A curtain wall glass installation posture automatic calibration method and device

Abstract

This invention relates to the field of attitude estimation technology, specifically to an automatic attitude calibration method and device for curtain wall glass installation, comprising: acquiring initial pose information of a curtain wall glass calibration plate and a curtain wall frame calibration plate; acquiring the relative pose of the curtain wall glass calibration plate and the curtain wall frame calibration plate, and performing pre-adjustment during installation; adjusting the installation angle of the calibration feature points of the curtain wall glass calibration plate and the curtain wall frame calibration plate according to the pose performance of the corresponding calibration feature points after adjusting the coordinate origin, and acquiring installation pose data of a first calibration point; acquiring installation pose data of a second calibration point according to the relative positional relationship of the corresponding calibration feature points of the curtain wall glass calibration plate and the curtain wall frame calibration plate; and performing staged calibration of the pre-adjusted curtain wall glass according to the installation pose data to obtain the attitude calibration result of the curtain wall glass installation. This invention fully considers the coordinate unification problem between different objects, ensuring better calibration results.

Description

Technical Field

[0001] This invention relates to the field of attitude estimation technology, and specifically to an automatic attitude calibration method and device for curtain wall glass installation. Background Technology

[0002] As modern architecture evolves towards higher floors and curved surfaces, the installation accuracy of curtain wall glass, a key component of the building's exterior envelope, directly impacts the building's aesthetic appearance and structural safety. Traditional curtain wall installation relies heavily on manual measurement and positioning, resulting in low efficiency, accumulated errors from repeated calibration, and high risks associated with working at heights. Particularly in the installation of irregularly shaped curtain walls and large-size unit panels, manual calibration struggles to meet millimeter-level positional matching requirements, leading to stress concentration and seal failure between the glass panels and the curtain wall frame.

[0003] In existing technologies, laser total stations combined with prism reflection can achieve three-dimensional coordinate positioning, but due to limitations in equipment size and multi-target tracking capabilities, they are difficult to adapt to dynamic construction scenarios. Inertial navigation-based pose estimation systems are susceptible to mechanical vibration interference, resulting in accumulated errors. Some studies have attempted to introduce machine vision technology, such as using feature point matching to achieve relative pose calculation, but feature loss is prone to occur on weakly textured curtain wall surfaces. Furthermore, these methods cannot consider the unification of multiple coordinate systems between the camera, glass, and mounting frame, leading to poor performance of automated pose calibration methods. Summary of the Invention

[0004] To address the problem that existing technologies cannot consider the unification of multiple coordinate systems between the camera, glass, and mounting frame, resulting in poor performance of automated attitude calibration methods, this invention aims to provide an automated attitude calibration method and device for curtain wall glass installation. The specific technical solution adopted is as follows:

[0005] In a first aspect, the present invention provides an automatic posture calibration method for curtain wall glass installation, comprising:

[0006] Obtain the initial pose information of the curtain wall glass calibration plate and the curtain wall frame calibration plate, wherein the initial pose information includes rotation dimension and translation dimension;

[0007] A matrix transformation operation is performed based on the initial pose information to obtain the relative pose of the curtain wall glass calibration plate and the curtain wall frame calibration plate; the relative pose is used to perform pre-adjustment for the installation of the curtain wall glass.

[0008] Based on the positional performance of the corresponding calibration feature points of the curtain wall glass calibration plate and the curtain wall frame calibration plate after adjusting the coordinate origin, the installation angle of the calibration feature points is adjusted to obtain the installation positional data of the first calibration point;

[0009] Based on the relative positional relationship of the calibration feature points corresponding to the curtain wall glass calibration plate and the curtain wall frame calibration plate, the installation posture data of the second calibration point is obtained;

[0010] Based on the installation posture data, the pre-adjusted curtain wall glass is calibrated in stages to obtain the posture calibration results of the curtain wall glass installation.

[0011] Preferably, obtaining the initial pose information of the curtain wall glass calibration plate and the curtain wall frame calibration plate specifically includes:

[0012] Obtain two-dimensional images of the curtain wall glass calibration plate and the curtain wall frame calibration plate;

[0013] For any two-dimensional image, based on the pixel coordinates of different pixels on the two-dimensional image and the spatial coordinates of the corresponding pixels in three-dimensional space, the origin of the coordinates of the calibration board in three-dimensional space is set as the centroid of the calibration board. The PnP algorithm is used to obtain the first rotation matrix and the first translation matrix of the pixel coordinate system of the two-dimensional image relative to the spatial coordinate system of the corresponding calibration board. The initial pose information includes the first rotation matrix and the first translation matrix.

[0014] Preferably, the step of performing a matrix transformation operation based on the initial pose information to obtain the relative pose of the curtain wall glass calibration plate and the curtain wall frame calibration plate specifically includes:

[0015] Based on the initial pose information of the curtain wall glass calibration plate, the homogeneous coordinate matrix of the first relative pose between the curtain wall glass calibration plate and the image acquisition device is determined.

[0016] Based on the initial pose information of the curtain wall frame calibration plate, perform matrix inverse transformation to determine the homogeneous coordinate matrix of the second relative pose from the image acquisition device to the curtain wall frame calibration plate.

[0017] Based on the product of the homogeneous coordinate matrices between the first relative pose and the second relative pose, the homogeneous coordinate matrix of the global relative pose between the curtain wall glass calibration plate and the curtain wall frame calibration plate is determined.

[0018] Based on the homogeneous coordinate matrix of the global relative pose, the relative pose of the curtain wall glass calibration plate and the curtain wall frame calibration plate is determined.

[0019] Preferably, the step of adjusting the installation angle of the calibration feature points based on the pose performance of the corresponding calibration feature points of the curtain wall glass calibration plate and the curtain wall frame calibration plate after adjusting the coordinate origin, and obtaining the installation pose data of the first calibration point, specifically includes:

[0020] Obtain any set of corner points corresponding to the curtain wall glass calibration plate and the curtain wall frame calibration plate, and record them as the first calibration feature point and the second calibration feature point respectively; the first calibration point shown includes the first calibration feature point of the curtain wall glass calibration plate and the second calibration feature point of the curtain wall frame calibration plate;

[0021] Set the origin of the three-dimensional coordinates of the curtain wall glass calibration plate as the first calibration feature point, and set the origin of the three-dimensional coordinates of the curtain wall frame calibration plate as the second calibration feature point. Obtain the first feature pose of the first calibration feature point and the second feature pose of the second calibration feature point respectively.

[0022] Perform a matrix transformation operation based on the first feature pose and the second feature pose to obtain the relative pose of the first calibration feature point and the second calibration feature point;

[0023] Based on the difference between the actual measured dimensions of the curtain wall glass and the actual measured dimensions of the curtain wall frame, and combined with the preset safety distance, the second translation matrix of the first calibration feature point is obtained; based on the spatial coordinate system of the curtain wall glass calibration plate and the preset rotation angle, the second rotation matrix of the first calibration feature point is obtained.

[0024] Under the relative pose of the first calibration feature point and the second calibration feature point, rotate sequentially according to the second rotation matrix and translate according to the second translation matrix to obtain the installation pose data of the first calibration point.

[0025] Preferably, the step of obtaining the second translation matrix of the first calibration feature point based on the difference between the actual measured dimensions of the curtain wall glass and the actual measured dimensions of the curtain wall frame, combined with a preset safety distance, specifically includes:

[0026] Obtaining the actual measured dimensions of the curtain wall glass includes the actual length of each side of the curtain wall glass; obtaining the actual measured dimensions of the curtain wall frame includes the actual length of each side of the curtain wall frame.

[0027] Based on the difference in the actual length of the corresponding side between the curtain wall glass and the curtain wall frame, the single-sided installation redundancy distance of each side of the curtain wall glass is determined, and the average of the single-sided installation redundancy distances of all sides of the curtain wall glass is taken as the balanced redundancy distance.

[0028] The curtain wall glass calibration plate is moved in the negative direction to balance the redundant distance, thus obtaining the second translation matrix of the first calibration feature point.

[0029] Preferably, the step of obtaining the second rotation matrix of the first calibration feature point based on the spatial coordinate system of the curtain wall glass calibration plate and a preset rotation angle specifically includes:

[0030] The curtain wall glass calibration plate is rotated along the x-axis by a preset rotation angle to obtain the x-axis rotation matrix; the curtain wall glass calibration plate is rotated in the opposite direction along the y-axis by a preset rotation angle to obtain the y-axis rotation matrix; the product of the x-axis rotation matrix and the y-axis rotation matrix is ​​used as the second rotation matrix of the first calibration feature point.

[0031] Preferably, obtaining the installation pose data of the second calibration point based on the relative positional relationship of the calibration feature points corresponding to the curtain wall glass calibration plate and the curtain wall frame calibration plate specifically includes:

[0032] Any corner point collinear with the first calibration feature point of the curtain wall glass calibration plate is taken as the third calibration feature point, and the corner point on the curtain wall frame calibration plate that matches the third calibration feature point is taken as the fourth calibration feature point; the second calibration point includes the third calibration feature point and the fourth calibration feature point;

[0033] The PnP algorithm is used to transform the coordinates of the third calibration feature point to obtain the current spatial coordinates of the third calibration feature point in the spatial coordinate system of the curtain wall frame calibration plate; the target spatial coordinates of the fourth calibration feature point in the spatial coordinate system of the curtain wall frame calibration plate are obtained.

[0034] Based on the current spatial coordinates and the coordinates of the first calibration feature point, determine the current spatial vector of the first calibration feature point pointing to the third calibration feature point in the current pose; based on the target spatial coordinates and the coordinates of the first feature point, determine the target spatial vector of the first calibration feature point pointing to the fourth calibration feature point in the current pose.

[0035] The third rotation matrix of the third calibration feature point is obtained based on the current spatial vector and the target spatial vector, and the installation pose data of the second calibration point includes the third rotation matrix.

[0036] Preferably, obtaining the third rotation matrix of the third calibration feature point based on the current spatial vector and the target spatial vector specifically includes:

[0037] Under the constraint condition R B Given v1 = v2, solve for the third rotation matrix R of the third calibration feature point. B v1 is the current space vector, and v2 is the target space vector.

[0038] Preferably, the step of performing phased calibration on the pre-adjusted curtain wall glass based on the installation posture data to obtain the posture calibration result of the curtain wall glass installation specifically includes:

[0039] The first stage of calibration is performed on the pre-adjusted curtain wall glass based on the installation posture data of the first calibration point.

[0040] The second stage of calibration is performed on the curtain wall glass after the first stage calibration based on the installation posture data of the second calibration point.

[0041] Obtain the third-stage relative pose of the current curtain wall glass calibration plate and curtain wall frame calibration plate, and use the third-stage relative pose to make final adjustments to the curtain wall glass to obtain the attitude calibration result of the curtain wall glass installation.

[0042] Secondly, the present invention provides an automatic attitude calibration device for curtain wall glass installation, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the computer program is executed by the processor, it implements the steps of an automatic attitude calibration method for curtain wall glass installation.

[0043] The embodiments of the present invention have at least the following beneficial effects:

[0044] This invention first acquires the initial pose information of the curtain wall glass calibration plate and the curtain wall frame calibration plate, including rotation and translation dimensions, providing a data foundation for subsequent feature analysis. By analyzing the relative pose information between the two, the relative poses of the curtain wall glass calibration plate and the curtain wall frame calibration plate are obtained; this achieves three-dimensional spatial pose alignment, solving the problem of multi-coordinate system unification, and then enabling pre-adjustment of the curtain wall glass installation based on the unified relative poses. Next, a phased calibration operation is performed on the corresponding calibration feature points on the curtain wall glass and the curtain wall frame. Firstly, the origin of the coordinate system needs to be adjusted, and the pose performance of the corresponding calibration feature points needs to be analyzed to determine the degree of translation required. The installation angle is also adjusted to analyze the degree of rotation required, determining the installation pose data of the first calibration point, which represents the first stage of the installation process. Secondly, the relative positional relationship of the corresponding calibration feature points needs to be analyzed, and the degree of rotation required for the calibration feature points is determined, determining the installation pose data of the second calibration point, which represents the second stage of the installation process. Finally, the phased calibration achieves automated posture calibration of the curtain wall glass installation. This invention fully considers the issue of coordinate unification between different objects and performs phased calibration and iterative optimization to ensure better calibration results. Attached Figure Description

[0045] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 This is a flowchart of the steps of an automatic posture calibration method for curtain wall glass installation provided by the present invention;

[0047] Figure 2 This is a flowchart of the first sub-step of an automatic posture calibration method for curtain wall glass installation provided by the present invention;

[0048] Figure 3 This is a flowchart of the second sub-step of an automatic posture calibration method for curtain wall glass installation provided by the present invention;

[0049] Figure 4 This is a schematic diagram of the first-stage installation result provided by the present invention;

[0050] Figure 5 This is a flowchart of the third sub-step of an automatic posture calibration method for curtain wall glass installation provided by the present invention;

[0051] Figure 6 This is a schematic diagram of the second-stage installation result provided by the present invention;

[0052] Figure 7 This is a structural schematic diagram of an automatic posture calibration device for curtain wall glass installation provided by the present invention;

[0053] Figure 8 This is a schematic diagram of the structure of a computer device provided by the present invention. Detailed Implementation

[0054] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of the SMS encryption method and system based on 5G technology proposed in accordance with the present invention.

[0055] Before introducing the specific solutions provided in the embodiments of this application, some of the uses in this application will be explained to facilitate understanding by those skilled in the art, and are not intended to limit the uses in this application.

[0056] (1) PnP Algorithm: The purpose is to solve the motion of 2D-3D point pairs. It solves the problem of estimating the camera pose in the reference coordinate system based on the two-dimensional pixel coordinates of feature points in the image and their corresponding three-dimensional spatial coordinates. Intuitively, when the camera observes an object in space, given the coordinates of n three-dimensional spatial points on the observed object (relative to a specified coordinate system O) and their two-dimensional projection positions, how to estimate the camera pose (i.e., the camera's attitude in coordinate system O) is the problem that PNP aims to solve. That is, it uses the known correspondence between the three-dimensional structure and the image to solve the relative relationship between the camera and the reference coordinate system. The pose mentioned above includes position and orientation, that is, a 6-DOF state, which is the rotation matrix and translation vector that need to be solved. Both are collectively referred to as the camera's extrinsic parameters.

[0057] (2) Curtain wall: This is the exterior wall cladding of a building, hung like a curtain, hence also called a suspended wall. A curtain wall typically consists of a metal supporting frame, a fixing structure, and curtain wall glass. Since a curtain wall is composed of multiple units, i.e., multiple curtain wall glass panels, and these panels are installed one by one during installation, curtain wall installation equipment is a well-known technology that can control the rotation, tilting, and translation of the curtain wall glass.

[0058] (3) Calibration plate: It is an important tool used in machine vision, image measurement, photogrammetry, 3D reconstruction and other fields. It is usually a flat plate with a fixed-spacing pattern array. By taking pictures of the calibration plate with a camera and combining it with calibration algorithms, a geometric model of the camera image can be established.

[0059] In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments may be combined in any suitable form.

[0060] In the description of the embodiments of the present invention, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. The "and / or" in the text is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. In addition, in the description of the embodiments of the present invention, "multiple" means two or more.

[0061] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature.

[0062] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0063] The following description, in conjunction with the accompanying drawings, details the specific scheme of the automatic posture calibration method and equipment for curtain wall glass installation provided by the present invention.

[0064] Please see Figure 1 The diagram illustrates a flowchart of an automatic attitude calibration method for curtain wall glass installation according to an embodiment of the present invention. The method includes the following steps:

[0065] Step S100: Obtain the initial pose information of the curtain wall glass calibration plate and the curtain wall frame calibration plate. The initial pose information includes rotation dimension and translation dimension.

[0066] The curtain wall glass to be installed is fixed by a clamping device, and then the curtain wall glass is lifted and moved to a preset position by the rotating shaft and telescopic rod of the clamping device. The clamping device refers to the clamping device on the curtain wall installation equipment for realizing the automated calibration and installation operation of the curtain wall glass. This is a well-known technology and will not be described in detail here.

[0067] For the curtain wall glass to be installed, a calibration plate B1 is smoothly affixed to the side facing the image acquisition device; calibration plate B2 is placed on the keel of the curtain wall mounting wall. The image acquisition device refers to the image acquisition device on the curtain wall mounting equipment, used to acquire image information of the curtain wall glass and curtain wall frame.

[0068] Furthermore, machine vision technology and two calibration plates can be used to perform coarse calibration of the curtain wall glass, obtaining the preliminary pose of the curtain wall glass. Specifically, the acquired image information includes two-dimensional images of the curtain wall glass calibration plate and two-dimensional images of the curtain wall frame calibration plate.

[0069] For any two-dimensional image, based on the pixel coordinates of different pixels on the two-dimensional image and the spatial coordinates of the corresponding pixels in three-dimensional space, the origin of the coordinates of the calibration board in three-dimensional space is set as the centroid of the calibration board. The PnP algorithm is used to obtain the first rotation matrix and the first translation matrix of the pixel coordinate system of the two-dimensional image relative to the spatial coordinate system of the corresponding calibration board. The initial pose information includes the first rotation matrix and the first translation matrix.

[0070] Taking the curtain wall glass calibration plate B1 as an example, the centroid of the curtain wall glass calibration plate B1 is set as the origin of the coordinate system. At the same time, corner detection technology is used to detect the two-dimensional image to obtain the spatial coordinates of three corner points on the curtain wall glass calibration plate B1 that are not on the same straight line in three-dimensional space. The spatial coordinates of these three corner points and the corresponding pixel coordinates in the two-dimensional image of the curtain wall glass calibration plate B1 are used to obtain the initial pose information [R1,t1] of the image acquisition device relative to the curtain wall glass calibration plate B1 through the PnP algorithm. Here, R1 represents the first rotation matrix of the pixel coordinate system of the image acquisition device relative to the spatial coordinate system of the curtain wall glass calibration plate B1, with a size of 3*3, and t1 represents the first translation matrix of the pixel coordinate system of the image acquisition device relative to the spatial coordinate system of the curtain wall glass calibration plate B1, with a size of 3*1.

[0071] The initial pose information [R2, t2] of the curtain wall frame calibration plate B2 can be obtained using the same method. Here, R2 represents the first rotation matrix (3*3) of the pixel coordinate system of the image acquisition device relative to the spatial coordinate system of the curtain wall frame calibration plate B2, and t2 represents the first translation matrix (3*1) of the pixel coordinate system of the image acquisition device relative to the spatial coordinate system of the curtain wall frame calibration plate B2. It can be understood that when obtaining the initial pose information [R2, t2] of the curtain wall frame calibration plate B2, the centroid of the curtain wall frame calibration plate B2 is set as the origin of the coordinate system.

[0072] Step S200: Perform a matrix transformation operation based on the initial pose information to obtain the relative pose of the curtain wall glass calibration plate and the curtain wall frame calibration plate; use the relative pose to perform pre-adjustment for the installation of the curtain wall glass.

[0073] Understandably, the initial pose information characterizes the relative pose performance of the image acquisition device with respect to each calibration plate. In order to achieve pose alignment in three-dimensional space, the relative pose performance between the curtain wall glass calibration plate B1 and the curtain wall frame calibration plate B2 can be analyzed using the image acquisition device as the intermediate point.

[0074] Furthermore, in order to achieve the alignment of the overall posture of the curtain wall glass with that of the curtain wall frame, and to limit the movement of the curtain wall glass to a safe installation distance, a safe distance can be moved based on the relative posture information between the curtain wall glass calibration plate and the curtain wall frame calibration plate. This ensures that the overall posture of the curtain wall glass and the curtain wall frame is aligned and maintained at a safe installation distance.

[0075] Finally, the relative pose of the curtain wall glass calibration plate and the curtain wall frame calibration plate can be used to pre-adjust the installation of the curtain wall glass. Taking a three-axis device as an example, the rotation matrix of the relative pose is decomposed into motion parameters of the three axes, namely Euler angles. The angle of the three-axis device is adjusted according to the Euler angles. The position is adjusted by the displacement of the robotic arm corresponding to the translation vector of the relative pose. The process of control and adjustment is a well-known technology and will not be described in detail here.

[0076] Step S300: Based on the positional performance of the corresponding calibration feature points of the curtain wall glass calibration plate and the curtain wall frame calibration plate after adjusting the coordinate origin, adjust the installation angle of the calibration feature points to obtain the installation positional data of the first calibration point.

[0077] After pre-adjustment, this invention performs phased calibration, gradually reducing errors through iterative optimization at each installation stage. In the first stage, calibration is performed on any fixed feature point. After adjusting the coordinate origin to the calibration feature point, the rotational pose of corresponding matching feature points on the curtain wall glass calibration plate and the curtain wall frame calibration plate is analyzed to adjust the installation angle. The dimensional differences between the curtain wall glass calibration plate and the curtain wall frame calibration plate are analyzed to adjust the translation distance, thereby determining the pose information corresponding to the calibration operation on a fixed feature point.

[0078] As can be seen in step S100, the corner points of the curtain wall glass calibration plate and the curtain wall frame calibration plate can be detected and obtained respectively using corner point matching operation. The corresponding matching relationship between the two corner points can be obtained using corner point matching operation. This is a well-known technology and will not be described in detail here.

[0079] Based on this, any set of corner points corresponding to the curtain wall glass calibration plate and the curtain wall frame calibration plate are obtained and recorded as the first calibration feature point and the second calibration feature point, respectively; the first calibration point shown includes the first calibration feature point of the curtain wall glass calibration plate and the second calibration feature point of the curtain wall frame calibration plate.

[0080] Furthermore, the origin of the coordinate system is adjusted. The origin of the three-dimensional coordinate system of the curtain wall glass calibration plate is set as the first calibration feature point, and the origin of the three-dimensional coordinate system of the curtain wall frame calibration plate is set as the second calibration feature point. The first feature pose of the first calibration feature point and the second feature pose of the second calibration feature point are obtained respectively. The method for obtaining the pose is the same as the method for obtaining the initial pose information in step S100.

[0081] By analyzing the relative pose between the first and second feature poses, the relative pose performance between the first calibration point and the second calibration feature point is determined. Based on this analysis, adjustments are made to the pose performance. Specifically, during the pose calibration of the first calibration point, to prevent overall positional deviation of the curtain wall glass, the pose is first rotated for adjustment. This rotation matrix is ​​called the avoidance matrix. The pose of the curtain wall glass is adjusted using the avoidance matrix to ensure the first calibration feature point is installed correctly.

[0082] Simultaneously, collisions can be prevented during the installation of the first calibration feature point of the curtain wall glass. Considering the balance between the differences in the actual measured dimensions of the curtain wall glass and the curtain wall frame, the translation adjustment matrix for the pose is determined. Based on the rotation and translation dimensions, the corresponding pose information between the first calibration feature point of the curtain wall glass calibration plate and the second calibration feature point of the curtain wall frame calibration plate can be obtained, which is the installation pose data of the first calibration point. The installation pose data of the first calibration point characterizes the pose information required to adjust the pose of a feature point on the curtain wall glass calibration plate to reach the target position.

[0083] Step S400: Based on the relative positional relationship of the calibration feature points corresponding to the curtain wall glass calibration plate and the curtain wall frame calibration plate, obtain the installation posture data of the second calibration point.

[0084] After the first stage of pose calibration is completed, the second stage of pose calibration is performed. This mainly takes into account the rotation and displacement of another corner point on the curtain wall glass calibration plate that needs to be moved to the target position. In order to avoid destroying the pose calibration results of the first stage, the position of the first calibration point needs to be fixed and the pose calibration process of the second calibration point is performed.

[0085] Specifically, any corner point collinear with the first calibration feature point of the curtain wall glass calibration plate is taken as the third calibration feature point, such as... Figure 4 As shown, for example, the lower right corner of the square curtain wall glass is used as the third calibration feature point. The corner point on the curtain wall frame marking plate that matches the third calibration feature point is used as the fourth calibration feature point; the second calibration point includes the third calibration feature point and the fourth calibration feature point.

[0086] Then, the PnP algorithm is used to transform the coordinates of the third calibration feature point to obtain the current spatial coordinates of the third calibration feature point in the spatial coordinate system of the curtain wall frame calibration plate; the target spatial coordinates of the fourth calibration feature in the spatial coordinate system of the curtain wall frame calibration plate are then obtained.

[0087] It is understandable that the spatial coordinates of the third calibration feature point correspond to the spatial coordinate system of the curtain wall glass calibration plate, while the target point that the third calibration feature point needs to move to is the fourth calibration feature point. The spatial coordinates of the fourth calibration feature point correspond to the spatial coordinate system of the curtain wall frame calibration plate. In order to achieve coordinate unification, it is necessary to use the relative pose between the curtain wall glass and the curtain wall frame at this time, and combine the PnP algorithm to perform coordinate transformation on the third calibration feature point, so as to transform the spatial coordinates of the third calibration feature point from the spatial coordinate system of the curtain wall glass calibration plate to the spatial coordinate system of the curtain wall frame calibration plate.

[0088] Furthermore, since the position of the first feature calibration point needs to be fixed, the relative position between the current actual position information and the first calibration feature point, as well as the relative position between the position information of the target point and the first feature calibration point, can be compared to determine the rotation that the third calibration feature point needs to be adjusted in the current second stage, and the final installation pose data of the third calibration feature point, which is also the installation pose data of the second calibration point, can be obtained.

[0089] Step S500: Based on the installation posture data, the pre-adjusted curtain wall glass is calibrated in stages to obtain the posture calibration results of the curtain wall glass installation.

[0090] The first stage of calibration is performed on the pre-adjusted curtain wall glass based on the installation posture data of the first calibration point, as described in subsequent embodiments and will not be elaborated further here. The second stage of calibration is performed on the pre-calibrated curtain wall glass based on the installation posture data of the second calibration point; this has also been described in subsequent embodiments and will not be elaborated further here. The third stage relative posture of the current curtain wall glass calibration plate and curtain wall frame calibration plate is obtained, and the curtain wall glass is finally adjusted using this third stage relative posture to obtain the posture calibration result of the curtain wall glass installation.

[0091] Understandably, after completing the calibration operations in the first and second stages, it is necessary to re-acquire the relative poses of the curtain wall glass calibration plate and the curtain wall frame calibration plate, which is referred to as the third stage relative pose. The method for obtaining the relative pose has been described in detail in the above steps and will not be repeated here.

[0092] It should be noted that the positions of the first and second calibration points are fixed at this time. Therefore, when obtaining the relative pose, the corresponding coordinates are still at the position of the first calibration point, such as... Figure 6 As shown, the position of one edge of the curtain wall glass is fixed, so there is no need to perform translation matrix adjustment operations. Only the angle of the curtain wall glass needs to be adjusted for gradual installation. That is, the translation matrix t included in the relative pose of the third stage. C =[0,0,0] T After the curtain wall glass reaches its installation position, subsequent operations such as fixing, sealing, and applying sealant are performed to complete the installation of the curtain wall glass.

[0093] Therefore, it can be understood that the main implementation process of the phased calibration in this embodiment includes: firstly, aligning the centroids of the curtain wall glass and the curtain wall frame to perform a pre-adjustment operation for attitude calibration; then, adjusting the coordinate origin and aligning a feature corner point of the curtain wall glass and the curtain wall frame to perform the first stage of attitude calibration; further, with the feature corner point fixed after the first stage adjustment, performing the second stage of attitude calibration to realize the installation operation of one side of the curtain wall glass. Finally, rotating and adjusting the curtain wall glass as a whole to realize the complete attitude calibration operation of the curtain wall glass, completing the installation operation of the curtain wall glass.

[0094] In summary, this invention sets calibration plates on both the curtain wall glass and the curtain wall frame. Using an image acquisition device combined with the PnP algorithm, the relative poses of the two are obtained, solving the problem of multi-coordinate system consistency and ensuring the consistency of pose calculation references. The relative poses of the curtain wall glass and the curtain wall frame are derived through matrix transformation (homogeneous coordinate matrix), achieving three-dimensional spatial pose alignment. Staged calibration (first calibration point → second calibration point → overall pose adjustment) is implemented, gradually reducing errors through iterative optimization of installation distances. Millimeter-level pose matching is achieved through collaborative calibration using the PnP algorithm and dual calibration plates, solving the error accumulation problem of traditional manual calibration. Coordinate system correction using calibration plates reduces manual intervention, lowers the complexity of the calibration process, and improves installation efficiency. It is compatible with machine vision, deep learning models, and industrial control modules, supporting feature recognition of weakly textured curtain wall surfaces.

[0095] In some embodiments, the above step S200 can be achieved through... Figure 2 The sub-steps shown are implemented as follows:

[0096] Step S201: Based on the initial pose information of the curtain wall glass calibration plate, determine the homogeneous coordinate matrix of the first relative pose between the curtain wall glass calibration plate and the image acquisition device.

[0097] Specifically, the homogeneous coordinate matrix T of the first relative pose of the image acquisition device with respect to the curtain wall glass calibration plate B1 B1→C It can be represented as Here, C represents the image acquisition device, and the homogeneous coordinate matrix unifies the mathematical expression of spatial transformations. This allows complex spatial transformations to be chained together, such as: camera → calibration plate → curtain wall center, or curtain wall glass → image acquisition device → curtain wall frame. Based on this, the image acquisition device can be used as an intermediate point to obtain the relative pose of the curtain wall glass with respect to the curtain wall frame.

[0098] Step S202: Perform inverse matrix transformation based on the initial pose information of the curtain wall frame calibration plate to determine the homogeneous coordinate matrix of the second relative pose from the image acquisition device to the curtain wall frame calibration plate.

[0099] Specifically, the homogeneous coordinate matrix T of the relative pose of the image acquisition device with respect to the curtain wall frame calibration plate B2. B2→C It can be represented as Using the inverse matrix transformation operation, the homogeneous coordinate matrix T of the relative pose is transformed. B2→C The homogeneous coordinate matrix of the second relative pose of the curtain wall frame calibration plate B2 with respect to the image acquisition device C can be expressed as: Let R2 be the transpose of matrix R2.

[0100] Step S203: Based on the product of the homogeneous coordinate matrices between the first relative pose and the second relative pose, determine the homogeneous coordinate matrix of the global relative pose between the curtain wall glass calibration plate and the curtain wall frame calibration plate.

[0101] Specifically, the homogeneous coordinate matrix of the global relative pose between the curtain wall glass calibration plate B1 and the curtain wall frame calibration plate B2, obtained by chain multiplication with the image acquisition device C as the intermediate point, can be expressed as:

[0102]

[0103] Among them, T B1→B2 T is the homogeneous coordinate matrix of the global relative pose between the curtain wall glass calibration plate B1 and the curtain wall frame calibration plate B2. B1→C Let T be the homogeneous coordinate matrix of the first relative pose of the image acquisition device C with respect to the curtain wall glass calibration plate B1. C→B2 The homogeneous coordinate matrix is ​​the second relative pose of the curtain wall frame calibration plate B2 with respect to the image acquisition device C.

[0104] Step S204: Based on the homogeneous coordinate matrix of the global relative pose, determine the relative pose of the curtain wall glass calibration plate and the curtain wall frame calibration plate.

[0105] Specifically, by homogeneous coordinate matrix It can be seen that the relative orientation of the curtain wall glass calibration plate and the curtain wall frame calibration plate can be expressed as: Where R3 represents the rotation matrix between the curtain wall glass calibration plate and the curtain wall frame calibration plate, and t3 represents the translation matrix between the curtain wall glass calibration plate and the curtain wall frame calibration plate.

[0106] It should be further explained that this embodiment sets a safe installation distance of 10cm. Therefore, a safe translation distance can be added to the relative pose of the curtain wall glass and the curtain wall frame. Based on this, the translation matrix corresponding to the relative pose between the curtain wall glass calibration plate and the curtain wall frame calibration plate can be updated, denoted as t3. ′ =t3+[0,0,-d0] T Where d0 represents the preset safety distance. Therefore, the relative pose between the updated curtain wall glass calibration plate and the curtain wall frame calibration plate can be represented as [R3, t3]. ′ ].

[0107] In some embodiments, the above step S300 can be achieved through... Figure 3 The sub-steps shown are implemented as follows:

[0108] Step S301: Perform a matrix transformation operation based on the first feature pose and the second feature pose to obtain the relative pose of the first calibration feature point and the second calibration feature point.

[0109] The first feature pose characterizes the pose information of the image acquisition device relative to the first calibration feature point of the curtain wall glass calibration plate after the coordinate origin transformation, and the second feature pose characterizes the pose information of the image acquisition device relative to the second calibration feature point of the curtain wall frame calibration plate after the coordinate origin transformation. Following the same method as step S200, and using the image acquisition device as the intermediate point, the relative poses [R] of the first and second calibration feature points can be obtained. p1 ,t p1 ], where R p1 Let t be the rotation matrix corresponding to the relative pose. p1 Let p1 be the translation matrix corresponding to the relative pose, and p1 represents the first calibration point.

[0110] Based on this, the relative pose of the first calibration feature point and the second calibration feature point characterizes the relative pose between the curtain wall glass and the curtain wall frame in the spatial coordinate system corresponding to the curtain wall glass and the curtain wall frame after adjusting the origin to the first calibration feature point and the second calibration feature point.

[0111] Step S302: Based on the difference between the actual measured dimensions of the curtain wall glass and the actual measured dimensions of the curtain wall frame, and combined with the preset safety distance, obtain the second translation matrix of the first calibration feature point.

[0112] The first step is to obtain the actual measured dimensions of the curtain wall glass, including the actual length of each side of the curtain wall glass, and to obtain the actual measured dimensions of the curtain wall frame, including the actual length of each side of the curtain wall frame.

[0113] This embodiment uses a square curtain wall glass as an example to illustrate the process, directly obtaining the actual length of each side of the square curtain wall glass and the actual length of each side of the curtain wall frame.

[0114] The second step is to determine the single-sided installation redundancy distance of each side of the curtain wall glass based on the difference in the actual length of the corresponding side between the curtain wall glass and the curtain wall frame, and take the average of the single-sided installation redundancy distances of all sides of the curtain wall glass as the balanced redundancy distance.

[0115] Specifically, the edge between the corner points where the curtain wall glass and the curtain wall frame have a matching relationship is called the corresponding edge of the curtain wall glass and the curtain wall frame. Considering that there may be errors in the length during the installation of the curtain wall glass, it can be evenly distributed to the left and right sides, therefore... As the single-sided installation redundancy distance of a certain edge of the curtain wall glass, l B1,i l represents the actual length of the i-th side of the curtain wall glass. B2,i This represents the actual length of the i-th side of the curtain wall frame.

[0116] As a concrete example, the balanced redundancy distance can be expressed by the formula: Where d1 represents the balanced redundancy distance, and n represents the total number of corresponding sides of the curtain wall glass and curtain wall frame. This represents the single-sided installation redundancy distance of the curtain wall glass and curtain wall frame corresponding to the i-th side.

[0117] The third step is to move the curtain wall glass calibration plate in the negative direction to balance the redundant distance, thereby obtaining the second translation matrix of the first calibration feature point.

[0118] The second translation matrix t4 of the first calibration feature point can be represented as t4 = [-d1, -d1, 0]. T The second translation matrix represents the compensation for dimensional deviations between the curtain wall glass and the curtain wall frame through reverse translation along the X-axis or Y-axis. The second translation matrix reflects how centered installation with error clearance is achieved by integrating dimensional differences across multiple sides.

[0119] In some embodiments, after moving the curtain wall glass calibration plate in the negative direction to balance the redundancy distance, it is then moved by a preset safety distance to obtain the second translation matrix t. ′ 4,t ′ 4 = t4 + [-d1, -d1, d0] T d0 is the preset safety distance, and d1 is the balanced redundancy distance, which can avoid errors to a certain extent.

[0120] Step S303: Obtain the second rotation matrix of the first calibration feature point based on the spatial coordinate system of the curtain wall glass calibration plate and the preset rotation angle.

[0121] To prevent collisions during the installation of the first calibration feature point of the curtain wall glass, an avoidance angle is set. In this embodiment, the avoidance angle is θ = 10°. Furthermore, the angle by which the curtain wall glass calibration plate rotates along the x-axis is set to a preset rotation angle to obtain the x-axis rotation matrix; the angle by which the curtain wall glass calibration plate rotates in the opposite direction along the y-axis is set to a preset rotation angle to obtain the y-axis rotation matrix; the product of the x-axis rotation matrix and the y-axis rotation matrix is ​​used as the second rotation matrix of the first calibration feature point.

[0122] The x-axis rotation matrix can be represented as: The y-axis rotation matrix can be represented as The second rotation matrix of the first calibration feature point can be represented as R4 = R x ·R y The curtain wall glass calibration plate rotates clockwise by θ around the x-axis and counterclockwise by θ around the y-axis, achieving the purpose of downward tilting and inward adjustment. By adjusting the tilt direction of the curtain wall glass, collisions during installation are avoided.

[0123] Step S304: Under the relative pose of the first calibration feature point and the second calibration feature point, rotate sequentially according to the second rotation matrix and translate sequentially according to the second translation matrix to obtain the installation pose data of the first calibration point.

[0124] Specifically, based on the relative pose information of the first and second calibration feature points, the installation pose data corresponding to the first calibration feature point is obtained by combining the degree of rotation and translation required for the first calibration feature point through operations such as cascading rotation and coordinate system alignment, which is also the installation pose data of the first calibration point.

[0125] The installation pose data corresponding to the first calibration feature point also includes rotation and translation matrices, where the rotation matrix R5 = R4·R p1 Translation matrix t5 = R4·t p1 +t4, the installation pose data corresponding to the first calibration feature point is [R5,t5]=[R4·R p1 ,R4·t p1 +t4], where t4 is the second translation matrix of the first calibration feature point, R4 is the second rotation matrix of the first calibration feature point, and R p1 Let t be the rotation matrix corresponding to the relative pose. p1 Let p1 be the translation matrix corresponding to the relative pose, and p1 represents the first calibration point.

[0126] After obtaining the installation posture data of the first calibration point, the curtain wall glass can be controlled in the first stage of the installation process using the installation posture data. The installation process has been described in step 200 and will not be repeated here. A schematic diagram of the installation result from top view is shown below. Figure 4 As shown, the feature corner point in the upper right corner has been aligned and installed.

[0127] In some embodiments, step S400 may be performed by Figure 5 The sub-steps shown are implemented as follows:

[0128] Step S401: Use the PnP algorithm to perform coordinate transformation on the third calibration feature point to obtain the current spatial coordinates of the third calibration feature point in the spatial coordinate system of the curtain wall frame calibration plate; obtain the target spatial coordinates of the fourth calibration feature point in the spatial coordinate system of the curtain wall frame calibration plate.

[0129] Before coordinate transformation, i.e., after the first stage of control and adjustment operations have been performed and the first calibration point has been fully installed, it is necessary to obtain the relative pose [R0, t0] between the curtain wall glass and the curtain wall frame, where R0 is the rotation matrix and t0 is the translation matrix. The specific method for obtaining this information is the same as in step S200. It should be understood that it is necessary to re-acquire the relative pose information of the image acquisition device relative to the curtain wall glass calibration plate and the relative pose information of the image acquisition device relative to the curtain wall frame calibration plate. Then, using the image acquisition device as an intermediate point, the relative pose [R0, t0] of the curtain wall glass relative to the curtain wall frame is obtained. It should also be noted that the origin of the coordinate system should still be set at the location of the first calibration point.

[0130] Obtain the spatial coordinates of the third calibration feature point in the spatial coordinate system of the curtain wall glass calibration plate. Use the relative pose [R0,t0] between the current glass calibration plate and the curtain wall frame calibration plate to perform coordinate transformation, and obtain the coordinates of the third calibration feature point in the spatial coordinate system of the curtain wall frame calibration plate as the current spatial coordinates.

[0131] As a concrete example, the process of coordinate transformation can be represented as P B ′ 0 = R0·P B0 +t0, P B ′ 0 represents the current spatial coordinates of the third calibration feature point, P B0 This indicates the spatial coordinates of the third calibration feature point in the spatial coordinate system of the curtain wall glass calibration plate.

[0132] Furthermore, the target spatial coordinates of the fourth calibration feature point in the spatial coordinate system of the curtain wall frame calibration plate can be directly obtained in the spatial coordinate system of the curtain wall frame calibration plate. The target spatial coordinates also represent the destination point that the third calibration feature point needs to reach for installation and adjustment.

[0133] Step S402: Based on the current spatial coordinates and the coordinates of the first calibration feature point, determine the current spatial vector of the first calibration feature point pointing to the third calibration feature point in the current pose; based on the target spatial coordinates and the coordinates of the first feature point, determine the target spatial vector of the first calibration feature point pointing to the fourth calibration feature point in the current pose.

[0134] Specifically, after the adjustment of the first calibration point in the first stage, the position of the first calibration point remains fixed. Ignoring minor installation errors, it is approximately regarded as the two calibration feature points of the first calibration point coinciding. That is, the first calibration feature point and the second calibration feature point on the curtain wall glass calibration plate coincide after the position calibration adjustment. Therefore, the first calibration point does not distinguish between the current position and the coordinate position, and both are the coordinate origin.

[0135] Then, the current spatial vector v1 in the current pose can be represented as v1 = P B ′ 0-P A P B ′ 0 represents the current spatial coordinates of the third calibration feature point, P A This represents the current coordinates of the first calibrated feature point, which is also the origin of the coordinate system. The target space vector v2 under the current pose can be expressed as v2 = P. B ′ 1-P A P B ′ 1 represents the target spatial coordinates of the fourth calibration feature point in the spatial coordinate system of the curtain wall frame calibration plate.

[0136] It is understandable that the magnitude of the current spatial vector v1 can be calculated using the current spatial coordinates of the third calibration feature point and the coordinates of the first calibration feature point, and the direction of the current spatial vector v1 is the direction from the first calibration feature point to the third calibration feature point. Similarly, the magnitude of the target spatial vector v2 can be calculated using the target spatial coordinates corresponding to the fourth calibration feature point and the coordinates of the first calibration feature point, and the direction of the target spatial vector v2 is the direction from the first calibration feature point to the fourth calibration feature point.

[0137] Step S403: Obtain the third rotation matrix of the third calibration feature point based on the current spatial vector and the target spatial vector. The installation pose data of the second calibration point includes the third rotation matrix.

[0138] Set the third rotation matrix to R. B The third rotation matrix R can be solved when the third rotation matrix, the current space vector, and the target space vector satisfy a certain relationship. B Specifically, under the constraint condition R B When v1 = v2, the third rotation matrix R of the third calibration feature point is solved using the SVD decomposition algorithm. B v1 is the current space vector, and v2 is the target space vector.

[0139] The simplified steps of SVD decomposition can be described as follows: first, calculate the covariance matrix. Perform singular value decomposition on the covariance matrix H to obtain the third rotation matrix R. B =V·U T V and U are both orthogonal matrices of singular value decomposition, which is a well-known technique and will not be discussed further here.

[0140] Finally, it should be noted that since the position of the first calibration point is already fixed, no translation operation is needed, only rotation. Therefore, the translation matrix corresponding to the installation pose data of the second calibration point is t. B =[0,0,0] T The rotation matrix corresponding to the installation pose data of the second calibration point is the third rotation matrix.

[0141] At this point, the second stage of the curtain wall glass installation process can be controlled using the installation posture data from the second calibration point. The installation process has already been described in step 200 and will not be repeated here. A schematic diagram of the side view of the installation result is shown below. Figure 7 As shown, it can be seen that the two feature corner points belonging to the same edge have achieved the alignment and installation operation.

[0142] This invention provides an automatic attitude calibration device for curtain wall glass installation, including a memory, a processor, and a computer program stored in the memory and running on the processor. When executed by the processor, the computer program implements the steps of an automatic attitude calibration method for curtain wall glass installation. Figure 7 As shown, the automatic posture calibration device for curtain wall glass installation includes:

[0143] The data acquisition module is used to acquire the initial pose information of the curtain wall glass calibration plate and the curtain wall frame calibration plate, wherein the initial pose information includes rotation dimension and translation dimension.

[0144] The pre-adjustment module is installed to perform matrix transformation operations based on the initial pose information to obtain the relative pose of the curtain wall glass calibration plate and the curtain wall frame calibration plate; and to perform pre-adjustment of the curtain wall glass installation using the relative pose.

[0145] The first-stage analysis module is used to adjust the installation angle of the calibration feature points based on the pose performance of the corresponding calibration feature points of the curtain wall glass calibration plate and the curtain wall frame calibration plate after adjusting the coordinate origin, and to obtain the installation pose data of the first calibration point.

[0146] The second-stage analysis module is used to obtain the installation posture data of the second calibration point based on the relative positional relationship of the calibration feature points corresponding to the curtain wall glass calibration plate and the curtain wall frame calibration plate.

[0147] The final attitude calibration module is used to perform phased calibration of the pre-adjusted curtain wall glass based on the installation posture data, and obtain the attitude calibration results of the curtain wall glass installation.

[0148] This application also provides a computer device; please refer to [link / reference]. Figure 8The illustration shows a schematic diagram of a computer device structure according to an embodiment of the present invention. The computer device includes a memory 801, a processor 802, and a computer program 803 stored in the memory 801 and running on the processor 802. When the processor 802 executes the computer program 803, the computer device can execute the aforementioned automatic attitude calibration method for curtain wall glass installation.

[0149] This application also provides a computer program product that, when run on a computer device, enables the computer device to execute the aforementioned automatic posture calibration method for curtain wall glass installation.

[0150] This application also provides a computer-readable storage medium storing computer program code. When the computer program code is run on a computer device, the computer device can execute the aforementioned automatic posture calibration method for curtain wall glass installation.

[0151] In the embodiments provided in this application, it should be understood that the computer device, computer program product and computer-readable storage medium provided are all used to perform the corresponding methods provided above, and therefore the beneficial effects they can achieve can be referred to the beneficial effects of the methods provided above, which will not be repeated here.

[0152] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. An automatic posture calibration method for curtain wall glass installation, characterized in that, The method includes the following steps: Obtain the initial pose information of the curtain wall glass calibration plate and the curtain wall frame calibration plate, wherein the initial pose information includes rotation dimension and translation dimension; A matrix transformation operation is performed based on the initial pose information to obtain the relative pose of the curtain wall glass calibration plate and the curtain wall frame calibration plate; the relative pose is used to perform pre-adjustment for the installation of the curtain wall glass. Based on the positional performance of the corresponding calibration feature points of the curtain wall glass calibration plate and the curtain wall frame calibration plate after adjusting the coordinate origin, the installation angle of the calibration feature points is adjusted to obtain the installation positional data of the first calibration point; the first calibration point includes the first calibration feature point of the curtain wall glass calibration plate and the second calibration feature point of the curtain wall frame calibration plate. Based on the relative positional relationship of the corresponding calibration feature points on the curtain wall glass calibration plate and the curtain wall frame calibration plate, the installation pose data of the second calibration point is obtained, specifically including: Any corner point collinear with the first calibration feature point of the curtain wall glass calibration plate is taken as the third calibration feature point, and the corner point on the curtain wall frame calibration plate that matches the third calibration feature point is taken as the fourth calibration feature point; the second calibration point includes the third calibration feature point and the fourth calibration feature point; The PnP algorithm is used to transform the coordinates of the third calibration feature point to obtain the current spatial coordinates of the third calibration feature point in the spatial coordinate system of the curtain wall frame calibration plate; the target spatial coordinates of the fourth calibration feature point in the spatial coordinate system of the curtain wall frame calibration plate are obtained. Based on the current spatial coordinates and the coordinates of the first calibration feature point, determine the current spatial vector of the first calibration feature point pointing to the third calibration feature point in the current pose; based on the target spatial coordinates and the coordinates of the first calibration feature point, determine the target spatial vector of the first calibration feature point pointing to the fourth calibration feature point in the current pose. The third rotation matrix of the third calibration feature point is obtained based on the current spatial vector and the target spatial vector, and the installation pose data of the second calibration point includes the third rotation matrix. Based on the installation posture data, the pre-adjusted curtain wall glass is calibrated in stages to obtain the posture calibration results of the curtain wall glass installation, including: The first stage of calibration is performed on the pre-adjusted curtain wall glass based on the installation posture data of the first calibration point. The second stage of calibration is performed on the curtain wall glass after the first stage calibration based on the installation posture data of the second calibration point. Obtain the third-stage relative pose of the current curtain wall glass calibration plate and curtain wall frame calibration plate, and use the third-stage relative pose to make final adjustments to the curtain wall glass to obtain the attitude calibration result of the curtain wall glass installation.

2. The automatic posture calibration method for curtain wall glass installation according to claim 1, characterized in that, The acquisition of the initial pose information of the curtain wall glass calibration plate and the curtain wall frame calibration plate specifically includes: Obtain two-dimensional images of the curtain wall glass calibration plate and the curtain wall frame calibration plate; For any two-dimensional image, based on the pixel coordinates of different pixels on the two-dimensional image and the spatial coordinates of the corresponding pixels in three-dimensional space, the origin of the coordinates of the calibration board in three-dimensional space is set as the centroid of the calibration board. The PnP algorithm is used to obtain the first rotation matrix and the first translation matrix of the pixel coordinate system of the two-dimensional image relative to the spatial coordinate system of the corresponding calibration board. The initial pose information includes the first rotation matrix and the first translation matrix.

3. The automatic posture calibration method for curtain wall glass installation according to claim 1, characterized in that, The step of performing a matrix transformation operation based on the initial pose information to obtain the relative pose of the curtain wall glass calibration plate and the curtain wall frame calibration plate specifically includes: Based on the initial pose information of the curtain wall glass calibration plate, the homogeneous coordinate matrix of the first relative pose between the curtain wall glass calibration plate and the image acquisition device is determined. Based on the initial pose information of the curtain wall frame calibration plate, perform matrix inverse transformation to determine the homogeneous coordinate matrix of the second relative pose from the image acquisition device to the curtain wall frame calibration plate. Based on the product of the homogeneous coordinate matrices between the first relative pose and the second relative pose, the homogeneous coordinate matrix of the global relative pose between the curtain wall glass calibration plate and the curtain wall frame calibration plate is determined. Based on the homogeneous coordinate matrix of the global relative pose, the relative pose of the curtain wall glass calibration plate and the curtain wall frame calibration plate is determined.

4. The automatic posture calibration method for curtain wall glass installation according to claim 2, characterized in that, The step of adjusting the installation angle of the calibration feature points based on the pose performance of the corresponding calibration feature points of the curtain wall glass calibration plate and the curtain wall frame calibration plate after adjusting the coordinate origin, and obtaining the installation pose data of the first calibration point, specifically includes: Obtain any set of corner points that correspond to the curtain wall glass calibration plate and the curtain wall frame calibration plate, and record them as the first calibration feature point and the second calibration feature point, respectively. Set the origin of the three-dimensional coordinates of the curtain wall glass calibration plate as the first calibration feature point, and set the origin of the three-dimensional coordinates of the curtain wall frame calibration plate as the second calibration feature point. Obtain the first feature pose of the first calibration feature point and the second feature pose of the second calibration feature point respectively. Perform a matrix transformation operation based on the first feature pose and the second feature pose to obtain the relative pose of the first calibration feature point and the second calibration feature point; Based on the difference between the actual measured dimensions of the curtain wall glass and the actual measured dimensions of the curtain wall frame, and combined with the preset safety distance, the second translation matrix of the first calibration feature point is obtained; based on the spatial coordinate system of the curtain wall glass calibration plate and the preset rotation angle, the second rotation matrix of the first calibration feature point is obtained. Under the relative pose of the first calibration feature point and the second calibration feature point, rotate sequentially according to the second rotation matrix and translate according to the second translation matrix to obtain the installation pose data of the first calibration point.

5. The automatic posture calibration method for curtain wall glass installation according to claim 4, characterized in that, The process of obtaining the second translation matrix of the first calibration feature point based on the difference between the actual measured dimensions of the curtain wall glass and the actual measured dimensions of the curtain wall frame, combined with a preset safety distance, specifically includes: Obtaining the actual measured dimensions of the curtain wall glass includes the actual length of each side of the curtain wall glass; obtaining the actual measured dimensions of the curtain wall frame includes the actual length of each side of the curtain wall frame. Based on the difference in the actual length of the corresponding side between the curtain wall glass and the curtain wall frame, the single-sided installation redundancy distance of each side of the curtain wall glass is determined, and the average of the single-sided installation redundancy distances of all sides of the curtain wall glass is taken as the balanced redundancy distance. The curtain wall glass calibration plate is moved in the negative direction to balance the redundant distance, thus obtaining the second translation matrix of the first calibration feature point.

6. The automatic posture calibration method for curtain wall glass installation according to claim 4, characterized in that, The step of obtaining the second rotation matrix of the first calibration feature point based on the spatial coordinate system of the curtain wall glass calibration plate and a preset rotation angle specifically includes: The curtain wall glass calibration plate is rotated along the x-axis by a preset rotation angle to obtain the x-axis rotation matrix; the curtain wall glass calibration plate is rotated in the opposite direction along the y-axis by a preset rotation angle to obtain the y-axis rotation matrix; the product of the x-axis rotation matrix and the y-axis rotation matrix is ​​used as the second rotation matrix of the first calibration feature point.

7. The automatic posture calibration method for curtain wall glass installation according to claim 1, characterized in that, The step of obtaining the third rotation matrix of the third calibration feature point based on the current spatial vector and the target spatial vector specifically includes: Under the condition of satisfying the constraints Next, solve for the third rotation matrix of the third calibration feature point. , For the current space vector, is the target space vector.

8. An automatic attitude calibration device for curtain wall glass installation, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the computer program is executed by the processor, it implements the steps of the automatic attitude calibration method for curtain wall glass installation as described in any one of claims 1-7.

Images