A hydraulic hoisting tower structure deformation measurement method based on edge feature direction code matching

Abstract

The application discloses a kind of based on edge feature direction code matching's hydraulic hoist tower structure deformation measurement method, belong to computer vision measurement field, this method includes the following steps: using subpixel edge detection method to the detection area image is handled, obtain subpixel edge line position and subpixel edge direction;Based on subpixel edge direction, using direction code encoding method to the edge direction angle is handled, obtain the edge direction code of discretization, and based on the edge direction code of discretization the edge direction code of left and right camera image is calculated;Based on binocular stereo vision system, using template matching method and parallax principle to the edge direction code of left and right camera image is matched, calculate the spatial three-dimensional coordinates of detection area matching coefficient maximum point;Based on the spatial three-dimensional coordinates change of same detection point before and after structure deformation, obtain the deformation of hydraulic hoist tower structure.

Description

Technical Field

[0001] This invention belongs to the field of computer vision measurement technology, and in particular relates to a method for measuring the deformation of a hydraulic lifting tower structure based on edge feature direction code matching. Background Technology

[0002] Currently, the primary method for measuring the deformation of hydraulic lifting towers is using total stations. In this process, key points on the tower are measured before and after lifting the load, and the deformation is calculated by measuring the distance the coordinates of the same point change. However, the use of total stations has significant limitations, such as cumbersome operation, low automation, and low measurement efficiency, making it difficult to monitor the deformation of hydraulic lifting towers. With the continuous advancement of measurement technology, building a measurement system based on advanced technology to achieve efficient, accurate, and highly automated tower deformation monitoring has become an urgent problem to be solved.

[0003] Visual measurement of structural deformation is an emerging measurement technology that has shown great application potential in the field of deformation measurement due to its advantages such as simple operation, non-contact nature, and strong real-time performance. The main principle of visual sensor-based structural deformation measurement methods is to monitor tower deformation through template matching technology. Specifically, templates are first matched in images of the tower before and after deformation. The peak point of the template matching coefficient is used as the tracking point, and then the amount of deformation is determined by calculating the distance between the tracking points before and after deformation.

[0004] However, in actual outdoor deformation measurement, the accuracy of visual measurement methods remains a major challenge. Template matching methods heavily rely on image grayscale information, while tower structures are typically located outdoors and are susceptible to interference from factors such as uniform coating and changes in lighting, leading to reduced image contrast. When the contrast of the target area is low, even small environmental changes can affect the accuracy of image template matching, thus reducing measurement precision. To improve measurement accuracy, current visual measurement methods for monitoring tower structure deformation typically require the installation of artificial markers on the structural surface. This approach not only increases the difficulty of measuring tower deformation but also reduces work efficiency. Therefore, this invention proposes a method for measuring the deformation of hydraulic lifting tower structures based on edge feature direction code matching. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention proposes a method for measuring the deformation of a hydraulic lifting tower structure based on edge feature direction code matching, thereby resolving the issues present in the prior art.

[0006] To achieve the above objectives, the present invention provides a method for measuring the deformation of a hydraulic lifting tower structure based on edge feature direction code matching, comprising:

[0007] The subpixel edge detection method is used to process the image of the detection area to obtain the subpixel edge line position and subpixel edge direction;

[0008] Based on the sub-pixel edge direction, the edge direction angle is processed by the direction code encoding method to obtain a discrete edge direction code, and the edge direction codes of the left and right camera images are calculated based on the discrete edge direction code.

[0009] Based on a binocular stereo vision system, the template matching method and the parallax principle are used to match the edge direction codes of the left and right camera images, and the spatial three-dimensional coordinates of the point with the maximum matching coefficient in the detection area are calculated.

[0010] The deformation of the hydraulic lifting tower structure is obtained based on the change in the three-dimensional coordinates of the same detection point before and after structural deformation.

[0011] Optionally, the process of processing the image of the detection region using a sub-pixel edge detection method to obtain the sub-pixel edge position and sub-pixel edge direction includes:

[0012] Based on the light intensity distribution characteristics of pixels within the detection area, a weighted average method is used to process edge pixels to obtain light intensity feature values;

[0013] Based on the light intensity feature value, the edge straightness coefficient is obtained by processing the ROI matrix;

[0014] Based on the edge straightness coefficient, the edge line position is processed using a straight line equation to obtain the sub-pixel edge line position;

[0015] Based on the sub-pixel edge line position, the edge direction is processed using a normal vector calculation method to obtain the sub-pixel edge direction.

[0016] Optionally, the calculation expressions for the first light intensity value and the second light intensity value are as follows:

[0017]

[0018] In the formula, C is the first light intensity value, O is the second light intensity value, d1, d2, and d3 are the first center distance, the second center distance, and the third center distance, respectively, and F... i+1,j-1 Let F be the light intensity of pixel (i+1, j-1). i,j-2 Let F be the light intensity of pixel (i,j-2). i+1,j-2 Let F be the light intensity of pixel (i+1, j-2). i-1,j+2 Let F be the light intensity of pixel (i-1, j+2). i,j+2Let F be the light intensity of pixel (i,j+2). i-1,j+1 Let (i-1,j+1) be the light intensity of pixel (i-1,j+1).

[0019] Optionally, the expression for processing the edge direction angle using the direction code encoding method is as follows:

[0020]

[0021] In the formula, C i,j For direction code, Δθ is the equal-width threshold, and θ i,j This is the direction angle.

[0022] Optionally, the process of calculating the edge orientation codes of the left and right camera images based on the discretized edge orientation codes includes:

[0023] Based on the detection area image acquired by the left camera, the direction code encoding method is used to process the sub-pixel edge direction to obtain the edge direction code of the left camera image;

[0024] Based on the detection area image acquired by the right camera, the same direction code encoding method is used to process the sub-pixel edge direction to obtain the edge direction code of the right camera image;

[0025] Based on a preset threshold angle, the edge direction codes of the left and right camera images are uniformly encoded using a direction code discretization method, resulting in re-encoded edge direction codes of the left and right camera images.

[0026] Optionally, the process of matching the edge direction codes of the left and right camera images using template matching and parallax principles includes:

[0027] Based on the edge direction codes of the left and right camera images, a digital image correlation matching algorithm is used to process the data and search for the global extreme point of the correlation coefficient.

[0028] Based on the extreme point pairs matched by the global extreme points, a binocular stereo vision geometric model is used for processing to calculate disparity information;

[0029] Based on the camera calibration parameters and the parallax information, a three-dimensional coordinate reconstruction algorithm is used to process the data and obtain the spatial three-dimensional coordinates of the matching point.

[0030] Optionally, the process of obtaining the deformation of the hydraulic lifting tower structure based on the spatial three-dimensional coordinate changes of the same detection point before and after structural deformation includes:

[0031] Based on the three-dimensional coordinate dataset of the detection points before deformation, a spatial point cloud registration method is used to process the data and establish a reference coordinate system.

[0032] Based on the 3D coordinate dataset of the detected points after deformation, the same spatial point cloud registration method is used to process the data to obtain the deformed coordinate system.

[0033] Based on the reference coordinate system and the deformed coordinate system, the displacement of each detection point is obtained by processing using the Euclidean distance calculation method.

[0034] Based on the distribution of the displacement, a statistical analysis algorithm is used to process the data and obtain the overall deformation characteristic parameters of the tower structure.

[0035] Compared with the prior art, the present invention has the following advantages and technical effects:

[0036] The hydraulic lifting tower structure deformation measurement method based on edge feature directional code matching provided by this invention has significant technical advantages. First, by introducing sub-pixel edge detection technology, it overcomes the accuracy limitations of traditional pixel-level detection, enabling more accurate identification of structural edge features and providing a more reliable data foundation for deformation measurement. Second, the innovative directional code encoding method effectively improves the robustness of feature matching, overcoming the problem of unstable matching results under complex lighting conditions in traditional methods. Third, the combination of binocular stereo vision and sub-pixel features enables non-contact measurement without manual targets, significantly simplifying on-site operation procedures. Furthermore, this method significantly improves computational efficiency by optimizing feature extraction and matching algorithms, meeting the needs of real-time monitoring in engineering projects. Compared to existing technologies, this invention significantly improves measurement accuracy, environmental adaptability, ease of operation, and system real-time performance, providing a more advanced technical solution for the structural safety monitoring of hydraulic lifting towers. Attached Figure Description

[0037] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0038] Figure 1 This is a flowchart of the hydraulic lifting tower structure deformation measurement method based on edge feature direction code matching according to an embodiment of the present invention;

[0039] Figure 2 This illustrates the relationship between the pixel area occupied by different light intensities and the pixel area in an embodiment of the present invention.

[0040] Figure 3 This is a schematic diagram of the sub-pixel edge detection principle in an embodiment of the present invention, wherein (a) is the selection of ROI, (b) is the establishment of a coordinate system, and (c) is the ideal light intensity distribution.

[0041] Figure 4 This diagram illustrates the determination of the position and direction of the edge line within a pixel in an embodiment of the present invention.

[0042] Figure 5 This is a schematic diagram of the direction code according to an embodiment of the present invention;

[0043] Figure 6 This is a schematic diagram illustrating the principle of calculating the three-dimensional coordinates of spatial points in binocular stereo vision measurement according to an embodiment of the present invention. Detailed Implementation

[0044] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0045] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0046] like Figure 1 As shown, this embodiment provides a method for measuring the deformation of a hydraulic lifting tower structure based on edge feature direction code matching. This invention utilizes binocular stereo vision theory, combined with sub-pixel edge detection technology and template matching technology, to propose a method for measuring the deformation of a hydraulic lifting tower structure based on sub-pixel edge direction code matching. This achieves visual measurement of the structural deformation of the hydraulic lifting tower structure without artificial targets. The method specifically includes the following steps:

[0047] S1. First, select a region of interest on the surface of the tracked object as the detection region. At the same time, perform morphological processing on the detection region to improve image quality, and perform sub-pixel edge detection on the image within the detection region of the left image.

[0048] S2. Then, an edge direction code is generated by extracting sub-pixel level edge directions;

[0049] S3. Next, template matching is performed on the direction codes of the left and right image detection areas before the tower structure deformation using the parallax triangle principle, and the three-dimensional spatial coordinates of the detection areas before deformation are calculated. Similarly, template matching is performed on the direction codes of the left and right image detection areas after the tower structure deformation using the parallax triangle principle, and the three-dimensional spatial coordinates of the detection areas after deformation are calculated.

[0050] S4. Finally, the deformation of the tower structure is obtained by the distance between the spatial coordinates of the measured position in the tower structure before and after deformation.

[0051] As a specific implementation method of this embodiment, the specific implementation process of S1-S4 includes:

[0052] S1. The subpixel edge detection method is used to process the image of the detection area to obtain the subpixel edge line position and subpixel edge direction. First, a region of interest (ROI) is selected on the surface of the tracked object as the detection area, and subpixel edge detection is performed on the image within the detection area. The subpixel edge detection algorithm is then optimized.

[0053] Furthermore, the process of processing the detection area image using a sub-pixel edge detection method to obtain the sub-pixel edge line position and sub-pixel edge direction includes: processing the edge pixels using a weighted average method based on the light intensity distribution characteristics of the pixels within the detection area to obtain light intensity feature values; processing the ROI matrix based on the light intensity feature values ​​to obtain edge straightness coefficients; processing the edge line position using a straight line equation based on the edge straightness coefficients to obtain the sub-pixel edge line position; and processing the edge direction using a normal vector calculation method based on the sub-pixel edge line position to obtain the sub-pixel edge direction.

[0054] The optimization of the sub-pixel edge detection algorithm in S1 specifically includes the following steps: Pixel elements (CCD, CMOS) are light integrating devices. Their main principle is to integrate the light intensity projected onto a pixel of a fixed area within a fixed time interval, and the output is the sampled value of the light intensity. Ideally, the edge projection of an object will divide its pixel unit into two parts with different light intensities. Therefore, a reasonable assumption is that when the edge line passes through pixel (i,j), the light intensity value conforms to the area-weighted average principle, as stated below:

[0055]

[0056] Among them, S C S represents the area corresponding to the portion of pixel (i,j) with a light intensity value of C. O Let h be the area corresponding to the portion of pixel (i,j) with a light intensity value of 0, and let h be the side length of pixel (i,j). 2 =S C +S O The relationship between the pixel area occupied by different light intensities and the pixel area is as follows: Figure 2 As shown.

[0057]

[0058] Based on the above principles, a sub-pixel edge detection algorithm for images is proposed. To apply the proposed method, a masking method is first used to detect which pixels belong to edges. Then, the sub-pixel position of the edge in each edge pixel is calculated. Finally, a sub-pixel edge line with accurate location is estimated.

[0059] Before applying subpixel-based edge detection, individual edge pixels of an image were calculated using traditional derivative masks, such as partial derivatives. After obtaining the individual edge pixels of the image, the row and column differences f of each pixel were calculated. x f y Then, according to f x f y The size relationship is determined by cropping pixel matrices of different sizes from the original image to determine the values ​​of intermediate parameters C, O, S1, S2, and S3, ultimately determining the coefficients of the edge lines and the specific locations of the edge pixels. Now, based on the f of the edge pixels... x f y The size relationships are discussed separately.

[0060] (1)|f x |≤|f y |

[0061] Since the edge lines of a structure can be straight lines or curves, for ease of calculation, this invention assumes that each edge line consists of small straight line segments within a pixel. The number of straight line segments in each edge line is the same as the number of pixels the edge line passes through. Therefore, this invention mainly studies straight line edge detection. Assume the expression for the edge line is y = a + bx, where the slope |b| ≤ 1. For simplicity, assume the slope of the line is between (0, 1), and the number of pixels to be detected is p. (i，j) , with p (i，j) Select a 3×5 pixel region (ROI) centered on the pixel, such as... Figure 3 As shown in (a), a rectangular coordinate system p-xy is established with the center of the ROI as the origin, as follows: Figure 3 As shown in (b). Ideally, the straight lines at the edges of the image divide the edge pixels into two parts with different light intensities. Assuming the light intensity gray values ​​are C and O respectively, the ideal light intensity distribution is as follows. Figure 3 As shown in (c).

[0062] Based on the method proposed by Trujillo-Pino et al., according to equation (2), it is assumed that F1, F2, and F3 are the sum of the gray values ​​of each column of pixels within the ROI, satisfying the equation:

[0063]

[0064] Where k = 1, 2, 3, and S1, S2, S3 represent the area of ​​the lower part of the straight edge of each column of pixels in the three columns of pixels within the ROI, respectively:

[0065]

[0066] The expressions for the linear coefficients a and b can be calculated based on equation (4):

[0067]

[0068] C and O are unknown. Due to the optical effects on the edges, the pixels at the edge of the structure are not clearly divided into two parts; the pixel values ​​undergo a change process, such as... Figure 3 As shown in (b). Therefore, the values ​​of C and O should be calculated from the light intensity values ​​of several pixels diagonally opposite the center pixel within the ROI. Since the distances between neighboring pixels and the center pixel are different, the contribution of the pixel's light intensity to the center pixel is also different, and its contribution conforms to the weighted average principle with respect to the pixel's distance. Specifically, taking the center pixel as the center of symmetry, three pixels at the diagonal of the edge line within the ROI are selected, and the center distances from the center pixel p(i,j) to the center of the three diagonal pixels are calculated, where the center distances are d1, d2, and d3, respectively. Figure 3 As shown in (b). Therefore, the values ​​of C and O can be calculated using the following formula:

[0069]

[0070] In the formula, C is the first light intensity value, O is the second light intensity value, d1, d2, and d3 are the first center distance, the second center distance, and the third center distance, respectively, and F... i+1,j-1 Let F be the light intensity of pixel (i+1, j-1). i,j-2 Let F be the light intensity of pixel (i,j-2). i+1,j-2 Let F be the light intensity of pixel (i+1, j-2). i-1,j+2 Let F be the light intensity of pixel (i-1, j+2). i,j+2 Let F be the light intensity of pixel (i,j+2). i-1,j+1 Let be the light intensity of pixel (i-1, j+1). In equations (7) and (8), since the pixel is a square with a side length of 1 pixel, the values ​​of d1, d2, and d3 are respectively... 2 pixels

[0071] (2)|f x |>|f y |

[0072] At this point, the slope of the line within the given range is |b| > 1. When the line is perpendicular to the horizontal axis, the slope b is infinite. Therefore, for ease of calculation, a horizontal window is selected within this range, and a 5×3 pixel region (ROI) is chosen. The line expression is then x = a + by. The calculation method for the line expression parameters is the same as described above.

[0073] In summary, based on the above calculation method and pixel information, the coefficients of the straight line within each edge pixel can be obtained. After obtaining the expression for the edge straight line, the sub-pixel position of the edge straight line within each edge pixel can be obtained using the following method: the intersection point p1 of the edge line and the y-axis is taken as the position of the edge line within the pixel, and N is the normal vector of the edge straight line within that pixel. The sub-pixel edge line position and edge direction are as follows: Figure 4 As shown.

[0074] S2. Based on the sub-pixel edge direction, a direction code encoding method is used to process the edge direction angle to obtain a discretized edge direction code. The edge direction codes of the left and right camera images are then calculated based on this discretized edge direction code. The sub-pixel edge direction code of the detection region image is calculated using the proposed edge direction code method. Specifically, the sub-pixel edge direction code algorithm in S2 includes the following steps:

[0075] After the sub-pixel edge detection algorithm described above, the angle between the edge line within each edge pixel and the x-axis in the pixel coordinate system can be calculated. Due to structural deformation and movement, to improve the robustness of the matching process, this embodiment proposes a direction code matching algorithm. Assume that in the image, the normal vector N of the sub-pixel edge line within the pixel at position i in row j is... i,j If (x, y), then the direction angle θ i,j It can be calculated using the following formula:

[0076]

[0077] The direction code is obtained by setting a constant width threshold Δθ, and the specific calculation formula is as follows:

[0078]

[0079] Examples of the direction codes proposed in this invention are as follows: Figure 5 As shown, the corresponding threshold Δθ = π / 8 in the example. Each edge pixel has an edge orientation angle, and each edge pixel is re-encoded using this method.

[0080] This method involves re-encoding all edge lines of the image at the measured location, then performing template matching between the images of the detection area before and after deformation. The coordinates and deformation amount of the measured location before and after deformation are calculated based on the location of the maximum correlation coefficient in the matching results.

[0081] Furthermore, the process of calculating the edge direction codes of the left and right camera images based on the discretized edge direction codes includes: processing the sub-pixel edge directions using a direction code encoding method based on the detection area image acquired by the left camera to obtain the edge direction codes of the left camera image; processing the sub-pixel edge directions using the same direction code encoding method based on the detection area image acquired by the right camera to obtain the edge direction codes of the right camera image; and uniformly encoding the edge direction codes of the left and right camera images using a direction code discretization method based on a preset threshold angle to obtain the edge direction codes of the left and right camera images after the edge lines have been re-encoded.

[0082] S3. Based on a binocular stereo vision system, the edge direction codes of the left and right camera images are matched using template matching and the parallax principle. The spatial three-dimensional coordinates of the point with the maximum matching coefficient in the detection region are calculated. Using the calculated edge direction codes of the left and right camera images, the three-dimensional coordinates of the point with the maximum matching coefficient in the detection region are calculated using the template matching principle and the parallax principle.

[0083] In S3, the template matching algorithm mainly refers to digital image correlation matching, which is achieved by searching for the global extremum of the correlation function. Specifically, it uses the parallax triangle principle to perform template matching on the direction codes of the detection areas in the left and right images before the tower structure deforms, calculating the three-dimensional spatial coordinates of the detection areas before deformation. Similarly, it uses the parallax triangle principle to perform template matching on the direction codes of the detection areas in the left and right images after the tower structure deforms, calculating the three-dimensional spatial coordinates of the detection areas after deformation.

[0084] The disparity principle and mathematical model in S3 specifically include the following steps: Binocular stereo vision, based on the disparity principle, obtains the coordinates of three-dimensional points through triangulation. The principle of calculating the three-dimensional coordinates of spatial points in binocular stereo vision is as follows: Figure 6 As shown.

[0085] The core geometric relationship of binocular stereo vision is as follows: the feature points of the measured object and the optical centers of the two cameras form a spatial triangle, and the projection points of these feature points on the left and right image planes provide angle measurement information. After obtaining the intrinsic parameter matrix, distortion coefficients, and extrinsic parameter matrix (including the rotation matrix R and translation vector T) of the binocular system through camera calibration, the three-dimensional spatial coordinates of the target point can be calculated using the coordinates of the matched feature points within the common viewing area.

[0086] In the binocular stereo vision of this invention, the origin of the left camera coordinate system O-xyz is used as the origin of the world coordinate system, and the camera does not rotate. Let the left camera image coordinate system be O. l -X l Y l The right camera image coordinate system is O r -X r Y rThen the camera projection transformation model is:

[0087]

[0088] Among them, S l S is the scale factor for the left camera. r Let S be the scale factor of the right camera, and S l / z l =S r / z r =1,f l f is the effective focal length of the left camera. r The effective focal length of the right camera. The O-xyz coordinate system and O r -X r Y r The relative positions of coordinate systems can be determined by the spatial transformation matrix M. lr express:

[0089]

[0090] Let the O-xyz coordinate system be intersected with O r -X r Y r The rotation and translation matrices of the coordinate system are as follows:

[0091]

[0092] From equations (1) to (5), it can be seen that the correspondence between spatial points in the O-xyz coordinate system and points in the left and right camera images is as follows:

[0093]

[0094] Therefore, the three-dimensional coordinates of a point in space can be represented as:

[0095]

[0096] Therefore, the focal lengths f of the left and right cameras are known. l f r The three-dimensional coordinates of the measured point can be calculated by combining the left and right image coordinates of the spatial point with the rotation and translation matrices of the two cameras.

[0097] Furthermore, the process of matching the edge direction codes of the left and right camera images using template matching and the principle of parallax includes: processing the edge direction codes of the left and right camera images using a digital image correlation matching algorithm to search for global extreme points of the correlation coefficient; processing the extreme point pairs matched by the global extreme points using a binocular stereo vision geometric model to calculate the parallax information; and processing the matching points using a three-dimensional coordinate reconstruction algorithm based on the camera calibration parameters and the parallax information to obtain the spatial three-dimensional coordinates of the matching points.

[0098] S4. By calculating the distance between the three-dimensional coordinates of the same detection location before and after structural deformation, the deformation of the hydraulic tower structure can be calculated.

[0099] Furthermore, the process of obtaining the deformation of the hydraulic lifting tower structure based on the spatial three-dimensional coordinate changes of the same detection point before and after structural deformation includes: processing the three-dimensional coordinate dataset of the detection points before deformation using a spatial point cloud registration method to establish a reference coordinate system; processing the three-dimensional coordinate dataset of the detection points after deformation using the same spatial point cloud registration method to obtain the post-deformation coordinate system; processing the Euclidean distance calculation method based on the reference coordinate system and the post-deformation coordinate system to obtain the displacement of each detection point; and processing the distribution of the displacement using a statistical analysis algorithm to obtain the overall deformation characteristic parameters of the tower structure.

[0100] This invention proposes an optimized subpixel edge detection algorithm to extract edges from the detection area of ​​the image of the object being tested, thereby improving edge localization accuracy. Based on the principle of triangular parallax, a binocular camera system is used to calculate the three-dimensional coordinates of edge feature points in the detection area. A subpixel edge direction code calculation method is proposed, which quantizes the gradient direction of edge points into discrete codes and performs template matching based on the direction codes, thereby improving matching efficiency and accuracy. The structural deformation displacement is calculated based on the change in the three-dimensional coordinates of the corresponding edge feature points before and after structural deformation.

[0101] The tower deformation measurement method based on edge feature direction code matching proposed in this invention mainly includes camera calibration, subpixel edge detection and edge direction extraction of the image at the measured location, coordinate extraction, and tower deformation extraction. The specific steps are as follows: First, the measured location is selected in the left camera image. To improve detection efficiency, the computational load is reduced by selecting a detection area, and the corresponding area is located in the right camera image using a template matching method. Then, subpixel edge detection is performed on the image at the measured location, and edge features (including edge pixel position and edge direction) are extracted. Finally, the edge direction code of the measured location is generated using the method proposed in this invention, the three-dimensional coordinates before and after deformation are calculated, and the tower deformation is calculated based on the distance of coordinate change. The main process of the proposed method is as follows: Figure 1 As shown. It's worth noting that before performing subpixel edge detection, to reduce the impact of image noise on the detection results, conventional morphological processing (mainly opening and closing operations) is typically performed on the image beforehand. Meanwhile, in Figure 1 In this process, a QR code is used to represent the general texture of the surface of the structure at the measured location, which is unrelated to the directional code algorithm proposed in this invention.

[0102] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for measuring the deformation of a hydraulic lifting tower structure based on edge feature direction code matching, characterized in that, Includes the following steps: The subpixel edge detection method is used to process the image of the detection area to obtain the subpixel edge line position and subpixel edge direction; Based on the sub-pixel edge direction, the edge direction angle is processed by the direction code encoding method to obtain a discrete edge direction code, and the edge direction codes of the left and right camera images are calculated based on the discrete edge direction code. Based on a binocular stereo vision system, the template matching method and the parallax principle are used to match the edge direction codes of the left and right camera images, and the spatial three-dimensional coordinates of the point with the maximum matching coefficient in the detection area are calculated. The deformation of the hydraulic lifting tower structure is obtained based on the change of the three-dimensional spatial coordinates of the same detection point before and after structural deformation. The process of calculating the edge direction codes of the left and right camera images based on the discretized edge direction codes includes: processing the sub-pixel edge directions using a direction code encoding method based on the detection area image acquired by the left camera to obtain the edge direction codes of the left camera image; processing the sub-pixel edge directions using the same direction code encoding method based on the detection area image acquired by the right camera to obtain the edge direction codes of the right camera image; and uniformly encoding the edge direction codes of the left and right camera images using a direction code discretization method based on a preset threshold angle to obtain the edge direction codes of the left and right camera images after the edge lines have been re-encoded. The process of matching the edge direction codes of the left and right camera images using template matching and the principle of parallax includes: processing the edge direction codes of the left and right camera images using a digital image correlation matching algorithm to search for global extreme points of the correlation coefficient; processing the extreme point pairs matched by the global extreme points using a binocular stereo vision geometric model to calculate parallax information; and processing the matching points using a three-dimensional coordinate reconstruction algorithm based on the camera calibration parameters and the parallax information to obtain the spatial three-dimensional coordinates of the matching points.

2. The method for measuring the deformation of a hydraulic lifting tower structure based on edge feature direction code matching according to claim 1, characterized in that, The process of processing the image of the detection region using a subpixel edge detection method to obtain the subpixel edge position and subpixel edge direction includes: Based on the light intensity distribution characteristics of pixels within the detection area, a weighted average method is used to process edge pixels to obtain light intensity feature values; Based on the light intensity feature value, the edge straightness coefficient is obtained by processing the ROI matrix; Based on the edge straightness coefficient, the edge line position is processed using a straight line equation to obtain the sub-pixel edge line position; Based on the sub-pixel edge line position, the edge direction is processed using a normal vector calculation method to obtain the sub-pixel edge direction.

3. The method for measuring the deformation of a hydraulic lifting tower structure based on edge feature direction code matching according to claim 2, characterized in that, The light intensity characteristic value includes a first light intensity value and a second light intensity value, and the calculation expressions for the first light intensity value and the second light intensity value are as follows: ； ； In the formula, C is the first light intensity value, O is the second light intensity value, and d1, d2, and d3 are the first center distance, the second center distance, and the third center distance, respectively. Let (i+1, j-1) be the light intensity. Let (i, j-2) be the light intensity. The light intensity of pixel (i+1, j-2) The light intensity of pixel (i-1, j+2) Let (i, j+2) be the light intensity. Let (i-1, j+1) be the light intensity of pixel (i-1, j+1).

4. The method for measuring the deformation of a hydraulic lifting tower structure based on edge feature direction code matching according to claim 1, characterized in that, The expression for processing edge direction angles using the direction code encoding method is as follows: ； In the formula, C i,j For direction code, ∆θ is the equal-width threshold, θ i, j This is the direction angle.

5. The method for measuring the deformation of a hydraulic lifting tower structure based on edge feature direction code matching according to claim 1, characterized in that, The process of obtaining the deformation of the hydraulic lifting tower structure based on the change of the spatial three-dimensional coordinates of the same detection point before and after structural deformation includes: Based on the three-dimensional coordinate dataset of the detection points before deformation, a spatial point cloud registration method is used to process the data and establish a reference coordinate system. Based on the 3D coordinate dataset of the detected points after deformation, the same spatial point cloud registration method is used to process the data to obtain the deformed coordinate system. Based on the reference coordinate system and the deformed coordinate system, the displacement of each detection point is obtained by processing using the Euclidean distance calculation method. Based on the distribution of the displacement, a statistical analysis algorithm is used to process the data and obtain the overall deformation characteristic parameters of the tower structure.

Images