three-dimensional solid model pre-generated by the UAV, locate the three-dimensional coordinates of the center point A of the bottom pier cap and the three-dimensional coordinates of the center point B of the top of the pier. Calculate the actual effective height and horizontal offset of the pier column; divide the horizontal offset by the actual effective height to obtain the relative sag deviation coefficient. Retrieve the preset allowable relative sag deviation coefficient and allowable horizontal offset; Substitute into the formula Obtain the verticality deviation value ;in The relative sag deviation coefficient is calculated. To allow relative sag deviation coefficient; To allow for horizontal offset, This represents the actual horizontal offset of the pier column.

[0011] Specifically, the safety dataset calculation logic for bridge and culvert work areas also includes: Using the pre-compression acceptance DSM as the benchmark model, the initial three-dimensional coordinates of each pre-set key monitoring point are extracted; the real-time three-dimensional coordinates of different key monitoring points within the current acquisition time point are extracted; the real-time three-dimensional coordinates and initial three-dimensional coordinates of different key monitoring points are combined to output the relative deformation of different key monitoring points; the maximum value of the relative deformation of each key monitoring point is taken as the highest relative deformation. Locate the center position of the supports at both ends of the precast beam, and read the actual coordinates of the first support and the second support at the beam end respectively; extract the pre-designed coordinates of the two supports, and calculate the actual alignment offset of the two supports respectively; take the maximum value of the two sets of actual alignment offsets as the control offset, retrieve the preset allowable control offset, and calculate the alignment deviation rate.

[0012] Specifically, the calculation logic for the safety dataset in the foundation pit work area includes: After registering the image at the current acquisition time point with the reference state image, extract the real-time planar coordinates of the same monitoring point; based on the initial coordinates and real-time coordinates, calculate the cumulative horizontal displacement of each monitoring point; statistically analyze the cumulative horizontal displacement of all monitoring points in the foundation pit, and take the maximum value as the foundation pit displacement of the foundation pit working area; After registering the DSM at the current acquisition time point with the reference state DSM, the real-time elevation value of the same monitoring point is extracted; the initial elevation value of the same monitoring point is subtracted from the real-time elevation value to obtain the cumulative settlement of a single monitoring point; the cumulative settlement of all monitoring points is statistically analyzed, and the maximum value is taken as the cumulative settlement of the soil in the foundation pit operation area; Identify all cracks from the images at the current acquisition time, calculate the length and average width of each crack, and establish a crack parameter ledger; calculate the sum of the products of the length and width of all cracks, divide it by the total area of ​​the crack monitoring area, and obtain the degree of crack development.

[0013] Specifically, the process for determining risk alarm signals: By pre-setting the hazard threshold index corresponding to each risk zone, the hazard index output by each risk zone at the current collection time point is compared with the corresponding hazard threshold index. If the hazard index of any risk zone is higher than the corresponding threshold index, a risk alarm signal for the target risk zone will be generated.

[0014] Specifically, the safety monitoring report includes the comparison results of each risk zone, the hazard index, and the collected image data.

[0015] The technical effects and advantages of this invention are as follows: A zoned, multi-parameter coupled quantitative assessment system for construction risks was constructed, achieving accurate quantification and scientific hierarchical evaluation of construction safety status. For the three major risk zones of roadbed, bridges and culverts, and foundation pits, a dedicated safety dataset adapted to their core hazard sources was constructed. Multi-parameter coupling association rules and weighted calculation models were established within each zone. Combined with the standard set preset by industry specifications, the system outputs a hazard index for each zone that can be directly used for risk assessment. This completely solves the shortcomings of traditional monitoring, which relies on independent analysis of single parameters and cannot comprehensively assess the overall safety status of the construction area. It represents a core breakthrough in the transformation of construction risk from qualitative description to quantitative evaluation, significantly improving the scientificity, comprehensiveness, and reliability of construction safety risk assessment. It achieves precise and standardized safety data collection and parameter extraction across the entire road construction area, effectively solving the technical shortcomings of traditional manual inspections, such as low efficiency, strong subjectivity, and incomplete coverage. Based on the hazard source characteristics of the three core risk zones of roadbed, bridges and culverts, and foundation pits, it formulates differentiated drone data collection schemes to accurately match the monitoring accuracy requirements of each area and generate standardized image sets with spatiotemporal tags. At the same time, it establishes a standardized monitoring parameter system for the core hazard sources of each zone, realizing the automated and quantitative extraction of multi-dimensional safety parameters, eliminating the subjective errors of manual interpretation, and enabling high-frequency, non-contact data collection across the entire long-distance construction area, significantly improving monitoring efficiency and data accuracy. It has achieved intelligent early warning and closed-loop full-process control of road construction safety monitoring, which has significantly improved the timeliness, accuracy and digitalization of construction safety management. Based on the hazard index output by each zone, it realizes automated risk classification alarm through preset threshold rules, and simultaneously integrates the monitoring comparison results, hazard index and original image data of each zone to generate standardized safety monitoring reports, which are accurately pushed to the management personnel of the corresponding risk zones. Attached Figure Description

[0016] Figure 1 This is a flowchart of the road construction safety monitoring method based on UAV imagery of 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] like Figure 1 As shown, the method for road construction safety monitoring based on UAV imagery is as follows: Regional data collection: Based on the road construction scenario, risk zones are pre-divided, and differentiated collection schemes are formulated for different zones. Image data collection of each construction zone is completed by using a drone equipped with an RTK high-precision positioning module and a full-frame visible light imaging module. After preprocessing, standardized image sets with spatiotemporal labels for each zone are obtained. The risk zones include roadbed operation zone, bridge and culvert operation zone, and foundation pit operation zone; Additional notes: For each risk zone, separate drone data acquisition parameters are set to ensure the accuracy of parameter extraction for each zone matches the requirements. Drone data acquisition parameters include, but are not limited to, flight altitude, forward overlap rate, lateral overlap rate, and acquisition frequency. During the acquisition process, POS positioning data, UTC timestamp, and flight attitude data corresponding to each frame of image are acquired simultaneously to ensure accurate spatial coordinate matching of data from each zone. For example: The navigation height for the roadbed operation area is set at 80m; The ground resolution is set to 3cm / pixel; The heading / lateral overlap is set to 75% / 65%; The flight path mode is set to orthographic flight path; The data collection frequency is set to once a week, and increased to once every 3 days during the rainy season and slope excavation period.

[0019] Preprocessing: Lens distortion correction, precise matching of POS data and image frames, light and color equalization processing, invalid frame removal, and image format standardization are performed sequentially on the raw image data collected from each partition to obtain a standardized spatiotemporal sequence image dataset with spatiotemporal positioning labels for each partition.

[0020] Zonal safety parameter extraction: For the core hazard sources associated with each risk zone, the monitoring parameters that are strongly linearly correlated with the safety level of the corresponding zone are extracted from the corresponding standardized image set to construct a safety dataset for each risk zone; The safety dataset for the roadbed operation area includes the roadbed slope deviation rate, the time-series uneven settlement of the roadbed top surface, the integrity of the slope protection structure, and the compliance rate of illegal surcharges on the roadbed. Standard cross sections are extracted from the DSM digital surface model preprocessed by UAV, along the direction of the roadbed slope. Extract the actual slope height and actual horizontal projection width of the slope from the cross section, and calculate the actual slope ratio. Actual slope ratio = actual vertical height of slope / actual horizontal width of slope; The design slope rate preset in the construction design document is retrieved as the benchmark value, and the deviation rate is calculated by combining it with the actual slope rate. The slope deviation rate is then output. .

[0021] After taking the arithmetic mean of the slopes of the entire roadbed operation area, the roadbed slope deviation rate is obtained.

[0022] To elaborate further, the construction area was acquired using imaging sensors mounted on UAVs, which collected multi-view, highly overlapping visible light images. After image preprocessing and aerial triangulation to calculate the exterior orientation elements, a dense point cloud of the construction area was generated based on a dense image matching algorithm. Then, after interpolation and meshing, a digital surface model (DSM) containing ground and feature elevation information was constructed.

[0023] Extract the DSM data of the top surface of the roadbed generated by the UAV at the previous historical acquisition time point before the current acquisition time point; Equally spaced monitoring points are pre-laid on the top surface of the roadbed to obtain the elevation values ​​of each point at adjacent data collection times. For the same monitoring point, the cumulative settlement of the target monitoring point is obtained by subtracting the elevation value of the previous historical monitoring point from the elevation value of the current collection time point. ;in and These represent the elevation value of the target monitoring point at the current data collection time and the elevation value of the target monitoring point at a historical data collection time, respectively.

[0024] Calculate the absolute value of the difference in cumulative settlement between two adjacent monitoring points, divide it by the horizontal distance between the two points to obtain the uneven settlement gradient, identify the maximum uneven settlement gradient in the entire road section of the roadbed operation area to obtain the time-series uneven settlement of the roadbed top surface. Based on images collected by drones, AI semantic segmentation is used to identify slope protection structures (including but not limited to frame beams, shotcrete surfaces, retaining walls, drainage facilities, etc., which are predefined by technical personnel). By using threshold segmentation, the areas of damaged, voided, eroded, and cracked regions in the slope protection structure are identified and summed to obtain the damaged and missing area. Based on the damaged and missing area and the preset total design protection area, the integrity of the slope protection structure is output. The lowest integrity of each slope segment in the roadbed operation area is taken as the integrity of the slope protection structure.

[0025] According to construction specifications, no-stack zones are pre-defined in the images collected by drones; for example, a 2m range at the top of the slope and a 1m range at the bottom of the slope. Determine whether there is any unauthorized loading in the prohibited loading area along the roadbed slope section by section; that is, use pre-edited AI to identify unauthorized loading of earthwork, building materials, machinery, debris, etc. within this area; The compliance rate of illegal loading on the roadbed is obtained by dividing the length of continuous sections without illegal loading by the total monitored length of the slope. Supplementary explanation of the meaning of various parameters in the roadbed work area The greater the deviation rate of the roadbed slope, the more the actual slope deviates from the design requirements, the lower the slope's anti-sliding stability, and the higher the risk of landslides and instability. The greater the time-series uneven settlement of the top surface of the roadbed, the more significant the difference in roadbed settlement, the easier it is to cause cracking and overall instability of the roadbed structure, and the higher the risk of structural damage. The greater the integrity of the slope protection structure, the less damage to the protection structure, the stronger the protection and restraint of the slope, and the better the safety status; conversely, the smaller the value, the more serious the damage and loss of the protection structure, and the higher the risk of slope deformation and collapse. The higher the compliance rate of illegal loading on the roadbed, the less illegal loading in the prohibited area, the more controllable the additional stress on the slope, and the better the safety status; conversely, the smaller the value, the more illegal loading, the worse the slope's anti-sliding stability, and the higher the risk of landslide.

[0026] The safety dataset for bridge and culvert work areas includes verticality deviation values, maximum relative deformation, and alignment deviation rate. A drone equipped with an RTK high-precision positioning module and a five-lens oblique photography module was used to collect a sequence of images with full-view and high overlap (heading ≥80%, side ≥70%) for the piers / abbutments to be monitored. The acquired images are subjected to distortion correction, aerial triangulation refinement, dense point cloud matching, and 3D reconstruction to generate millimeter-precision 3D solid models of piers / abutments.

[0027] The three-dimensional coordinates of the center point A on the top surface of the foundation at the bottom of the pier are determined from the three-dimensional solid model pre-generated by the UAV. The three-dimensional coordinates of the center point B at the top of the pier ; Calculate the actual effective height of the pier column ; and horizontal offset The relative sag deviation coefficient is obtained by dividing the horizontal offset by the actual effective height. Retrieve the allowable relative sag deviation coefficient and allowable horizontal offset as preset in the standard construction documents; Substitute into the formula Obtain the verticality deviation value ;in The relative sag deviation coefficient is calculated. To allow relative sag deviation coefficient; To allow for horizontal offset, This represents the actual horizontal offset of the pier column; As a supplementary explanation, after calculating the verticality deviation of all piers / abutments in the bridge and culvert work area, the maximum value is taken as the verticality deviation value of the zone.

[0028] Using the state after the completion of the cast-in-place support / hanging basket erection and pre-stressing acceptance as the initial reference state, a high-precision DSM digital surface model of the reference state is collected by UAV.

[0029] Using the pre-stressed acceptance DSM as the benchmark model, the initial three-dimensional coordinates of each pre-set key monitoring point are extracted; key monitoring points include, but are not limited to, the main node at the top of the support, the main truss node of the hanging basket, and the mid-span control point, etc. Extract the real-time 3D coordinates of different key monitoring points within the current acquisition time point; By combining the real-time and initial three-dimensional coordinates of different key monitoring points, the relative deformation of different key monitoring points is output; Using formula The initial and real-time three-dimensional coordinates of the key monitoring points are marked as follows: and ;in = This represents the cumulative vertical settlement. , representing the cumulative horizontal displacement; and These are the maximum allowable vertical deformation and the preset design horizontal displacement span, respectively, based on the specifications.

[0030] The maximum relative deformation at each key monitoring point is taken as the highest relative deformation. That is, all cast-in-place supports and hanging baskets in the bridge and culvert operation area are calculated one by one, and the maximum value is taken as the final monitoring value of the parameter for that area.

[0031] Coordinate registration was performed on the high-precision DOM images acquired by the UAV, establishing a one-to-one correspondence between the image pixel coordinates and the construction design plane coordinates, and system error correction was completed.

[0032] Using AI target recognition, the center positions of the supports at both ends of the precast beam are located, and the actual coordinates of the first support at each end of the beam are read. Actual coordinates of the second support at the beam end ; Extract the pre-design coordinates of the two corresponding supports from the construction design drawings; that is... And calculate the actual alignment offset of the two supports respectively; Alignment offset .

[0033] Take the maximum value of the two sets of actual alignment deviations as the control deviation, retrieve the allowable control deviation preset by the construction specifications, and calculate the alignment deviation rate. The alignment deviation rate is obtained by dividing the control deviation by the allowable control offset.

[0034] For each of the precast beams erected in the bridge and culvert work area, the calculation was performed, and the maximum value was taken as the final monitoring value of the alignment deviation rate of the zone.

[0035] Supplementary explanation of the meaning of various parameters in the bridge and culvert work area. The larger the verticality deviation value, the more the verticality of the pier / abutment deviates from the design specifications, the greater the additional bending moment of the structure, and the higher the risk of overturning and fracture. The greater the maximum relative deformation, the further the support / cantilever system deviates from the stable state, and the higher the risk of system collapse during construction processes such as concrete pouring. The greater the alignment deviation rate, the more the beam support deviates from the design requirements, and the higher the risk of accidents such as support eccentricity, beam instability, and even bridge erecting machine overturning and beam falling.

[0036] The safety dataset for the foundation pit work area includes the foundation pit displacement, cumulative soil settlement, and crack development level. Along the top line of the retaining structure of the foundation pit, monitor targets are set up at equal intervals (or fixed feature points such as the external corner of the support structure or pre-embedded markers are selected), the initial plane coordinates of all monitoring points are extracted, and a reference coordinate system for displacement calculation is established.

[0037] Based on the baseline image before the excavation of the foundation pit, a monitoring point is set up at a preset distance along the top of the foundation pit slope, and the initial coordinates of all monitoring points are extracted. After registering the image at the current acquisition time point with the reference state image, extract the real-time planar coordinates of the same monitoring point; Based on the initial coordinates and real-time coordinates, calculate the cumulative horizontal displacement of each monitoring point; The cumulative horizontal displacement of all monitoring points in the foundation pit is statistically analyzed, and the maximum value is taken as the foundation pit displacement of the foundation pit operation area. Based on the baseline DSM before the foundation pit excavation, monitoring points are set up in the pre-marked excavation depth influence zone according to a set ratio (10m×10m) grid, and the initial elevation values ​​of all monitoring points are extracted. After registering the DSM at the current acquisition time point with the reference state DSM, extract the real-time elevation value of the same monitoring point; The cumulative settlement of a single monitoring point is obtained by subtracting the real-time elevation value from the initial elevation value of the same monitoring point. The cumulative settlement of all monitoring points was statistically analyzed, and the maximum value was taken as the cumulative settlement of the soil in the foundation pit operation area. Using AI semantic segmentation, all cracks are identified from the images at the current acquisition time; harmless cracks with a width of <0.2mm are removed; the length and average width of each crack are calculated, and a crack parameter ledger is established. The degree of crack development is obtained by summing the products of the length and width of all cracks and dividing them by the total area of ​​the crack monitoring region. Crack development degree ;in Let be the length of the i-th crack. Let S be the average width of the i-th crack, S be the total area of ​​the crack monitoring area, and n be the total number of cracks.

[0038] Supplementary explanation of the meaning of various parameters in the foundation pit work area. The greater the displacement of the foundation pit, the more severe the horizontal deformation of the foundation pit retaining structure, the more significant the plastic deformation of the retaining structure, and the higher the risk of overall collapse and instability of the foundation pit. The greater the cumulative settlement of the soil, the more serious the disturbance of the surrounding soil caused by the excavation of the foundation pit, the easier it is to cause soil loss behind the retaining structure, ground cracking, damage to surrounding structures, and even the higher the risk of foundation pit instability. The greater the degree of crack development, the denser and larger the cracks in the monitoring area, the more serious the damage to the integrity of the soil / retaining structure, and the higher the risk of foundation pit instability and landslide.

[0039] Quantification of construction safety level: Extract the safety dataset corresponding to each risk zone, construct parameter coupling association rules in different zones, and use the parameter coupling association rules to perform coupling analysis on the safety datasets of different risk zones, and output the hazard index of different risk zones corresponding to the road construction site; Specifically; The hazard index of the roadbed operation area is obtained by comprehensively processing the roadbed slope deviation rate, the time-series uneven settlement of the roadbed top surface, the integrity of the slope protection structure, and the compliance rate of illegal surcharges on the roadbed, combined with a preset standard set for the roadbed operation area. The specific calculation process is as follows: Hazard index of roadbed operation area ; in , , as well as These represent the deviation rate of roadbed slope, the amount of time-series uneven settlement of the top surface of the roadbed, the integrity of the slope protection structure, and the compliance rate of illegal surcharges on the roadbed, respectively. , , as well as The standards set for roadbed operation areas include the allowable deviation rate of roadbed slope specifications, the allowable settlement of roadbed top surface time unevenness specifications, the pass and integrity of slope protection structure specifications, and the pass and compliance rate of roadbed illegal surcharge specifications. , , as well as These are preset weighting coefficients, and their sum is one.

[0040] The hazard index of bridge and culvert work areas is obtained by comprehensively processing the verticality deviation value, the maximum relative deformation, and the alignment deviation rate in combination with a preset standard set for bridge and culvert work areas. The specific calculation process is as follows: Hazard index of bridge and culvert work area ; , as well as These represent the perpendicularity deviation value, the maximum relative deformation, and the alignment deviation rate, respectively. , as well as The allowable deviation value of verticality, the maximum allowable relative deformation, and the allowable alignment deviation rate are included in the standard set for bridge and culvert operation areas. , , These are preset weighting coefficients, and their sum is one.

[0041] The hazard index of the foundation pit operation area is obtained by comprehensively processing the foundation pit displacement, cumulative soil settlement, and crack development degree in combination with the preset standard set of foundation pit operation areas; The specific calculation process is as follows: Hazard index of foundation pit operation area ; , as well as These represent the foundation pit displacement, cumulative soil settlement, and crack development degree, respectively. , as well as The standards set for foundation pit operation areas include the allowable displacement of foundation pits according to specifications, the allowable cumulative settlement of soil according to specifications, and the allowable degree of crack development according to specifications. , , These are preset weighting coefficients, and their sum is one.

[0042] Safety construction level determination: Based on the hazard index output by each risk zone, combined with the preset threshold determination rules, risk alarms are completed for each risk zone, and a safety monitoring report is generated and sent to the management personnel of the respective risk zone; Specifically: That is, by pre-setting the hazard threshold index corresponding to each risk zone, the hazard index output by each risk zone at the current collection time point is compared with the corresponding hazard threshold index. If the hazard index of any risk zone is higher than the corresponding threshold index, a risk alarm signal for the target risk zone is generated. The comparison results of each risk zone, the hazard index, and the collected image data are filled into a pre-built report template to generate a safety monitoring report on road construction at the current collection time.

[0043] The above formulas are all dimensionless calculations. Dimensionless calculations can be performed using various methods such as standardization, which will not be elaborated here. The formulas are derived from software simulations based on a large amount of collected data, and the preset parameters in the formulas can be set by those skilled in the art according to the actual situation.

[0044] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, ATA hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid-state ATA hard disk.

[0045] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0046] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0047] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0048] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0049] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0050] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable ATA hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0051] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for monitoring road construction safety based on UAV imagery, characterized in that, include: Step 1: Based on the road construction scenario, risk zones are pre-divided, and differentiated data acquisition plans are developed for different zones. Image data of each construction zone is collected by drones to obtain standardized image sets with spatiotemporal labels for each zone. The risk zones include roadbed operation zone, bridge and culvert operation zone, and foundation pit operation zone; Step 2: For the core hazards associated with each risk zone, extract the monitoring parameters related to the safety level of the corresponding zone from the corresponding standardized image set, and construct the safety dataset for each risk zone; Step 3: Extract the safety dataset corresponding to each risk zone, construct parameter coupling association rules within different zones, and use the parameter coupling association rules to perform coupling analysis on the safety datasets of different risk zones, outputting the hazard index of different risk zones corresponding to the road construction site; Step 4: Based on the hazard index output by each risk zone and combined with the preset threshold judgment rules, complete the risk alarm for each risk zone, and integrate and generate a safety monitoring report to be sent to the management personnel of the respective risk zone.

2. The road construction safety monitoring method based on UAV imagery as described in claim 1, characterized in that: The security datasets for each risk partition include: The safety dataset for the roadbed work area includes the roadbed slope deviation rate, the time-series uneven settlement of the roadbed top surface, the integrity of the slope protection structure, and the compliance rate of illegal surcharges on the roadbed. The safety dataset for bridge and culvert work areas includes verticality deviation values, maximum relative deformation, and alignment deviation rate. The safety dataset for the foundation pit operation area includes the foundation pit displacement, cumulative soil settlement, and degree of crack development.

3. The road construction safety monitoring method based on UAV imagery as described in claim 2, characterized in that: Output the hazard index of different risk zones corresponding to the road construction site: The hazard index of the roadbed operation area is obtained by comprehensively processing the roadbed slope deviation rate, the time-series uneven settlement of the roadbed top surface, the integrity of the slope protection structure, and the compliance rate of illegal surcharges on the roadbed, combined with a preset standard set for the roadbed operation area. The hazard index of bridge and culvert work areas is obtained by comprehensively processing the verticality deviation value, the maximum relative deformation, and the alignment deviation rate in combination with a preset standard set for bridge and culvert work areas. The hazard index of the foundation pit operation area is obtained by comprehensively processing the foundation pit displacement, cumulative soil settlement, and crack development degree in combination with the preset standard set of foundation pit operation areas.

4. The road construction safety monitoring method based on UAV imagery as described in claim 2, characterized in that: The calculation logic for the safety dataset of the roadbed work area includes: From the DSM digital surface model of the UAV, a standard cross section is extracted along the direction of the roadbed slope; the actual slope height and actual horizontal projection width of the slope are extracted from the cross section, and the actual slope ratio is calculated. The preset design slope is retrieved as the benchmark value, and the deviation rate is calculated by combining it with the actual slope, and the slope deviation rate is output. The arithmetic mean of the slopes of the entire roadbed operation area is then taken to obtain the roadbed slope deviation rate. Extract the DSM data of the roadbed top surface generated by the UAV at the previous historical acquisition time point before the current acquisition time point; pre-deploy equally spaced monitoring points on the roadbed top surface and obtain the elevation values ​​of each point at adjacent acquisition time points; for the same monitoring point, subtract the elevation value of the previous historical acquisition time point from the elevation value of the current acquisition time point to obtain the cumulative settlement of the target monitoring point; calculate the absolute value of the difference between the cumulative settlement of two adjacent monitoring points and divide it by the horizontal distance between the two points to obtain the uneven settlement gradient; identify the maximum uneven settlement gradient in the entire roadbed operation area to obtain the temporal uneven settlement of the roadbed top surface.

5. The road construction safety monitoring method based on UAV imagery as described in claim 2, characterized in that: The calculation logic for the safety dataset of the roadbed work area also includes: Based on images collected by drones, the slope protection structure is identified through AI semantic segmentation. Through threshold segmentation, the areas of damaged, void, eroded, and cracked areas in the slope protection structure are identified and summed to obtain the damaged and missing area. Combining the damaged and missing area with the preset total design protection area, the integrity of the slope protection structure is output. In the images collected by the drone, no-stack zones are pre-defined, and the presence of stacking is determined segment by segment along the roadbed slope. The length of continuous segments without illegal stacking is counted and divided by the total monitored length of the slope to obtain the compliance rate of illegal stacking on the roadbed.

6. The road construction safety monitoring method based on UAV imagery as described in claim 5, characterized in that: The calculation logic for the safety dataset in the bridge and culvert work area includes: From the three-dimensional solid model pre-generated by the UAV, locate the three-dimensional coordinates of the center point A of the bottom pier cap and the three-dimensional coordinates of the center point B of the top of the pier. Calculate the actual effective height and horizontal offset of the pier column; divide the horizontal offset by the actual effective height to obtain the relative sag deviation coefficient. Retrieve the preset allowable relative sag deviation coefficient and allowable horizontal offset; Substitute into the formula Obtain the verticality deviation value ;in The relative sag deviation coefficient is calculated. To allow relative sag deviation coefficient; To allow for horizontal offset, This represents the actual horizontal offset of the pier column.

7. The road construction safety monitoring method based on UAV imagery as described in claim 6, characterized in that: The safety dataset calculation logic for bridge and culvert work areas also includes: Using the pre-compression acceptance DSM as the benchmark model, the initial three-dimensional coordinates of each pre-set key monitoring point are extracted; the real-time three-dimensional coordinates of different key monitoring points within the current acquisition time point are extracted; the real-time three-dimensional coordinates and initial three-dimensional coordinates of different key monitoring points are combined to output the relative deformation of different key monitoring points; the maximum value of the relative deformation of each key monitoring point is taken as the highest relative deformation. Locate the center position of the supports at both ends of the precast beam, and read the actual coordinates of the first support at the beam end and the actual coordinates of the second support at the beam end respectively; Extract the pre-designed coordinates of the two supports and calculate the actual alignment offset of the two supports respectively; take the maximum value of the two sets of actual alignment offsets as the control offset, retrieve the preset allowable control offset, and calculate the alignment deviation rate.

8. The road construction safety monitoring method based on UAV imagery as described in claim 7, characterized in that: The calculation logic for the safety dataset of the foundation pit work area includes: After registering the image at the current acquisition time point with the reference state image, extract the real-time planar coordinates of the same monitoring point; based on the initial coordinates and real-time coordinates, calculate the cumulative horizontal displacement of each monitoring point; statistically analyze the cumulative horizontal displacement of all monitoring points in the foundation pit, and take the maximum value as the foundation pit displacement of the foundation pit working area; After registering the DSM at the current acquisition time point with the reference state DSM, the real-time elevation value of the same monitoring point is extracted; the initial elevation value of the same monitoring point is subtracted from the real-time elevation value to obtain the cumulative settlement of a single monitoring point; the cumulative settlement of all monitoring points is statistically analyzed, and the maximum value is taken as the cumulative settlement of the soil in the foundation pit operation area; Identify all cracks from the images at the current acquisition time, calculate the length and average width of each crack, and establish a crack parameter ledger; calculate the sum of the products of the length and width of all cracks, divide it by the total area of ​​the crack monitoring area, and obtain the degree of crack development.

9. The road construction safety monitoring method based on UAV imagery as described in claim 3, characterized in that: The process of determining risk alarm signals: By pre-setting the hazard threshold index corresponding to each risk zone, the hazard index output by each risk zone at the current collection time point is compared with the corresponding hazard threshold index. If the hazard index of any risk zone is higher than the corresponding threshold index, a risk alarm signal for the target risk zone will be generated.

10. The road construction safety monitoring method based on UAV imagery as described in claim 3, characterized in that: The safety monitoring report includes the comparison results of each risk zone, the hazard index, and the collected image data.

Images