Virtual-real fused intelligent inspection method for construction site

Abstract

The invention provides a virtual-real fusion construction site intelligent inspection method, and the method comprises the steps: building a boundary collision body block of a construction site through obtaining a BIM model of the construction site; obtaining pose parameters of a tower crane camera on the construction site, and calculating rotation parameters of the camera based on the BIM model and the boundary collision body block; defining a supervision object weight and a tower crane camera inspection control matrix; and based on the supervision object weight and the tower crane camera monitoring picture identification result, the camera inspection control matrix is updated in real time, and the tower crane camera is intelligently controlled to execute inspection. The visual angle weight score is dynamically updated according to the construction site supervision video and the intelligent identification result, so that the stay time of the tower crane camera under each visual angle is adjusted, the picture of the camera is always focused on the main supervision object of the construction site, the monitoring coverage range is optimized, and the effectiveness and efficiency of construction site supervision are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of construction management, and particularly relates to an intelligent inspection method for a virtual-reality integrated construction site. Background Art

[0002] With the continuous development of the construction industry, the scale of construction sites is increasing day by day, and the safety risks during the construction process are becoming more and more complex. The traditional safety supervision method for construction sites relies on manual inspections and fixed-point video monitoring. Although it can ensure construction safety to a certain extent, due to the low inspection frequency, limited coverage, high labor cost, and lack of focus in information collection, it is difficult to efficiently and accurately discover potential safety hazards, and it is often difficult to cope with the complex and changeable construction environment, resulting in management lag. Summary of the Invention

[0003] The purpose of the present invention is to provide an intelligent inspection method for a virtual-reality integrated construction site.

[0004] To solve the above problems, the present invention provides an intelligent inspection method for a virtual-reality integrated construction site, including:

[0005] Step S1: Obtain the BIM model of the construction site and construct the boundary collision volume of the construction site;

[0006] Step S2: Obtain the pose parameters of the tower crane camera at the construction site, and calculate the rotation parameters of the camera based on the BIM model and the boundary collision volume;

[0007] Step S3: Define the weight of the supervision object and the inspection control matrix of the tower crane camera;

[0008] Step S4: Based on the weight of the supervision object and the recognition result of the monitoring screen of the tower crane camera, update the inspection control matrix of the camera in real time, and intelligently control the tower crane camera to perform inspections.

[0009] Further, in the above method, Step S1: Obtain the BIM model of the construction site and construct the boundary collision volume of the construction site, including:

[0010] Step S1.1: Obtain and integrate the site layout model and the building structure model of the construction site to obtain the BIM model of the construction site, and present the BIM model through a graphics engine, where the positive direction of the X-axis of the spatial coordinate system where the BIM model is located is due east, the positive direction of the Y-axis is due north, and the positive direction of the Z-axis is vertically upward;

[0011] Step S1.2: Draw a closed spatial polyline along the top of the fence at the construction site boundary in the BIM model, and connect the spatial polyline with its polygon projection on the horizontal ground through vertical projection lines to construct the boundary collision volume of the construction site, and denote the vertices of the spatial polyline as V i, where \(i = 1, 2, \ldots, n\) and \(n\) is a positive integer.

[0012] Further, in the above method, step S2: Obtain the pose parameters of the tower crane camera at the construction site, and calculate the rotation parameters of the camera based on the BIM model and the boundary collision blocks, including:

[0013] Step S2.1: Obtain the video monitoring data and pose parameters of the tower crane camera at the construction site. When the tower crane camera is installed, set the horizontal and vertical rotation angles of the camera pan-tilt to 0°, measure the three-dimensional position \(T\) of the camera pan-tilt, and adjust the default lens orientation of the camera to due north;

[0014] Step S2.2: Connect the three-dimensional position \(T\) of the tower crane camera pan-tilt and the vertex \(V\) of the construction site boundary collision block i , and calculate the range \(t=(min(t_1, \ldots, t i , \ldots, t n ), ) of the vertical rotation angle of the tower crane camera, where \(t i is the angle between the vector and the XOY plane, and FOV y represents the vertical field of view angle of the camera, which is an internal parameter of the camera;

[0015] Step S2.3: Calculate the vertical direction rotation angle \(\Delta t\) of the tower crane camera during a single rotation based on the vertical rotation angle range \(t\) of the tower crane camera:

[0016]

[0017] Step S2.4: Calculate the horizontal direction rotation angle \(\Delta p\) of the tower crane camera during a single rotation:

[0018]

[0019] where FOV x represents the horizontal field of view angle of the camera, which is an internal parameter of the camera.

[0020] Further, in the above method, step S3: Define the supervision object weight and the tower crane camera inspection control matrix, including:

[0021] Step S3.1: According to the actual management requirements, sort out the supervision objects \(W = \{w o \}\), and define the weights \(S = \{s wo \}\), \(o = 1, 2, \ldots, k\), where \(k\) is a positive integer;

[0022] Step S3.2: Define the inspection control matrix \(R\) of the tower crane camera mn, where m is the maximum number of horizontal rotations of the camera, n is the maximum number of vertical rotations of the camera, The element r of the i-th row and j-th column of matrix R mn represents the sum of the weights of all supervised objects in the captured image when the horizontal rotation angle p ij of the camera is p = Δp×(i - 1) and the vertical rotation angle t i = Δt×(j - 1), where j t is the number of supervised objects w in the current captured image. t wo is the number of supervised objects w o in the current captured image.

[0023] Further, in the above method, the supervised objects in step S3.1 at least include workers, tower cranes, and olive trucks.

[0024] Further, in the above method, step S4: Based on the weights of the supervised objects and the recognition results of the tower crane camera monitoring images, the camera inspection control matrix is updated in real time, and the tower crane camera is intelligently controlled to perform inspections, including:

[0025] Step S4.1: Initialize the values of the elements in the inspection control matrix R mn of the tower crane camera to 1, and define the unit residence time d of the tower crane camera;

[0026] Step S4.2: For a non-negative element r mn in the inspection control matrix R ij of the tower crane camera, control the camera to rotate to the horizontal rotation angle p i = Δp×(i - 1) and the vertical rotation angle t i = Δt×(j - 1);

[0027] Step S4.3: After the tower crane camera rotates to the specified angle, determine the simulated shooting ray with the three-dimensional position T of the camera as the endpoint and the current orientation as the direction, and use the collision detection algorithm to determine whether there is an intersection between the simulated shooting ray and the boundary collision block of the construction site. If there is an intersection, go to step S4.4; otherwise, it indicates that the current camera view is an area outside the construction site that does not require supervision, and update r ij = -1;

[0028] Step S4.4: Calculate the residence time d×r ij of the tower crane camera in the current view, and control the camera to stay for the corresponding duration in the current view. During this period, automatically detect the supervised objects in the camera monitoring image through the target recognition algorithm, calculate and update

[0029] Step S4.5: Traverse the inspection control matrix Rmn Repeat steps S4.2 to S4.4 to complete one round of on-site construction inspection, and update to obtain a new tower crane camera inspection control matrix R. mn ;

[0030] Step S4.6: Repeat step S4.5 to perform intelligent on-site inspection of the tower crane camera.

[0031] Furthermore, in the above method, in step S4.4, it further includes:

[0032] If the intelligent recognition algorithm detects an abnormal situation of the supervised object in the monitoring screen of the tower crane camera, an alarm message will be pushed to relevant users through the business system.

[0033] Compared with the prior art, the present invention obtains the BIM model of the construction site and constructs the boundary collision blocks of the construction site; obtains the pose parameters of the tower crane camera at the construction site, and calculates the rotation parameters of the camera based on the BIM model and the boundary collision blocks; defines the weights of the supervised objects and the tower crane camera inspection control matrix; based on the weights of the supervised objects and the recognition results of the tower crane camera monitoring screen, the camera inspection control matrix is updated in real time, and the tower crane camera is intelligently controlled to perform inspections. The present invention dynamically updates the perspective weight scores according to the construction site supervision video and the intelligent recognition results, thereby adjusting the residence time of the tower crane camera at each perspective, so that the camera screen always focuses on the main supervised objects of the construction site, optimizing the monitoring coverage and improving the effectiveness and efficiency of construction site supervision.

[0034] The present invention optimizes the monitoring coverage through the parametric control of the camera and image recognition technology, and improves the effectiveness of construction site supervision; through image recognition technology, intelligent early warning and abnormal situation recognition are realized; combined with BIM technology, a digital twin model of the construction site is provided, which is convenient for managers to make decisions; through intelligent technology and automated equipment, the management efficiency and safety level of the construction site are improved, realizing automated and intelligent supervision of the construction site. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 is a flowchart of the intelligent on-site inspection method for the virtual-real fusion construction site according to an embodiment of the present invention;

[0036] Figure 2 is a calculation schematic diagram for solving the minimum value of the vertical rotation angle of the camera according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0037] To make the above objects, features, and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below with reference to the drawings and specific embodiments.

[0038] AsFigure 1 As shown in the figure, the present invention provides an intelligent inspection method for the virtual-real fusion construction site, including:

[0039] Step S1: Obtain the BIM model of the construction site and construct the boundary collision volume of the construction site;

[0040] Step S1.1: Obtain and integrate the site layout model and the building structure model of the construction site to obtain the BIM model of the construction site, and present the BIM model through a graphics engine. The positive direction of the X-axis of the spatial coordinate system where the BIM model is located is due east, the positive direction of the Y-axis is due north, and the positive direction of the Z-axis is vertically upward;

[0041] Step S1.2: Draw a closed spatial polyline along the top of the enclosure wall of the construction site boundary in the BIM model, and connect the spatial polyline with its polygonal projection on the horizontal ground through vertical projection lines to construct the boundary collision volume of the construction site. Denote the vertices of the spatial polyline as V i , where i = 1, 2, …, n, and n is a positive integer;

[0042] Specifically, for example, draw a closed spatial polyline along the top of the enclosure wall of the construction site boundary in the BIM model, and connect the spatial polyline with its polygonal projection on the horizontal ground through vertical projection lines to construct the boundary collision volume of the construction site. The vertex coordinates of the spatial polyline are V1 = (-18, -15, 0), V2 = (-30, 162, 0), V3 = (267, 161, 0), V4 = (305, -25.8, 0), V5 = (130, -36, 0).

[0043] Step S2: Obtain the pose parameters of the tower crane camera at the construction site, and calculate the rotation parameters of the camera based on the BIM model and the boundary collision volume;

[0044] Step S2.1: Obtain the video monitoring data and pose parameters of the tower crane camera at the construction site. When the tower crane camera is installed, set the horizontal and vertical rotation angles of the camera pan-tilt head to 0 degrees, measure the three-dimensional position T of the camera pan-tilt head, and adjust the default lens orientation of the camera to due north;

[0045] Specifically, for example, install a camera on the tower crane at the construction site. When installing, set both the horizontal rotation angle (pan) and the vertical rotation angle (tilt) of the camera pan-tilt head to 0°, adjust the camera lens to face due north, and measure the three-dimensional position information of the camera, denoted as point T = (118, 14, 39);

[0046] Step S2.2: Connect the three-dimensional position T of the tower crane camera pan-tilt head and the vertex V of the boundary collision volume of the construction site i, calculate the range of the vertical rotation angle of the tower crane camera \(t=(min(t1,\ldots,t i ,\ldots,t n ), ), where \(t i is the angle between the vector and the XOY plane, and FOV y represents the vertical field of view angle of the camera, which is an internal parameter of the camera;

[0047] Specifically, for example, connect the three-dimensional position \(T\) of the tower crane camera pan-tilt and the vertex \(V\) of the construction site boundary collision block i , calculate the angles between the vector and the XOY plane, which are \(t1 = 15.67^{\circ}\), \(t2 = 10.56^{\circ}\), \(t3 = 10.55^{\circ}\), \(t4 = 11.53^{\circ}\), \(t5 = 37.18^{\circ}\) respectively, and directly obtain the vertical field of view angle FOV y \(= 7.61^{\circ}\) in the internal parameters of the camera, and calculate the range of the vertical rotation angle of the tower crane camera \(t=(10.55^{\circ}, 86.195^{\circ})\);

[0048] Step S2.3: Based on the range \(t\) of the vertical rotation angle of the tower crane camera, calculate the vertical rotation angle \(\Delta t\) when the tower crane camera rotates once:

[0049]

[0050] Specifically, for example, based on the range \(t\) of the vertical rotation angle of the tower crane camera, calculate the vertical rotation angle \(\Delta t\) when the tower crane camera rotates once:

[0051]

[0052] Step S2.4: Calculate the horizontal rotation angle \(\Delta p\) when the tower crane camera rotates once:

[0053]

[0054] Among them, FOV x represents the horizontal field of view angle of the camera, which is an internal parameter of the camera;

[0055] Specifically, for example, directly obtain the horizontal field of view angle FOV x \(= 13.49^{\circ}\) in the internal parameters of the camera, and thus calculate the horizontal rotation angle \(\Delta p\) when the tower crane camera rotates once:

[0056]

[0057] Step S3: Define the weights of the supervised objects and the inspection control matrix of the tower crane camera;

[0058] Step S3.1: According to the actual management requirements, sort out the supervision objects W = {w o}, and define the weights of the supervision objects S = {s wo}, o = 1, 2,..., k, where k is a positive integer;

[0059] Specifically, for example, according to the actual management requirements, sort out the supervision object set W = {worker, tower crane, olive vehicle}, and define the weights of the supervision objects S = {s 工人 = 1, s 塔吊 = 2, s 橄榄车 = 3};

[0060] Step S3.2: Define the inspection control matrix R mn of the tower crane camera, where m is the maximum number of horizontal rotations of the camera, n is the maximum number of vertical rotations of the camera, The element r mn in the i-th row and j-th column of the matrix R ij represents the sum of the weights of all supervision objects in the captured image when the horizontal rotation angle p i = Δp × (i - 1) and the vertical rotation angle t j = Δt × (j - 1), t wo is the number of supervision objects w o in the current captured image;

[0061] Specifically, for example, define the inspection control matrix R mn of the tower crane camera, where m is the maximum number of horizontal rotations of the camera, n is the maximum number of vertical rotations of the camera, The element r mn in the i-th row and j-th column of the matrix R ij represents the sum of the weights of all supervision objects in the captured image when the horizontal rotation angle p i = Δp × (i - 1) and the vertical rotation angle t j = Δt × (j - 1), t wo is the number of supervision objects w o in the current captured image.

[0062] Step S4: Based on the weights of the supervision objects and the recognition results of the tower crane camera monitoring images, update the camera inspection control matrix in real time and intelligently control the tower crane camera to perform inspections;

[0063] Step S4.1: Initialize the inspection control matrix R of the tower crane cameramn The values of the elements in it are 1, and the unit residence time d of the tower crane camera is defined;

[0064] Specifically, for example, initialize the inspection control matrix R of the tower crane camera mn The values of the elements in it are 1, and the unit residence time d of the tower crane camera is defined as d = 5 minutes;

[0065] Step S4.2: For a non - negative element r mn in the inspection control matrix R of the tower crane camera ij , control the camera to rotate to the horizontal rotation angle p i = Δp×(i - 1), and the vertical rotation angle t i = Δt×(j - 1);

[0066] Step S4.3: After the tower crane camera rotates to the specified angle, determine the simulated shooting ray with the three - dimensional position T of the camera as the endpoint and the current orientation as the direction. Use the collision detection algorithm to judge whether there is an intersection between the simulated shooting ray and the boundary collision block of the construction site. If there is an intersection, go to Step S4.4. Otherwise, it indicates that the current camera view is an area outside the construction site that does not need to be supervised, and update r ij = - 1;

[0067] Step S4.4: Calculate the residence time d×r of the tower crane camera at the current view ij , and control the camera to stay at the current view for the corresponding duration. During this period, automatically detect the supervised objects in the camera monitoring screen through the target recognition algorithm, calculate and update If the intelligent recognition algorithm detects abnormal situations of the supervised objects in the tower crane camera monitoring screen, alarm information will be pushed to relevant users through the business system;

[0068] Step S4.5: Traverse the inspection control matrix R of the tower crane camera mn , repeat Steps S4.2 to S4.4 to complete a round of inspection of the construction site, and update to obtain a new inspection control matrix R of the tower crane camera mn ;

[0069] Step S4.6: Repeat Step S4.5 to perform intelligent inspection of the construction site by the tower crane camera.

[0070] Specifically, for example, when the tower crane camera first rotates to pan = 189.84°, tilt = 22.49°, it will stay at the current position for a time d×r 28,3 = 5 minutes. During the stay, automatically detect the supervised objects in the camera monitoring screen through the target recognition algorithm, calculate and update the element r 28,3= max(1, 8) = 8. At this time, if there are any abnormal situations with the supervised objects in the monitoring video of the tower crane camera, the business system will initiate an alarm push to relevant users.

[0071] In summary, through the parametric control of the camera and image recognition technology, the present invention optimizes the monitoring coverage and improves the effectiveness of construction site supervision; through image recognition technology, intelligent early warning and abnormal situation recognition are achieved; combined with BIM technology, a digital twin model of the construction site is provided to facilitate decision-making by management personnel; through intelligent technology and automated equipment, the management efficiency and safety level of the construction site are improved, realizing automated and intelligent supervision of the construction site.

[0072] In this specification, the various embodiments are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. For the same or similar parts among the various embodiments, reference can be made to each other.

[0073] Those skilled in the art can further realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the components and steps of the examples have been generally described according to their functions in the above description. Whether these functions are executed in a hardware or software manner 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 to exceed the scope of the present invention.

[0074] Obviously, those skilled in the art can make various modifications and variations to the invention without departing from the spirit and scope of the invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. An intelligent inspection method for the construction site with virtual-real fusion, characterized in that, It includes the following steps: Step S1: Obtain the BIM model of the construction site and construct the boundary collision volume of the construction site; Step S2: Obtain the pose parameters of the tower crane camera at the construction site and calculate the rotation parameters of the camera based on the BIM model and the boundary collision volume; Step S3: Define the supervision object weights and the tower crane camera inspection control matrix; Step S4: Based on the supervision object weights and the recognition results of the tower crane camera monitoring images, update the camera inspection control matrix in real time and intelligently control the tower crane camera to perform inspections.

2. The intelligent inspection method for the virtual-reality integrated construction site according to claim 1, characterized in that Step S1: Obtain the BIM model of the construction site and construct the boundary collision volume of the construction site, including: Step S1.1: Obtain and integrate the site layout model and the building structure model of the construction site to obtain the BIM model of the construction site, and present the BIM model through a graphics engine. The positive direction of the X-axis of the spatial coordinate system where the BIM model is located is due east, the positive direction of the Y-axis is due north, and the positive direction of the Z-axis is vertically upward; Step S1.2: Draw a closed spatial polyline along the top of the fence at the construction site boundary in the BIM model, and connect the spatial polyline with its polygonal projection on the horizontal ground through vertical projection lines to construct the boundary collision volume of the construction site. Denote the vertices of the spatial polyline as V i , where i = 1, 2, …, n, and n is a positive integer.

3. The intelligent inspection method for the virtual-real fusion construction site according to claim 1, wherein, Step S2: Obtain the pose parameters of the tower crane camera at the construction site and calculate the rotation parameters of the camera based on the BIM model and the boundary collision volume, including: Step S2.1: Obtain the video monitoring data and pose parameters of the tower crane camera at the construction site. When the tower crane camera is installed, set the horizontal and vertical rotation angles of the camera pan-tilt to 0°, measure the three-dimensional position T of the camera pan-tilt, and adjust the default lens orientation of the camera to due north; Step S2.2: Connect the three-dimensional position T of the tower crane camera pan-tilt and the vertex V of the construction site boundary collision volume i , and calculate the range of the vertical rotation angle of the tower crane camera where t i is the angle between the vector and the XOY plane, and FOV y represents the vertical field of view angle of the camera, which is an internal parameter of the camera; Step S2.3: Based on the vertical rotation angle range t of the tower crane camera, calculate the vertical rotation angle Δt during a single rotation of the tower crane camera; Step S2.4: Calculate the horizontal rotation angle Δp during a single rotation of the tower crane camera; Among them, FOV x represents the horizontal field of view angle of the camera, which is an internal parameter of the camera.

4. The intelligent inspection method for the virtual-real fusion construction site according to claim 1, wherein Step S3: Define the supervision object weights and the tower crane camera inspection control matrix, including: Step S3.1: According to the actual management requirements, sort out the supervised objects \(W = \{w o \}\), and define the weights of the supervised objects \(S=\{s wo \}\), \(o = 1, 2, \ldots, k\), where \(k\) is a positive integer; Step S3.2: Define the inspection control matrix R of the tower crane camera mn , where m is the maximum number of horizontal rotations of the camera, n is the maximum number of vertical rotations of the camera, Matrix R mn The element r in the i-th row and j-th column of ij represents the sum of the weights of all supervised objects in the captured image when the horizontal rotation angle p i = Δp×(i - 1) and the vertical rotation angle t j = Δt×(j - 1), t wo is the number of supervised objects w o in the current captured image.

5. The intelligent inspection method for the virtual-real fusion construction site according to claim 4, wherein In step S3.1, the supervision objects at least include workers, tower cranes, and olive trucks.

6. The intelligent inspection method for the virtual-real fusion construction site as described in claim 1, wherein Step S4: Based on the supervision object weights and the recognition results of the tower crane camera monitoring images, update the camera inspection control matrix in real time and intelligently control the tower crane camera to perform inspections, including: Step S4.1: Initialize the inspection control matrix R of the tower crane camera mn with the value of each element being 1, and define the unit residence time d of the tower crane camera; Step S4.2: Check the control matrix R of the tower crane camera mn A non-negative element r in ij , control the camera to rotate to the horizontal rotation angle p i =Δp×(i-1), vertical rotation angle t i =Δt×(j-1); Step S4.3: After the tower crane camera rotates to the specified angle, determine a simulated shooting ray with the three-dimensional position T of the camera as the endpoint and the current orientation as the direction. Use the collision detection algorithm to determine whether there is an intersection between the simulated shooting ray and the boundary collision blocks of the construction site. If there is an intersection, go to step S4.4; otherwise, it indicates that the current camera view is an area outside the construction site that does not require supervision, and update r ij =-1; Step S4.4: Calculate the residence time d×r of the tower crane camera at the current perspective ij , and control the camera to stay at the current perspective for the corresponding duration. During this period, automatically detect the supervised objects in the monitoring screen of the camera through the target recognition algorithm, calculate and update Step S4.5: Traverse the tower crane camera inspection control matrix R mn , repeat steps S4.2 to S4.4 to complete one round of construction site inspection, and update to obtain a new tower crane camera inspection control matrix R mn ; Step S4.6: Repeat step S4.5 to perform intelligent inspections of the construction site by the tower crane camera.

7. The intelligent inspection method for the virtual-reality integrated construction site according to claim 6, wherein In step S4.4, it also includes: If the intelligent recognition algorithm detects an abnormal situation of the supervision object in the tower crane camera monitoring image, an alarm message will be pushed to the relevant users through the business system.

Images