Building deformation monitoring method and system based on multiple parameters

Abstract

The application relates to the technical field of physical sensors, and discloses a building deformation monitoring method and system based on multiple parameters, which comprises the following steps: collecting a current projection image of a building through a collection device, and obtaining a historical projection image as a comparison projection image; dividing the projection image into projection parts and fixed parts; performing corner point matching between the projection parts and analyzing displacement information to obtain a homography matrix; obtaining displacement information of an X-axis direction in the current monitoring, and a rotation angle of a Y-axis and Z-axis rotation direction; performing corner point matching on the fixed parts to obtain a rotation angle of an X-axis rotation direction in the current monitoring, and displacement information of a Y-axis direction and a Z-axis direction; constructing conversion coefficients of the three axis directions to obtain deformation variables of the three axis directions, and realizing deformation monitoring of the building. The application aims to obtain rotation and translation information through changes of building projection images to monitor deformation variables of the whole building.

Description

A Multi-Parameter-Based Method and System for Building Deformation Monitoring Technical Field

[0001] This invention relates to the field of physical sensor technology, and more specifically to a method and system for monitoring building deformation based on multiple parameters. Background Technology

[0002] Deformation of high-rise buildings caused by uneven loads or foundation settlement requires real-time monitoring to prevent collapse accidents. Monitoring building deformation can promptly identify potential safety hazards and provide data support for maintenance, reinforcement, or emergency decision-making, thereby protecting people's lives and property. Furthermore, deformation monitoring can verify design parameters, provide feedback on construction quality, study normal deformation patterns, and predict deformation, improving the safe operation and operational efficiency of large buildings. Through monitoring, data can be accumulated to optimize design methods, providing a basis for formulating relevant standards and improving designs. Simultaneously, deformation monitoring helps improve building quality by promptly identifying and addressing problems during construction.

[0003] Existing methods for monitoring building deformation include using lidar to acquire 3D point clouds or attaching targets and using visual photography. However, the placement and quality of the attached targets affect the monitoring results, and target movement can lead to misjudgments. Furthermore, the acquired target images are affected by changes in lighting. Lidar, on the other hand, has limited measurement dimensions and requires excessive computation, making it unsuitable for accurate and rapid monitoring of building deformation. Summary of the Invention

[0004] This invention provides a building deformation monitoring method and system based on multiple parameters to solve the problems of excessive computation or numerous interference factors leading to misjudgments in existing lidar or visual photography target-mounted systems. The specific technical solution adopted is as follows:

[0005] This invention proposes a building deformation monitoring method based on multiple parameters, which includes the following steps:

[0006] The system acquires current projected images of the building using acquisition devices and obtains historical projected images for comparison.

[0007] The projected image is divided into a projected portion and a fixed portion. Based on the projected portions of the current projected image and the comparison projected image, corner point matching is performed between the projected portions and displacement information is analyzed to obtain the homography matrix. By performing singular value decomposition on the homography matrix, the rotation matrix and translation matrix are obtained, thereby obtaining the displacement information along the X-axis and the rotation angles along the Y-axis and Z-axis under the current monitoring.

[0008] The fixed portion of the current projected image is adjusted based on the homography matrix to obtain the adjusted image of the current projected image; by performing corner point matching between the fixed portion of the adjusted image and the fixed portion of the comparison projected image, the rotation angle of the X-axis rotation direction, as well as the displacement information of the Y-axis and Z-axis under the current monitoring are obtained based on the position changes of the coordinate points in the matched portion.

[0009] Based on the pixel representation and actual physical properties of the projected image, conversion coefficients for three axes are constructed, and deformation in three axes is obtained based on displacement information in the three axes. Combined with rotation angles in three rotation directions, deformation monitoring of the building is achieved.

[0010] Optionally, the specific method for dividing the projected image into a projected portion and a fixed portion includes:

[0011] The projected image is split by color segmentation. The range of red in the three RGB channels is extracted, and the part of the projected image that is red is used as the projected part; the other parts of the projected image are used as the fixed part.

[0012] Optionally, the specific method for obtaining the homography matrix includes:

[0013] Corner detection is performed on the projected portion of the current projected image and the comparison portion of the comparison projected image to obtain several corner points of the two projected portions.

[0014] By matching the corner points of the two projection parts using the RANSAC algorithm, we can obtain the coordinate point pairs corresponding to the same object in different projection parts.

[0015] Homogeneous equations are constructed using the matched coordinate point pairs to obtain the homography matrix H. The homography matrix H is then solved by constructing a system of equations using four sets of point pairs.

[0016] Optionally, the specific methods for obtaining the rotation matrix and translation matrix include:

[0017] Singular value decomposition is performed on the homography matrix to obtain the rotation matrix R and the translation matrix T.

[0018] Optionally, the displacement information along the X-axis and the rotation angles along the Y and Z axes under the current monitoring are obtained using the following method:

[0019] Based on the translation information of the three axes in the translation matrix, extract the displacement information of the X-axis;

[0020] The rotation matrix R is obtained by multiplying the matrices of the three rotation directions; by obtaining the matrices of the three rotation directions, three angles are obtained, which are the roll angles. Pitch angle and yaw angle ; and the pitch angle Yaw angle is the rotation angle that represents the direction of rotation along the Y-axis. The rotation angle is the direction of rotation along the Z-axis.

[0021] Optionally, the specific method for obtaining the adjusted image of the current projected image includes:

[0022] For any pixel in the current projected image, its coordinates are represented as: Adjustments are made in conjunction with the homography matrix H. The adjustment calculation method is as follows:

[0023]

[0024] Will Use these as the adjusted coordinates for that pixel; obtain the adjusted coordinates for all pixels in the current projected image to obtain the initial adjusted image;

[0025] The initial adjusted image is compensated for the missing pixels relative to the comparison projection image by an interpolation algorithm. The initial adjusted image after interpolation is recorded as the adjusted image of the current projection image.

[0026] Optionally, the specific method for obtaining the rotation angle of the X-axis rotation direction and the displacement information of the Y-axis and Z-axis under the current monitoring is as follows:

[0027] For the adjusted image, the fixed part and the projected part are obtained by color segmentation. Corner detection and corner matching are performed on the fixed part of the adjusted image and the fixed part of the comparison projected image to obtain several matching point pairs in the two fixed parts.

[0028] For any pair of matching points, the difference between the x-coordinate of the fixed part of the adjusted image and the x-coordinate of the fixed part of the projected image is used as the lateral displacement of the pair of matching points; the average of the lateral displacements of all matching points is used as the displacement information along the Y-axis.

[0029] Based on the changes in the ordinates of the points in the fixed part of the adjusted image in the matching point pair and the points in the fixed part of the comparison projection image, the longitudinal displacement of each matching point pair is obtained, and the average of the longitudinal displacements of all matching point pairs is used as the displacement information along the Z-axis.

[0030] For any pair of matching points, obtain the ratio of the longitudinal displacement to the lateral displacement of the pair of matching points, use the ratio as the input of the arctangent function, and use the resulting angle as the roll angle of the pair of matching points; take the mean of the roll angles of all pairs of matching points as the rotation angle in the X-axis rotation direction.

[0031] Optionally, the specific method for constructing the conversion factors for the three axes includes:

[0032] The acquisition device projects the building onto a projection plane, measures the maximum length and width of the projected image on the projection plane, obtains the maximum vertical and horizontal pixel count of the projected image, and records the distance between the acquisition device and the projection plane.

[0033] The ratio of the maximum length to the maximum number of pixels in the vertical direction is used as the Z-axis conversion factor; the ratio of the maximum width to the maximum number of pixels in the horizontal direction is used as the Y-axis conversion factor; the Euclidean norm of the maximum number of pixels in the vertical direction and the maximum number of pixels in the horizontal direction is obtained, and the ratio of the distance between the acquisition device and the projection plane to the Euclidean norm is used as the X-axis conversion factor.

[0034] Optionally, the specific method for obtaining the deformation in the three axes based on the displacement information in the three axes includes:

[0035] The displacement information of the X, Y and Z axes under the current monitoring is multiplied by the conversion coefficient of the corresponding axis, and the product is used as the deformation of the three axes under the current monitoring.

[0036] The present invention also proposes a building deformation monitoring system based on multiple parameters, the system comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of the above method.

[0037] The beneficial effects of this invention are as follows: This invention acquires projected images of fixed objects indoors. Based on the changes in the projected images under different monitoring times, it first quantifies the translation and rotation information of the projected portion and obtains a homography matrix for calibration, eliminating the influence of coordinate system changes on the judgment of the fixed portion. Then, it adjusts the tilt of the fixed portion and analyzes the changes in YZ-axis plane displacement, finally acquiring displacement and rotation information in the three axes and rotation direction. Specifically, by dividing the current projected image into a projected portion and a fixed portion using a comparison projected image, the homography matrix of the projected portion is obtained based on corner point matching, quantifying the translation and rotation information between the two projected portions. The rotation information is obtained by decomposing the homography matrix. The system employs transformation and translation matrices to extract information about the changes in the X-axis and rotational directions of the Y and Z axes of the projected portion, while maintaining a fixed YZ-axis plane. Based on the homography matrix, the current projected image is tilted to adjust the fixed portion using known scale information. This is then matched with the fixed portion of a comparison projected image to determine the changes in the YZ axes of the building itself, thereby extracting information about the changes in the Y-axis, Z-axis, and X-axis rotational directions. This allows for the analysis of displacement and rotational changes of external buildings based on the projection changes of a fixed object, achieving the monitoring of building deformation, with minimal computational load and a convenient calculation method. Attached Figure Description

[0038] To more clearly illustrate the technical solutions 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.

[0039] Figure 1 is a schematic flowchart of a building deformation monitoring method based on multiple parameters provided in an embodiment of the present invention;

[0040] Figure 2 is a schematic diagram of the three axes and three rotational directions in three-dimensional space;

[0041] Figure 3 shows an example of a data acquisition device;

[0042] Figure 4 is a schematic diagram of the process for obtaining deformation based on projected images. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] Please refer to Figure 1, which shows a flowchart of a building deformation monitoring method based on multiple parameters provided in an embodiment of the present invention. The method includes the following steps:

[0045] Step S001: Acquire the current projected image of the building using the acquisition device, and obtain historical projected images as comparison images.

[0046] The purpose of this embodiment is to obtain the relative displacement changes inside the room by projecting the data indoors. It does not require the fabrication and pasting of targets. It achieves deformation measurement through real-time projection and monitors the overall deformation of the indoor walls in three-dimensional space, including deformation in three axes and three rotational directions, i.e., six degrees of freedom.

[0047] It should be noted that the projector projects the same pattern onto the indoor wall, and then obtains the spatial displacement changes by comparing the differences in the projected images, thereby realizing the measurement and monitoring of indoor deformation. The translation of the wall does not affect the projected image, so in addition to the projected image, it is also necessary to combine the projected part of the image and the wall part itself. The calibration and deformation conversion of the target are achieved through the projection and the wall part. The projection can realize the calibration function, that is, the pixels of the projected image are converted into corresponding physical quantities, and the wall part can determine the deformation of the wall.

[0048] It should be further explained that, as shown in Figure 2, the direction towards the camera is the X-axis, the opposite direction to the right is the Y-axis, and the up and down direction is the Z-axis. The three rotation directions are rotations around the corresponding axes. Among them, the projection method can only obtain displacement information in the X-axis direction, the Z-axis rotation direction, and the Y-axis rotation direction. For the displacement in the Z-axis and Y-axis directions, since the wall actually moves but the projection does not move accordingly, it is necessary to combine the wall image and the projection image to obtain the displacement and deformation information in the remaining two axes and the rotation direction around the X-axis.

[0049] Specifically, the acquisition device, as shown in Figure 3, obtains the current projected image by projecting the same pattern onto an indoor wall. It keeps the camera position unchanged from the position when the historical projected image was acquired, acquires the historical projected image, and uses it as a comparison projected image.

[0050] Furthermore, the building deformation monitoring system includes a projection module for projecting laser patterns; a rotation module for controlling the projection direction; an image acquisition module for acquiring projected images; a distance measurement module for assisting focusing, achieving automatic focusing, and obtaining distance information from the camera to the object; a gyroscope module for acquiring the current device's attitude information; and a calculation module for processing the acquired projected images and calculating the corresponding deformation data. It should be noted that due to the certain computational requirements, the monitoring is not real-time but has a lag and is offline.

[0051] Furthermore, as shown in Figure 4, it is the processing flow for obtaining deformation based on the projection image. Projection image A is a comparison projection image of projection image B. A is a historical projection image relative to B. The collected projection image contains the projected pattern and the wall itself, so it is divided into two parts, namely the projection part and the fixed part (building part).

[0052] It should be noted that by comparing the projected images, the transformation matrix, i.e., the homography matrix, can be obtained, and the rotation and translation matrices can be decomposed to extract some deformation information in three-dimensional space. The homography matrix is ​​used to adjust the current projected image to eliminate deformation in some dimensions. Then, the remaining deformation information in three-dimensional space is obtained through the building part of the image. Finally, the conversion coefficient between pixel space and physical quantity is obtained based on the laser projection part, and the complete deformation physical quantity in three-dimensional space is obtained.

[0053] Furthermore, a corresponding camera is configured below the laser head to capture the projected image; the projection structure is rotatable and can obtain the projected image at the corresponding angle according to the set rotation angle, commonly used to obtain the projected image of the four walls of an interior room; at the same time, after the projected image is acquired, conventional image preprocessing is required, including filtering and noise reduction processing.

[0054] It should be further noted that the projection result of the projected image is affected by the changes in projection distance and projection surface. Changes in the distance to the projection surface and the tilt angle of the projection surface will cause corresponding changes in the projected image. By extracting the differences in the projected image, the relationship between image differences and deformation information can be constructed to realize the monitoring of building deformation.

[0055] Step S002: Divide the projected image into a projected part and a fixed part; based on the projected parts of the current projected image and the comparison projected image, perform corner point matching and analyze displacement information between the projected parts to obtain the homography matrix; by performing singular value decomposition on the homography matrix, obtain the rotation matrix and translation matrix, and then obtain the displacement information of the X-axis under the current monitoring, as well as the rotation angles of the Y-axis and Z-axis rotation directions.

[0056] It should be noted that some of the deformations of the building walls being monitored can be reflected in changes in the projected image, such as near objects appearing larger and far objects appearing smaller. However, some deformations do not cause changes in the projected image. Therefore, it is necessary to split the projected image into a projection portion of a specific color (red, green) and the building portion with the wall fixed (the wall portion).

[0057] Preferably, in one embodiment of the present invention, the specific method for dividing the projected image into a projection portion and a fixed portion includes:

[0058] The projected image is split by color segmentation. After acquiring the RGB three-channel color projected image, the range of a specific color, such as red, in the three RGB channels is determined. In this embodiment, the range of red in the three RGB channels is extracted, and the part of the projected image that is red is taken as the projected part. The specific range of red in the three RGB channels is determined based on the existing range, which will not be described in detail in this embodiment. The other parts in the projected image are taken as fixed parts.

[0059] It should be further explained that during the acquisition of projection images from different monitoring sessions, the camera position remains unchanged. Therefore, the change in the projection image is caused by the motion of the object relative to the camera. The transformation between two images is a planar homography transformation, which can be represented by a 3×3 matrix (homography matrix). This matrix encodes the rigid body motion (rotation + translation) of the object from position 1 to position 2. Therefore, it is necessary to acquire the homography matrix of the projection portion of the projection images from different monitoring sessions, and then obtain the deformation information by processing the matrix. The homography matrix, also known as the homography matrix or projection matrix, is a linear transformation matrix in Euclidean space that maps one plane to another. It can describe the mapping relationship between two planes.

[0060] Preferably, in one embodiment of the present invention, based on the projected portions of the current projected image and the comparison projected image, corner point matching is performed between the projected portions and displacement information is analyzed to obtain a homography matrix. The specific method includes:

[0061] Harris corner detection is performed on the projected portion of the current projected image and the comparison portion of the comparison projected image to obtain several corner points in both projections. The corner points are then matched using the RANSAC (Random Sample Consensus) algorithm, which employs iterative and fitting methods. By identifying corner points with identical descriptions in both projections, pairs of coordinate points corresponding to the same object are obtained across different projections. Each pair of points takes the form of... ,in This represents the corner coordinates in the projected portion AT of the contrasted projected image. This indicates the coordinates of the corner points that were successfully matched in the projection portion BT of the current projected image.

[0062] Furthermore, a homography equation is constructed using the matched corner points, i.e., the coordinate pairs, to obtain the homography matrix H, where the homography matrix H has the following form:

[0063]

[0064] The homography matrix H is solved by constructing a system of equations using four sets of points (since there are eight unknowns in the homography matrix H, a homogeneous equation is constructed using four sets of points to solve it). The system of equations is as follows:

[0065]

[0066] Solving the system of equations yields eight unknowns, which in turn lead to the corresponding homography matrix H. It should be noted that Harris corner detection, the RANSAC algorithm, and the homography matrix solution are all existing algorithms, and the specific calculation process will not be described in this embodiment.

[0067] It should be further noted that the homography matrix is ​​a linear transformation matrix under the corner matching result, which also includes the translation and rotation information between corners. Therefore, it needs to be decomposed to separate the rotation and translation information.

[0068] Preferably, in one embodiment of the present invention, by performing singular value decomposition on the homography matrix to obtain the rotation matrix and translation matrix, the displacement information along the X-axis and the rotation angles along the Y-axis and Z-axis under the current monitoring are obtained. The specific method includes:

[0069] Singular value decomposition of the homography matrix yields the rotation matrix R and the translation matrix T, in the following matrix forms:

[0070]

[0071]

[0072] in, , and These represent translation information in the three axial directions: X-axis, Y-axis, and Z-axis, respectively.

[0073] It should be further explained that, for the changes in the projection part, since the plane forming the projection part is fixed (YZ axis plane), the changes in the projection part in the three-dimensional coordinate system are mainly manifested as the displacement of the X axis and the axial rotation changes of the Y and Z axes. Therefore, the above displacement and rotation information are extracted in the rotation matrix and translation matrix.

[0074] Furthermore, based on the translation information along the three axes in the translation matrix, the displacement information along the X-axis is extracted. This extraction process is performed using existing methods and will not be elaborated upon in this embodiment. Simultaneously, the rotation matrix R is obtained by multiplying the matrices of the three rotation directions, expressed as:

[0075]

[0076] in, This is the roll angle, which is the rotation angle in the X-axis rotation direction; This is the pitch angle, which is the rotation angle in the direction of rotation along the Y-axis; The yaw angle is the rotation angle along the Z-axis. By obtaining the matrices for the three rotation directions, three angles are obtained, including the pitch angle. Yaw angle is the rotation angle that represents the direction of rotation along the Y-axis. The rotation angle is the direction of rotation along the Z-axis.

[0077] Thus, by dividing the current projected image and the comparison projected image into a projected part and a fixed part, the homography matrix of the projected part is obtained based on corner matching, and the translation and rotation information between the two projected parts is quantified; by decomposing the homography matrix, the rotation matrix and translation matrix are obtained, thereby extracting the information on the changes in the X-axis and the rotation direction of the Y-axis and Z-axis of the projected part under the premise that the YZ-axis plane is fixed.

[0078] Step S003: Adjust the fixed part of the current projected image based on the homography matrix to obtain the adjusted image of the current projected image; by performing corner point matching between the fixed part of the adjusted image and the fixed part of the comparison projected image, and based on the position change of the coordinate points in the matched part, obtain the rotation angle of the X-axis rotation direction, as well as the displacement information of the Y-axis and Z-axis under the current monitoring.

[0079] It should be noted that partial deformation information is extracted by changing the projected portion, but partial deformation does not cause changes in the projected portion. Therefore, changes in the fixed portion, i.e., the wall itself, need to be utilized. Since the scale information of the projected portion is known, i.e., the projection plane is fixed, the changes in the wall portion are extracted by calibrating the projected portion. This requires tilting the fixed portion of the current projected image based on the homography matrix, adjusting the fixed portion under different coordinate systems to the same coordinate system, which can eliminate the changes in the rotational directions of the X-axis and the Y and Z axes.

[0080] Preferably, in one embodiment of the present invention, adjusting a fixed portion of the current projected image based on the homography matrix to obtain an adjusted image of the current projected image includes the following specific method:

[0081] For any pixel in the current projected image, its coordinates are represented as: Adjustments are made in conjunction with the homography matrix H. The adjustment calculation method is as follows:

[0082]

[0083] Will The initial adjusted image is obtained by taking the adjusted coordinates of all pixels in the current projected image as the coordinates of the pixel. The initial adjusted image is then obtained by taking the adjusted coordinates of all pixels in the current projected image as the coordinates of the pixel. The pixel values ​​of the missing pixels in the initial adjusted image relative to the comparison projected image are obtained by interpolation. The initial adjusted image after interpolation is recorded as the adjusted image of the current projected image.

[0084] It should be further noted that, generally speaking, the adjusted image and the comparison projection image are consistent in orientation in all directions except for the Y-axis, Z-axis and X-axis rotation direction. Therefore, the change information is quantified by the coordinate changes of the fixed part of the adjusted image and the fixed part of the comparison projection image.

[0085] Preferably, in one embodiment of the present invention, by performing corner point matching between a fixed portion of the adjusted image and a fixed portion of the comparison projection image, and based on the positional changes of the coordinate points in the matched portion, the rotation angle of the X-axis rotation direction and the displacement information of the Y-axis and Z-axis under the current monitoring are obtained. The specific method includes:

[0086] For the adjusted image, the fixed part and the projected part are obtained by color segmentation. Corner detection and corner matching are performed on the fixed part (building part) BT-J of the adjusted image and the fixed part (building part) AJ of the comparison projected image to obtain several matching point pairs in the two fixed parts.

[0087] Furthermore, for any pair of matching points, the difference between the x-coordinate of the fixed portion of the adjusted image and the x-coordinate of the fixed portion of the projected image is used as the lateral displacement of the pair of matching points; the average of the lateral displacements of all pairs of matching points is used as the displacement information along the Y-axis.

[0088] It should be noted that, The two-dimensional coordinates actually represent the coordinate information in the adjusted image. In three-dimensional space, this corresponds to changes in the YZ plane; that is, lateral displacement in the two-dimensional coordinate system is actually a change in the Y-axis in the three-dimensional coordinate system. Similarly, longitudinal displacement is a change in the Z-axis. Based on the changes in the Y and Z axes, the rotation direction (roll angle) along the X-axis is obtained using inverse trigonometric functions. The changes.

[0089] Furthermore, based on the changes in the ordinates of the points in the fixed portion of the adjusted image and the points in the contrasted projected image in the matching point pair, the longitudinal displacement of each matching point pair is obtained, and the average of the longitudinal displacements of all matching point pairs is used as the displacement information along the Z-axis.

[0090] Furthermore, for any pair of matching points, the ratio of the longitudinal displacement to the lateral displacement of the pair of matching points is obtained, and the ratio is used as the input of the arctangent function. The angle obtained from the output is used as the roll angle of the pair of matching points. The mean of the roll angles of all pairs of matching points is used as the rotation angle in the X-axis rotation direction.

[0091] At this point, the current projected image is tilted based on the homography matrix to adjust the fixed part using the known scale information of the projected part. By matching it with the fixed part of the comparison projected image, the changes in the Y and Z axes of the building part itself are made, and then the information on the changes in the Y-axis, Z-axis and X-axis rotation direction is extracted.

[0092] Step S004: Based on the pixel performance and actual physical performance of the projected image, construct conversion coefficients for three axes, and obtain deformation for three axes based on displacement information of the three axes. Combine the rotation angles of the three rotation directions to realize deformation monitoring of the building.

[0093] It should be noted that since the displacement information in the three axes is actually obtained based on the changes in the projected image, its dimension is in pixels and not the actual physical quantity. It is also necessary to combine the acquisition equipment to obtain the conversion coefficient between the projected image and the real physical quantity, and then obtain the deformation of the three axes of displacement information. The calculation is performed by obtaining the relationship function between the pixels of the projected image and the object distance. Since the projected image is known, the conversion coefficient is obtained by the degree of near-large and far-small.

[0094] Specifically, the acquisition device projects the building onto a projection plane (wall), measures the maximum length (vertical, which is the actual physical size) and maximum width (horizontal, which is the actual physical size) of the projected image on the projection plane, and simultaneously obtains the maximum number of vertical and horizontal pixels of the projected image. It also records the distance between the acquisition device and the projection plane, which is directly obtained through the distance sensor in the acquisition device.

[0095] Furthermore, the ratio of the maximum length to the maximum number of vertical pixels is used as the Z-axis conversion factor; the ratio of the maximum width to the maximum number of horizontal pixels is used as the Y-axis conversion factor; the Euclidean norm of the maximum number of vertical pixels and the maximum number of horizontal pixels (i.e., the root of the sum of squares) is obtained, and the ratio of the distance between the acquisition device and the projection plane to the Euclidean norm is used as the X-axis conversion factor; wherein the unit of the three-axis conversion factors in this embodiment is millimeters / pixel.

[0096] Furthermore, by multiplying the displacement information of the X, Y, and Z axes under the current monitoring with the corresponding conversion coefficients, the product is taken as the deformation of the three axes under the current monitoring. At the same time, the rotation angle of the rotation direction of the three axes has been obtained, thus realizing the deformation monitoring of the building.

[0097] Thus, by acquiring projected images of fixed indoor objects, and based on the changes in the projected images under different monitoring times, the translation and rotation information of the projected part is first quantified, and the homography matrix is ​​obtained for calibration to eliminate the influence of coordinate system changes on the judgment of the fixed part. Then, the tilt of the fixed part is adjusted, and the displacement changes of the YZ axis plane are analyzed. Finally, displacement and rotation information in the three axes and rotation direction are obtained. Thus, with a small amount of computation and a convenient calculation method, the displacement and rotation changes of external buildings can be analyzed based on the projection changes of fixed objects, thereby achieving the monitoring of building deformation.

[0098] Another embodiment of the present invention provides a building deformation monitoring system based on multiple parameters. The system includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the above-described method steps S001 to S004.

[0099] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A building deformation monitoring method based on multiple parameters, characterized in that, The method includes the following steps: acquiring a current projected image of the building using a data acquisition device, and obtaining historical projected images as comparison images; splitting the projected image using color segmentation, extracting the range of red in the RGB channels, and representing the red portion of the projected image as the projected portion; treating the other portions of the projected image as fixed portions; based on the projected portions of the current and comparison images, performing corner point matching and analyzing displacement information between the projected portions to obtain a homography matrix; performing singular value decomposition on the homography matrix to obtain a rotation matrix and a translation matrix, thereby obtaining the displacement information along the X-axis and the rotation angles along the Y and Z axes under the current monitoring; and based on the homography... The property matrix adjusts the fixed portion of the current projected image to obtain the adjusted image. Corner point matching is performed between the fixed portion of the adjusted image and the fixed portion of the comparison projected image. Based on the positional changes of coordinate points in the matched portions, the rotation angle along the X-axis and the displacement information along the Y and Z axes under the current monitoring are obtained. The specific method includes: for the adjusted image, obtaining the fixed and projected portions through color segmentation; performing corner point detection and matching between the fixed portions of the adjusted image and the fixed portions of the comparison projected image to obtain several pairs of matching points in the two fixed portions; for any pair of matching points, the abscissa of the point in the fixed portion of the adjusted image in that pair is determined. The difference obtained by subtracting the x-coordinate of the points in the fixed portion of the comparison projection image is taken as the lateral displacement of the matching point pair; the mean of the lateral displacements of all matching point pairs is taken as the displacement information along the Y-axis; based on the change in y-coordinate of the points in the fixed portion of the adjusted image and the points in the fixed portion of the comparison projection image in the matching point pair, the longitudinal displacement of each matching point pair is obtained, and the mean of the longitudinal displacements of all matching point pairs is taken as the displacement information along the Z-axis; for any matching point pair, the ratio of the longitudinal displacement to the lateral displacement of the matching point pair is obtained, and the ratio is used as the input of the arctangent function, and the resulting angle is taken as the roll angle of the matching point pair; the mean of the roll angles of all matching point pairs is taken as the rotation angle in the X-axis rotation direction; Based on the pixel representation and actual physical properties of the projected image, conversion coefficients are constructed along three axes, and deformation variables along these three axes are obtained based on displacement information. Combined with rotation angles in three rotational directions, deformation monitoring of the building is achieved. The homography matrix is ​​obtained through the following specific methods: corner point detection is performed on the projected portion of the current projected image and the comparison portion of the comparison projected image to obtain several corner points of the two projected portions; the corner points of the two projected portions are matched using the RANSAC algorithm to obtain coordinate point pairs corresponding to the same object in different projected portions; homogeneous equations are constructed using the matched coordinate point pairs to obtain the homography matrix H; and the homography matrix H is solved by constructing a system of equations using four sets of point pairs.

2. The building deformation monitoring method based on multiple parameters according to claim 1, characterized in that, The specific method for obtaining the rotation matrix and translation matrix includes: performing singular value decomposition on the homography matrix to obtain the rotation matrix R and the translation matrix T.

3. The building deformation monitoring method based on multiple parameters according to claim 2, characterized in that, The displacement information along the X-axis and the rotation angles along the Y and Z axes under the current monitoring are obtained as follows: The displacement information along the X-axis is extracted from the translation information of the three axes in the translation matrix; the rotation matrix R is obtained by multiplying the matrices of the three rotation directions; by obtaining the matrices of the three rotation directions, three angles are obtained, namely the roll angle. Pitch angle and yaw angle ; and the pitch angle Yaw angle is the rotation angle that represents the direction of rotation along the Y-axis. The rotation angle is the direction of rotation along the Z-axis.

4. The building deformation monitoring method based on multiple parameters according to claim 1, characterized in that, The specific method for obtaining the adjusted image of the current projected image is as follows: for any pixel in the current projected image, its coordinates are represented as... Adjustments are made in conjunction with the homography matrix H. The adjustment calculation method is as follows: Will The initial adjusted image is obtained by taking the adjusted coordinates of all pixels in the current projected image and using an interpolation algorithm to compensate for the missing pixels in the initial adjusted image relative to the comparison projected image. The initial adjusted image after interpolation is recorded as the adjusted image of the current projected image.

5. The building deformation monitoring method based on multiple parameters according to claim 1, characterized in that, The specific method for constructing the conversion coefficients for the three axes includes: the acquisition device projects the building onto a projection plane, measures the maximum length and maximum width of the projected image on the projection plane, obtains the maximum vertical and horizontal pixel counts of the projected image, and records the distance between the acquisition device and the projection plane; the ratio of the maximum length to the maximum vertical pixel count is used as the Z-axis conversion coefficient; the ratio of the maximum width to the maximum horizontal pixel count is used as the Y-axis conversion coefficient; the Euclidean norm of the maximum vertical and horizontal pixel counts is obtained, and the ratio of the distance between the acquisition device and the projection plane to the Euclidean norm is used as the X-axis conversion coefficient.

6. The building deformation monitoring method based on multiple parameters according to claim 5, characterized in that, The method for obtaining the deformation of the three axes based on the displacement information of the three axes includes: multiplying the displacement information of the X, Y and Z axes under the current monitoring by the conversion coefficient of the corresponding axis, and the product is used as the deformation of the three axes under the current monitoring.

7. A building deformation monitoring system based on multiple parameters, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the building deformation monitoring method based on multiple parameters as described in any one of claims 1-6.

Images