Building construction site inspection management system and method based on AI

Abstract

The invention discloses a building construction site inspection management system and method based on AI, and relates to the technical field of construction inspection. Collecting detection data through inspection equipment in the inspection process; establishing a three-dimensional point cloud model based on the detection data after data cleaning; constructing a space coordinate system in the three-dimensional point cloud model, and analyzing the space position information of the inspection target; analyzing an optimal detection road section of the inspection equipment according to the spatial position information of the inspection equipment and the inspection target; and searching and adjusting the real-time inspection angle of the inspection equipment based on the inspection parameter sequence, further adjusting the real-time inspection angle corresponding to the real-time detection data which does not meet the condition, and updating the real-time inspection angle. The quality of detection data is improved, and the overall analysis efficiency of the inspection system is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of construction inspection, and specifically to an AI-based building construction site inspection management system and method. Background Art

[0002] In modern urban construction, the safety of building construction sites has always been a focus of attention. In order to ensure the safety of workers and improve construction efficiency, more and more building construction sites have adopted intelligent building construction site safety inspection systems. The intelligent building construction site safety inspection system can comprehensively monitor the safety status of the building construction site and timely discover and solve potential safety hazards.

[0003] The data acquisition devices of the existing safety inspection systems are often in the data acquisition stage from the start to the end of the inspection. Some of the collected data has low quality, resulting in low analysis value and easy to cause data accumulation in the database, reducing the overall analysis efficiency of the inspection system.

[0004] Therefore, the present invention discloses an AI-based building construction site inspection management system and method to solve the above problems. Summary of the Invention

[0005] The purpose of the present invention is to provide an AI-based building construction site inspection management system and method to solve the problems raised in the prior art.

[0006] To achieve the above purpose, the present invention provides the following technical solution: An AI-based building construction site inspection management method, the method includes the following steps:

[0007] S1: Plan the operation path of the inspection device according to the inspection task; control the inspection device to conduct safety inspections and keep the detection heights of the data collectors of the inspection device consistent, and collect detection data through the inspection device during the inspection process;

[0008] S2: Clean the detection data during the operation of the inspection device, and establish a three-dimensional point cloud model based on the detection data after data cleaning; construct a spatial coordinate system in the three-dimensional point cloud model and analyze the spatial position information of the inspection target;

[0009] S3: Analyze the best detection section of the inspection device according to the spatial position information of the inspection device and the inspection target;

[0010] S4: Form a group of inspection parameters according to the spatial position information of the best detection section and the corresponding inspection parameters; search and adjust the real-time inspection angle of the inspection device based on the inspection parameter sequence, further adjust the real-time inspection angle corresponding to the real-time detection data that does not meet the conditions, and update it.

[0011] According to the above solution, in S1, the data collector of the inspection device includes a scanning instrument and a camera; the detection height includes the detection height of the scanning instrument and the detection height of the camera; the detection height of the scanning instrument represents the vertical distance between the scanning instrument and the ground where the inspection device is located; the detection height of the camera represents the vertical distance between the camera and the ground where the inspection device is located; the detection data includes three-dimensional point cloud data and image data.

[0012] According to the above solution, in S2, it includes the following content:

[0013] S201: Extract the three-dimensional point cloud data in the detection data, and remove the three-dimensional point cloud data corresponding to the moving objects during the scanning process;

[0014] S202: Use the three-dimensional point cloud data after data cleaning to establish a three-dimensional point cloud model; in the construction of any plane of the three-dimensional point cloud model, use the RANSA algorithm to analyze the plane that meets the preset requirements and has the largest number of points, and record it as the optimal plane; record the point cloud within the optimal plane as the optimal point cloud, and count the number of point clouds within the radius threshold for any optimal point cloud. If the number of point clouds within the radius threshold is greater than or equal to the quantity threshold, save the optimal point cloud; if the number of point clouds within the radius threshold is less than the quantity threshold, remove the optimal point cloud;

[0015] Through the scanning instrument of the inspection device, the three-dimensional space information and surface model on the operation path of the inspection device can be quickly obtained, and a three-dimensional point cloud model can be constructed; analyzing the number of point clouds for the optimal point cloud can further improve the accuracy of the three-dimensional point cloud model; with the support of the digital terrain model and high-resolution images, high-efficiency and high-precision line inspection applications can be realized;

[0016] S203: Construct a spatial coordinate system in the three-dimensional point cloud model, extract the spatial coordinates of the boundary of the inspection target in the three-dimensional point cloud model, analyze the spatial coordinates of the centroid point of the inspection target, and record them as (a, b, c); where a, b, and c respectively represent the values of the centroid point of the inspection target on the spatial coordinate axes in the three-dimensional point cloud model.

[0017] According to the above solution, in S3, it includes the following content:

[0018] S301: Record the spatial coordinates of the camera in the inspection device as (x, y, z); where x, y, and z represent the spatial coordinate variables in the three-dimensional point cloud model; extract the spatial coordinates corresponding to the camera when the inspection device is on the operation path, and analyze the actual maximum observation area of the inspection target corresponding to the camera at different spatial coordinates, and record it as S(x, y, z); the actual maximum observation area represents the largest surface area size that can be projected and displayed in the image data under the corresponding inspection target volume; one set of spatial coordinates of a camera corresponds to one actual maximum observation area;

[0019] S302: Obtain the focal length f of the camera of the inspection device and the actual size P of a single pixel, analyze the distance between the camera of the inspection device and the centroid of the inspection target, denoted as L(x, y, z); calculate the pixel area PI(x, y, z) of the inspection target displayed in the image data at different spatial coordinates of the camera according to each actual maximum observation area; the specific calculation formula is:

[0020] L(x,y,z)=[(x - a) 2 +(y - b) 2 +(z - c) 2 1 / 2 ;

[0021] PI(x,y,z)=S(x,y,z)×f 2 ÷[L(x,y,z)×P] 2 ;

[0022] S303: Extract the pixel areas of the actual maximum observation areas corresponding to the camera at different spatial coordinates when the inspection device is on the operation path, and form a pixel area set; analyze the analysis reference values of each pixel area in the pixel area set:

[0023] ARV i =[1÷(2πα 2 ) 1 / 2 ×exp[-(PI i -β) 2 ÷(2×α 2 )];

[0024] where ARV i represents the i-th pixel area in the pixel area set, β represents the mean value of the pixel area set, and α represents the standard deviation of the pixel area set;

[0025] S304: Set an analysis reference value threshold for the analysis reference values, extract the spatial coordinates of the camera corresponding to the pixel areas where the analysis reference values are equal to the analysis reference value threshold, denoted as the spatial boundary, and record the operation path of the inspection device between the two spatial boundaries as the best detection section. When the spatial coordinates of the camera of the inspection device belong to the best detection section, turn on the camera to perform the detection task; the pixel area corresponding to the spatial coordinates in the best detection section is greater than the pixel area threshold.

[0026] By analyzing the maximum observation area of the inspection target under different position information and the distance between the inspection target and the inspection device, analyzing the pixel area in the image data, calculating the analysis reference value according to the pixel area, obtaining the best detection section, reducing the amount of detection data collected, improving the quality of the detection data, avoiding data accumulation in the database, and improving the overall analysis efficiency of the inspection system; ​

[0027] According to the above solution, in S4, the following contents are included:

[0028] S401: Extract the spatial coordinates of the best detection section and the inspection angle θ corresponding to the spatial coordinates of the best detection section in the actual maximum observation area of the inspection target, and form an inspection parameter group (x, y, z, θ); the inspection angle is the included angle between the horizontal plane where the camera is located and the shooting direction of the camera;

[0029] S402: Form an inspection parameter sequence with all inspection parameter groups according to the operation direction of the inspection device, and store the inspection parameter sequence in the database; after the camera is turned on, use the spatial coordinates of the camera to find the corresponding inspection parameter group in the database, and adjust the real-time inspection angle of the inspection device camera according to the inspection parameter group;

[0030] S403: Extract the real-time image data collected based on the inspection parameter group, and analyze the real-time pixel area of the inspection target in the real-time image data. The real-time pixel area is equal to the actual size of a single pixel multiplied by the number of pixels of the inspection target; if the real-time pixel area is greater than or equal to the pixel area multiplied by the preset proportional coefficient, no processing is performed; if the real-time pixel area is less than the pixel area multiplied by the preset proportional coefficient, adjust the real-time inspection angle until the real-time pixel area is greater than or equal to the pixel area multiplied by the preset proportional coefficient; update the adjusted real-time inspection angle to the inspection parameter group.

[0031] This application forms an inspection parameter group and an inspection parameter sequence, uses the spatial coordinates of the camera to find the corresponding inspection parameter group in the database, and adjusts the real-time inspection angle of the inspection device camera according to the inspection parameter group; it can effectively improve the acquisition speed of the real-time inspection angle and further improve work efficiency; further analyze the real-time image data and adjust the real-time inspection angle, which can timely detect the situation of low data quality and further improve the quality of subsequent data collection.

[0032] Another aspect of this application provides an AI-based inspection management system for construction sites. The system is implemented by applying the above-mentioned AI-based inspection management method for construction sites. The system includes an inspection data collection module, a spatial model construction module, an optimal section analysis model, and a parameter adjustment and update module;

[0033] The inspection data collection module is used to plan the operation path of the inspection device according to the inspection task, control the inspection device to perform safety inspections and keep the detection heights of the data collectors of the inspection device consistent; collect detection data through the inspection device during the inspection process;

[0034] The spatial model construction module is used to clean the detection data during the operation of the inspection equipment, and establish a three-dimensional point cloud model based on the detection data after data cleaning; construct a spatial coordinate system in the three-dimensional point cloud model, and analyze the spatial position information of the inspection target;

[0035] The optimal route analysis model is used to analyze the optimal detection route of the inspection equipment according to the spatial position information of the inspection equipment and the inspection target;

[0036] The parameter adjustment and update module is used to form an inspection parameter group according to the spatial position information of the optimal detection route and the corresponding inspection parameters; search and adjust the real-time inspection angle of the inspection equipment based on the inspection parameter sequence, further adjust the real-time inspection angle corresponding to the real-time detection data that does not meet the conditions, and perform an update.

[0037] According to the above solution, the inspection data acquisition module includes an operation route planning unit and an acquisition control unit;

[0038] The operation route planning unit is used to plan the operation route of the inspection equipment according to the inspection task;

[0039] The acquisition control unit is used to control the inspection equipment to perform safe inspections and keep the detection heights of the data collectors of the inspection equipment consistent, and collect detection data through the inspection equipment during the inspection process; the data collectors of the inspection equipment include scanning instruments and cameras; the detection heights include the detection heights of the scanning instruments and the cameras; the detection height of the scanning instrument represents the vertical distance between the scanning instrument and the ground where the inspection equipment is located; the detection height of the camera represents the vertical distance between the camera and the ground where the inspection equipment is located; the detection data includes three-dimensional point cloud data and image data.

[0040] According to the above solution, the spatial model construction module includes a three-dimensional point cloud model construction unit and an inspection target spatial analysis unit;

[0041] The three-dimensional point cloud model construction unit is used to extract the three-dimensional point cloud data in the detection data, and remove the three-dimensional point cloud data corresponding to the moving objects during the scanning process; use the three-dimensional point cloud data after data cleaning to establish a three-dimensional point cloud model;

[0042] The inspection target spatial analysis unit is used to construct a spatial coordinate system in the three-dimensional point cloud model, extract the spatial coordinates of the boundary of the inspection target in the three-dimensional point cloud model, and analyze the spatial coordinates of the center of gravity point of the inspection target.

[0043] According to the above solution, the optimal route analysis model includes an inspection target image analysis unit and an optimal detection route judgment unit;

[0044] The inspection target image analysis unit is used to analyze the actual maximum observation area of the inspection target corresponding to different spatial coordinates of the camera; according to each actual maximum observation area, calculate the pixel area of the inspection target displayed in the image data at different spatial coordinates of the camera.

[0045] The best detection section judgment unit is used to analyze the analysis reference value of each pixel area in the pixel area set; extract the spatial coordinates of the camera corresponding to the pixel area whose analysis reference value is equal to the analysis reference value threshold and record them as spatial boundaries, and record the operation path of the inspection device between the two spatial boundaries as the best detection section.

[0046] According to the above solution, the parameter adjustment and update module includes a real-time control unit and a parameter update unit.

[0047] The real-time control unit is used to form an inspection parameter sequence with all inspection parameter groups in the operation direction of the inspection device, store the inspection parameter sequence in the database; after the camera is turned on, use the spatial coordinates of the camera to find the corresponding inspection parameter group in the database, and adjust the real-time inspection angle of the inspection device camera according to the inspection parameter group.

[0048] The parameter update unit is used to analyze the real-time pixel area of the inspection target in the real-time image data, and the real-time pixel area is equal to the actual size of a single pixel multiplied by the number of pixels of the inspection target; if the real-time pixel area is greater than or equal to the pixel area multiplied by the preset proportionality coefficient, no processing is performed; if the real-time pixel area is less than the pixel area multiplied by the preset proportionality coefficient, adjust the real-time inspection angle until the real-time pixel area is greater than or equal to the pixel area multiplied by the preset proportionality coefficient; update the adjusted real-time inspection angle to the inspection parameter group.

[0049] Compared with the prior art, the beneficial effects of the present invention are as follows: Through the scanning instrument of the inspection device, three-dimensional spatial information and surface models on the operation path of the inspection device can be quickly obtained to construct a three-dimensional point cloud model; Analyzing the number of points in the optimal point cloud can further improve the accuracy of the three-dimensional point cloud model; With the support of a digital terrain model and high-resolution images, high-efficiency and high-precision line inspection applications can be realized; By analyzing the maximum observation area of the inspection target under different position information and the distance between the inspection target and the inspection device, analyzing the pixel area in the image data, calculating and analyzing the reference value according to the pixel area, the optimal detection section can be obtained, reducing the amount of detection data collected, improving the quality of the detection data, avoiding data accumulation in the database, and improving the overall analysis efficiency of the inspection system; In this application, by forming an inspection parameter group and an inspection parameter sequence, the corresponding inspection parameter group is searched in the database using the spatial coordinates of the camera, and the real-time inspection angle of the inspection device camera is adjusted according to the inspection parameter group; It can effectively improve the acquisition speed of the real-time inspection angle and further improve work efficiency; Further analyzing the real-time image data and adjusting the real-time inspection angle can timely detect the situation of low data quality and further improve the quality of subsequent data collection. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The accompanying drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention, and do not constitute a limitation to the present invention. In the accompanying drawings:

[0051] Figure 1 is a schematic flow chart of a method for inspecting and managing a construction site based on AI according to the present invention;

[0052] Figure 2 is a schematic structural diagram of a system for inspecting and managing a construction site based on AI according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0053] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0054] Please refer to Figure 1 , the present invention provides a technical solution: A method for inspecting and managing a construction site based on AI, the method includes the following steps:

[0055] S1: Plan the operation path of the inspection device according to the inspection task; control the inspection device to conduct safety inspections and keep the detection heights of the data collectors of the inspection device consistent. During the inspection process, collect detection data through the inspection device;

[0056] In S1, the data collectors of the inspection device include scanning instruments and cameras; the detection heights include the detection height of the scanning instrument and the detection height of the camera; the detection height of the scanning instrument represents the vertical distance between the scanning instrument and the ground where the inspection device is located; the detection height of the camera represents the vertical distance between the camera and the ground where the inspection device is located; the detection data includes three-dimensional point cloud data and image data.

[0057] S2: Clean the detection data during the operation of the inspection device, and establish a three-dimensional point cloud model based on the detection data after data cleaning; construct a spatial coordinate system in the three-dimensional point cloud model, and analyze the spatial position information of the inspection target;

[0058] In S2, it includes the following contents:

[0059] S201: Extract the three-dimensional point cloud data in the detection data, and remove the three-dimensional point cloud data corresponding to moving objects during the scanning process;

[0060] S202: Use the three-dimensional point cloud data after data cleaning to establish a three-dimensional point cloud model; in the construction of any plane in the three-dimensional point cloud model, use the RANSA algorithm to analyze the plane that meets the preset requirements and has the largest number of points, which is recorded as the optimal plane; record the point cloud in the optimal plane as the optimal point cloud, and count the number of point clouds within the radius threshold for any optimal point cloud. If the number of point clouds within the radius threshold is greater than or equal to the quantity threshold, save the optimal point cloud; if the number of point clouds within the radius threshold is less than the quantity threshold, remove the optimal point cloud;

[0061] S203: Construct a spatial coordinate system in the three-dimensional point cloud model, extract the spatial coordinates of the boundary of the inspection target in the three-dimensional point cloud model, analyze the spatial coordinates of the centroid point of the inspection target, and record it as (a, b, c); where a, b, and c respectively represent the values of the centroid point of the inspection target on the spatial coordinate axes in the three-dimensional point cloud model.

[0062] S3: Analyze the best detection section of the inspection device according to the spatial position information of the inspection device and the inspection target;

[0063] In S3, it includes the following contents:

[0064] S301: Denote the spatial coordinates of the camera in the inspection device as (x, y, z), where x, y, and z represent the spatial coordinate variables in the three-dimensional point cloud model. Extract the spatial coordinates corresponding to the camera when the inspection device is on the running path, and analyze the actual maximum observation area of the inspection target corresponding to the camera at different spatial coordinates, denoted as S(x, y, z). The actual maximum observation area represents the maximum surface area size that can be projected and displayed in the image data under the corresponding inspection target volume. One set of spatial coordinates of a camera corresponds to one actual maximum observation area.

[0065] S302: Obtain the focal length f of the camera of the inspection device and the actual size P of a single pixel, and analyze the distance between the camera of the inspection device and the centroid of the inspection target, denoted as L(x, y, z). Calculate the pixel area PI(x, y, z) of the inspection target displayed in the image data at different spatial coordinates of the camera according to each actual maximum observation area. The specific calculation formula is:

[0066] L(x,y,z)=[(x - a) 2 +(y - b) 2 +(z - c) 2 1 / 2 ;

[0067] PI(x,y,z)=S(x,y,z)×f 2 ÷[L(x,y,z)×P] 2 ;

[0068] Example 1: In this example, the focal length f of the camera is 0.05 m; the physical size of the camera sensor is 6.4 mm × 4.8 mm; the resolution is 640 × 480 pixels; the distance L(x, y, z) between the camera of the inspection device and the centroid of the inspection target is 5 m; the actual maximum observation area S(x, y, z) is 40 cm 2 ;

[0069] Then P = sensor width ÷ resolution width = 6.4 mm ÷ 640 = 0.01 mm = 10 -5 m;

[0070] P 2 =10 -10 m 2 ; L(x,y,z) 2 =25 m 2 ; f 2 =0.0025 m 2 ; S(x,y,z)=0.0004 m 2 ;

[0071] PI(x,y,z)=0.0004×0.0025÷[10​-10 × 25] = 400 pixels 2 ;

[0072] S303: When extracting the inspection device on the running path, the pixel area of the actual maximum observation area corresponding to the camera at different spatial coordinates is formed into a pixel area set; analyze the analysis reference value of each pixel area in the pixel area set:

[0073] ARV i = [1 ÷ (2πα 2 ) 1 / 2 × exp[-(PI i - β) 2 ÷ (2 × α 2 )];

[0074] Where ARV i represents the i-th pixel area in the pixel area set, β represents the mean value of the pixel area set, and α represents the standard deviation of the pixel area set;

[0075] S304: Set an analysis reference value threshold for the analysis reference value, extract the spatial coordinates of the camera corresponding to the pixel area where the analysis reference value is equal to the analysis reference value threshold and record it as the spatial boundary, and record the running path of the inspection device between the two spatial boundaries as the best detection section. When the spatial coordinates of the camera of the inspection device belong to the best detection section, turn on the camera to perform the detection task; the pixel area corresponding to the spatial coordinates in the best detection section is greater than the pixel area threshold.

[0076] S4: Form an inspection parameter group according to the spatial position information of the best detection section and the corresponding inspection parameters; based on the inspection parameter sequence, find and adjust the real-time inspection angle of the inspection device, further adjust the real-time inspection angle corresponding to the real-time detection data that does not meet the conditions, and update it.

[0077] In S4, it includes the following content:

[0078] S401: Extract the spatial coordinates of the best detection section and the inspection angle θ corresponding to the spatial coordinates of the best detection section in the actual maximum observation area of the inspection target, and form an inspection parameter group (x, y, z, θ); the inspection angle is the angle between the horizontal plane where the camera is located and the camera shooting direction;

[0079] S402: Form an inspection parameter sequence with all inspection parameter groups according to the running direction of the inspection device, store the inspection parameter sequence in the database; after turning on the camera, use the spatial coordinates of the camera to find the corresponding inspection parameter group in the database, and adjust the real-time inspection angle of the inspection device camera according to the inspection parameter group;

[0080] S403: Extract the real-time image data collected based on the patrol parameter group, analyze the real-time pixel area of the patrol target in the real-time image data, where the real-time pixel area is equal to the actual size of a single pixel multiplied by the number of pixels of the patrol target; if the real-time pixel area is greater than or equal to the pixel area multiplied by the preset proportionality coefficient, no processing is performed; if the real-time pixel area is less than the pixel area multiplied by the preset proportionality coefficient, adjust the real-time patrol angle until the real-time pixel area is greater than or equal to the pixel area multiplied by the preset proportionality coefficient; update the adjusted real-time patrol angle to the patrol parameter group.

[0081] Please refer to Figure 2 , the present invention provides a technical solution: an AI-based patrol management system for construction sites, which includes a patrol data collection module, a spatial model construction module, an optimal route analysis model, and a parameter adjustment and update module;

[0082] The patrol data collection module is used to plan the operation path of the patrol device according to the patrol task, control the patrol device to perform safety patrols and keep the detection heights of the data collectors of the patrol device consistent; collect detection data through the patrol device during the patrol process;

[0083] The spatial model construction module is used to clean the detection data during the operation of the patrol device, and establish a three-dimensional point cloud model based on the detection data after data cleaning; construct a spatial coordinate system in the three-dimensional point cloud model and analyze the spatial position information of the patrol target;

[0084] The optimal route analysis model is used to analyze the optimal detection route of the patrol device according to the spatial position information of the patrol device and the patrol target;

[0085] The parameter adjustment and update module is used to form a patrol parameter group according to the spatial position information of the optimal detection route and the corresponding patrol parameter sequence; find and adjust the real-time patrol angle of the patrol device based on the patrol parameter sequence, further adjust the real-time patrol angle corresponding to the real-time detection data that does not meet the conditions, and perform an update.

[0086] The patrol data collection module includes an operation path planning unit and a collection control unit;

[0087] The operation path planning unit is used to plan the operation path of the patrol device according to the patrol task;

[0088] The acquisition control unit is used to control the inspection equipment to conduct safety inspections and keep the detection heights of the data collectors of the inspection equipment consistent. During the inspection process, detection data is collected through the inspection equipment; the data collectors of the inspection equipment include scanning instruments and cameras; the detection heights include the detection height of the scanning instrument and the detection height of the camera; the detection height of the scanning instrument represents the vertical distance between the scanning instrument and the ground where the inspection equipment is located; the detection height of the camera represents the vertical distance between the camera and the ground where the inspection equipment is located; the detection data includes three-dimensional point cloud data and image data.

[0089] The spatial model construction module includes a three-dimensional point cloud model construction unit and an inspection target space analysis unit;

[0090] The three-dimensional point cloud model construction unit is used to extract the three-dimensional point cloud data in the detection data and remove the three-dimensional point cloud data corresponding to moving objects during the scanning process; using the three-dimensional point cloud data after data cleaning, a three-dimensional point cloud model is established;

[0091] The inspection target space analysis unit is used to construct a spatial coordinate system in the three-dimensional point cloud model, extract the spatial coordinates of the boundaries of the inspection target in the three-dimensional point cloud model, and analyze the spatial coordinates of the centroid point of the inspection target.

[0092] The optimal section analysis model includes an inspection target image analysis unit and an optimal detection section judgment unit;

[0093] The inspection target image analysis unit is used to analyze the actual maximum observation area of the inspection target corresponding to different spatial coordinates of the camera; according to each actual maximum observation area, calculate the pixel area of the inspection target displayed in the image data at different spatial coordinates of the camera;

[0094] The optimal detection section judgment unit is used to analyze the analysis reference value of each pixel area in the pixel area set; extract the spatial coordinates of the camera corresponding to the pixel area whose analysis reference value is equal to the analysis reference value threshold and record it as the spatial boundary, and record the operation path of the inspection equipment between the two spatial boundaries as the optimal detection section.

[0095] The parameter adjustment and update module includes a real-time control unit and a parameter update unit;

[0096] The real-time control unit is used to form an inspection parameter sequence for all inspection parameter groups according to the operation direction of the inspection equipment, store the inspection parameter sequence in the database; after turning on the camera, use the spatial coordinates of the camera to find the corresponding inspection parameter group in the database, and adjust the real-time inspection angle of the inspection equipment camera according to the inspection parameter group;

[0097] The parameter update unit is used to analyze the real-time pixel area of the inspection target in the real-time image data. The real-time pixel area is equal to the actual size of a single pixel multiplied by the number of pixels of the inspection target. If the real-time pixel area is greater than or equal to the pixel area multiplied by the preset proportionality coefficient, no processing is performed. If the real-time pixel area is less than the pixel area multiplied by the preset proportionality coefficient, the real-time inspection angle is adjusted until the real-time pixel area is greater than or equal to the pixel area multiplied by the preset proportionality coefficient. The adjusted real-time inspection angle is updated to the inspection parameter group.

[0098] It should be noted that in this document, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device.

[0099] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and can be implemented in other specific forms without departing from the spirit or basic characteristics of the present invention. Therefore, in any regard, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present invention. Any reference signs in the claims should not be regarded as limiting the claimed rights.

Claims

1. An AI-based inspection and management method for construction sites, characterized in that, The method includes the following steps: S1: Plan the operation path of the inspection device according to the inspection task; control the inspection device to perform safety inspections and keep the detection heights of the data collectors of the inspection device consistent. During the inspection process, collect detection data through the inspection device; S2: Clean the detection data during the operation of the inspection device, and establish a three-dimensional point cloud model based on the detection data after data cleaning; construct a spatial coordinate system in the three-dimensional point cloud model, and analyze the spatial position information of the inspection target; S3: Analyze the best detection section of the inspection device according to the spatial position information of the inspection device and the inspection target; S4: Form a group of inspection parameters according to the spatial position information of the best detection section and the corresponding inspection parameters; Based on the inspection parameter sequence, find and adjust the real-time inspection angle of the inspection device, further adjust the real-time inspection angle corresponding to the real-time detection data that does not meet the conditions, and update it.

2. The method for inspecting and managing a construction site based on AI according to claim 1, characterized in that: In S1, the data collectors of the inspection device include a scanning instrument and a camera; the detection heights include the scanning instrument detection height and the camera detection height; the scanning instrument detection height represents the vertical distance between the scanning instrument and the ground where the inspection device is located; the camera detection height represents the vertical distance between the camera and the ground where the inspection device is located; the detection data includes three-dimensional point cloud data and image data.

3. The method for inspecting and managing a construction site based on AI according to claim 2, characterized in that: In S2, it includes the following contents: S201: Extract the three-dimensional point cloud data in the detection data, and remove the three-dimensional point cloud data corresponding to the moving objects during the scanning process; S202: Use the three-dimensional point cloud data after data cleaning to establish a three-dimensional point cloud model; in any plane construction of the three-dimensional point cloud model, use the RANSA algorithm to analyze the plane that meets the preset requirements and has the largest number of points, and record it as the optimal plane; record the point cloud in the optimal plane as the optimal point cloud, and count the number of point clouds within the radius threshold for any optimal point cloud. If the number of point clouds within the radius threshold is greater than or equal to the quantity threshold, save the optimal point cloud; if the number of point clouds within the radius threshold is less than the quantity threshold, remove the optimal point cloud; S203: Construct a spatial coordinate system in the three-dimensional point cloud model, extract the spatial coordinates of the boundary of the inspection target in the three-dimensional point cloud model, analyze the spatial coordinates of the center of gravity point of the inspection target, and record it as (a, b, c); where a, b, and c respectively represent the values of the center of gravity point of the inspection target on the spatial coordinate axes in the three-dimensional point cloud model.

4. The method for inspecting and managing a construction site based on AI according to claim 3, characterized in that: In S3, it includes the following contents: S301: Record the spatial coordinates of the camera in the inspection device as (x, y, z); where x, y, and z represent the spatial coordinate variables in the three-dimensional point cloud model; extract the spatial coordinates corresponding to the camera when the inspection device is on the operation path, and analyze the actual maximum observation area of the inspection target corresponding to the camera at different spatial coordinates, and record it as S(x, y, z); the actual maximum observation area represents the maximum surface area size that can be projected and displayed in the image data under the corresponding inspection target volume; one set of spatial coordinates of a camera corresponds to one actual maximum observation area; S302: Obtain the focal length f of the camera of the inspection device and the actual size P of a single pixel, analyze the distance between the camera of the inspection device and the centroid of the inspection target, denoted as L(x, y, z); calculate the pixel area PI(x, y, z) of the inspection target displayed in the image data at different spatial coordinates of the camera according to each actual maximum observation area; the specific calculation formula is: L(x,y,z) = [(x - a) 2 + (y - b) 2 + (z - c) 2 1 / 2 ;​ PI(x, y, z) = S(x, y, z) × f 2 ÷ [L(x, y, z) × P] 2 ; S303: Extract the pixel areas of the actual maximum observation areas corresponding to the camera at different spatial coordinates when the inspection device is on the operation path, and form a pixel area set; analyze the analysis reference values of each pixel area in the pixel area set: ARV i = [1÷(2πα 2 ) 1 / 2 ×exp[-(PI i -β) 2 ÷(2×α 2 )]; where ARV i represents the i-th pixel area in the set of pixel areas, β represents the mean of the set of pixel areas, and α represents the standard deviation of the set of pixel areas; S304: Set an analysis reference value threshold for the analysis reference values, extract the spatial coordinates of the camera corresponding to the pixel areas whose analysis reference values are equal to the analysis reference value threshold and denote them as spatial boundaries, and denote the operation path of the inspection device between the two spatial boundaries as the best detection section. When the spatial coordinates of the camera of the inspection device belong to the best detection section, turn on the camera to perform the detection task; the pixel area corresponding to the spatial coordinates in the best detection section is greater than the pixel area threshold.

5. The method for inspecting and managing a construction site based on AI according to claim 4, wherein: In S4, it includes the following content: S401: Extract the spatial coordinates of the best detection section and the inspection angle θ corresponding to the spatial coordinates of the best detection section in the actual maximum observation area of the inspection target, and form an inspection parameter group (x, y, z, θ); the inspection angle is the angle between the horizontal plane where the camera is located and the shooting direction of the camera; S402: Form an inspection parameter sequence with all the inspection parameter groups in the operation direction of the inspection device, and store the inspection parameter sequence in the database; after turning on the camera, use the spatial coordinates of the camera to find the corresponding inspection parameter group in the database, and adjust the real-time inspection angle of the camera of the inspection device according to the inspection parameter group; S403: Extract the real-time image data collected based on the inspection parameter group, analyze the real-time pixel area of the inspection target in the real-time image data, and the real-time pixel area is equal to the actual size of a single pixel multiplied by the number of pixels of the inspection target; if the real-time pixel area is greater than or equal to the pixel area multiplied by the preset proportionality coefficient, no processing is performed; if the real-time pixel area is less than the pixel area multiplied by the preset proportionality coefficient, adjust the real-time inspection angle until the real-time pixel area is greater than or equal to the pixel area multiplied by the preset proportionality coefficient; update the adjusted real-time inspection angle to the inspection parameter group.

6. An AI-based inspection management system for construction sites, which is implemented by applying the AI-based inspection management method for construction sites according to any one of claims 1-5, characterized in that The system includes an inspection data acquisition module, a spatial model construction module, a best section analysis model, and a parameter adjustment and update module; The inspection data acquisition module is used to plan the operation path of the inspection device according to the inspection task, control the inspection device to perform safe inspections and keep the detection heights of the data collectors of the inspection device consistent; Collect detection data through the inspection device during the inspection process; The spatial model construction module is used to perform data cleaning on the detection data during the operation of the inspection device, establish a three-dimensional point cloud model based on the detection data after data cleaning; construct a spatial coordinate system in the three-dimensional point cloud model and analyze the spatial position information of the inspection target; The optimal section analysis model is used to analyze the optimal detection section of the inspection device according to the spatial position information of the inspection device and the inspection target; The parameter adjustment and update module is used to form an inspection parameter group according to the spatial position information of the optimal detection section and the corresponding inspection parameters; Based on the inspection parameter sequence, find and adjust the real-time inspection angle of the inspection device, further adjust the real-time inspection angle corresponding to the real-time detection data that does not meet the conditions, and perform an update.

7. An AI-based inspection and management system for construction sites according to claim 6, characterized in that: The inspection data acquisition module includes an operation path planning unit and an acquisition control unit; The operation path planning unit is used to plan the operation path of the inspection device according to the inspection task; The acquisition control unit is used to control the inspection device to perform a safety inspection and keep the detection heights of the data collectors of the inspection device consistent. During the inspection process, detection data is collected through the inspection device; the data collectors of the inspection device include a scanning instrument and a camera; the detection heights include the detection height of the scanning instrument and the detection height of the camera; the detection height of the scanning instrument represents the vertical distance between the scanning instrument and the ground where the inspection device is located; the detection height of the camera represents the vertical distance between the camera and the ground where the inspection device is located; the detection data includes three-dimensional point cloud data and image data.

8. The AI-based construction site inspection management system according to claim 6, wherein: The spatial model construction module includes a three-dimensional point cloud model construction unit and an inspection target spatial analysis unit; The three-dimensional point cloud model construction unit is used to extract the three-dimensional point cloud data in the detection data and remove the three-dimensional point cloud data corresponding to the moving objects during the scanning process; using the three-dimensional point cloud data after data cleaning, a three-dimensional point cloud model is established; The inspection target spatial analysis unit is used to construct a spatial coordinate system in the three-dimensional point cloud model, extract the spatial coordinates of the boundary of the inspection target in the three-dimensional point cloud model, and analyze the spatial coordinates of the centroid point of the inspection target.

9. The AI-based construction site inspection management system according to claim 6, characterized in that: The optimal section analysis model includes an inspection target image analysis unit and an optimal detection section judgment unit; The inspection target image analysis unit is used to analyze the actual maximum observation area of the inspection target corresponding to different spatial coordinates of the camera; According to each actual maximum observation area, calculate the pixel area of the inspection target displayed in the image data at different spatial coordinates of the camera; The optimal detection section judgment unit is used to analyze the analysis reference value of each pixel area in the pixel area set; Extract the spatial coordinates of the camera corresponding to the pixel area whose analysis reference value is equal to the analysis reference value threshold and record them as the spatial boundary, and record the operation path of the inspection device between the two spatial boundaries as the optimal detection section.

10. The AI-based inspection and management system for construction sites according to claim 6, characterized in that: The parameter adjustment and update module includes a real-time control unit and a parameter update unit; The real-time control unit is used to form an inspection parameter sequence with all the inspection parameter groups according to the operation direction of the inspection device, store the inspection parameter sequence in the database; after the camera is turned on, use the spatial coordinates of the camera to find the corresponding inspection parameter group in the database, and adjust the real-time inspection angle of the camera of the inspection device according to the inspection parameter group; The parameter update unit is used to analyze the real-time pixel area of the inspection target in the real-time image data, and the real-time pixel area is equal to the actual size of a single pixel multiplied by the number of pixels of the inspection target; If the real-time pixel area is greater than or equal to the pixel area multiplied by the preset proportionality coefficient, no processing is performed; If the real-time pixel area is less than the pixel area multiplied by the preset proportionality coefficient, adjust the real-time inspection angle until the real-time pixel area is greater than or equal to the pixel area multiplied by the preset proportionality coefficient; update the adjusted real-time inspection angle to the inspection parameter group.

Images