Method for identifying multi-target cable force based on digital image technology

Abstract

The invention provides a method for identifying multi-target cable force based on a digital image technology, and the method comprises the steps: firstly selecting a target cable, and collecting the image data of a bridge cable; detecting and identifying positions and contours of a plurality of target cables, and determining coordinates of the target cables; performing central difference method second derivative processing on the original displacement data, and calculating to obtain acceleration data of the target cable; dividing the acceleration data into a plurality of sections, performing window function processing and fast Fourier transform on each section of signal, and averaging fast Fourier transform results of all sections to obtain an average power spectral density; the modal frequency is identified by decomposing the power spectral density matrix through the eigenvalue; extracting a time domain signal related to the fundamental frequency; detecting a peak value in the power spectrum, and picking up a fundamental frequency of the cable structure; and calculating the cable force of the target cable. According to the method, efficient and accurate identification of multi-target cable force can be realized, limitation caused by traditional sensor installation is avoided, and synchronous monitoring of a plurality of target cables can be realized.

Description

Technical Field

[0001] The invention belongs to the technical field of bridge engineering, and in particular relates to a method for identifying multi-target cable forces based on digital image technology. Background Art

[0002] As the key load-bearing component of a bridge, the accurate measurement and monitoring of its cable force is of great significance to the safety and durability of the bridge. Once the cable force is abnormal, it may lead to structural instability, imbalanced force distribution, and even serious consequences such as bridge collapse. Therefore, the monitoring and evaluation of cable force occupies a core position in bridge safety monitoring. Although traditional measurement methods can obtain relatively accurate cable force data under certain conditions, their common problem is that they are highly dependent on measurement equipment and the measurement operation is complicated. Summary of the invention

[0003] In view of this, the present invention aims to overcome the defects in the prior art and proposes a method for identifying multi-target cable forces based on digital image technology.

[0004] To achieve the above object, the technical solution created by the present invention is implemented as follows:

[0005] A method for identifying multi-target cable forces based on digital image technology comprises the following steps:

[0006] S1. Select a target cable, use a camera to shoot a video of multiple target cables of the bridge, and collect image data of the bridge cables;

[0007] S2, detecting and identifying the positions and contours of multiple target cables in the acquired image data, determining the target cable coordinates; and obtaining the original displacement data of the target cables;

[0008] S3, performing the second-order derivative processing of the central difference method on the original displacement data, and calculating the acceleration data of the target cable;

[0009] S4, dividing the acceleration data into several segments, performing window function processing and fast Fourier transform on each segment signal, averaging the fast Fourier transform results of all segments to obtain an average power spectral density;

[0010] S5. Identify the modal frequencies by eigenvalue decomposition of the power spectral density matrix;

[0011] S6, filtering using a filter to extract a time domain signal related to the base frequency;

[0012] S7, detecting the peak in the power spectrum and picking up the fundamental frequency f of the cable structure;

[0013] S8. According to the identified fundamental frequency f, the cable force F of the target cable is calculated using the following formula: Wherein, F is the cable force, L is the length of the inclined cable, m is the mass per unit length of the inclined cable, f is the fundamental frequency of the cable, EI is the bending stiffness of the inclined cable, and π is pi.

[0014] Furthermore, the YOLOv8 model is used to detect and identify the location and contour of the target cable in the collected image data.

[0015] Furthermore, the power spectral density is calculated using the Welch method.

[0016] Furthermore, in step S1, the camera acquisition frequency is higher than 32 Hz and the resolution is higher than 1080P.

[0017] Furthermore, in step S2, YOLOv8 uses an anchor-free detection strategy and uses a loss function to weight the bounding box loss, classification loss, and confidence loss.

[0018] Further, in step S2, the target index coordinate X(t x , t y ), and obtain the displacement vibration data as X.

[0019] Further, in step S3, the formula for calculating the target cable acceleration data is as follows: Where Δt is the time interval between adjacent data points.

[0020] Further, in step S4, the fast Fourier transform calculation formula is: Where N is the total number of data points, A(n) is the nth acceleration data point, k is the frequency point, and j is the imaginary unit.

[0021] Furthermore, in step S5, during eigenvalue decomposition, the enhanced frequency domain decomposition method identifies the modal frequency by decomposing the power spectral density (PSD) matrix through eigenvalue decomposition. The formula of the power spectral density is: yy (f) = E[Y(f)Y * (f)] where: S yy (f) is the power spectral density at frequency f; Y(f) is the frequency component Y after Fourier transformation * (f) is the conjugate complex number of Y(f); E is the expectation operation, which represents the average of multiple response signals.

[0022] Further, in step S5, in EFDD, eigenvalue decomposition is performed: S yy (f)=V(f)Λ(f)V * (f)] where: V(f) is the modal matrix (eigenvector matrix); Λ(f) is a diagonal matrix, in which the frequency corresponding to the maximum eigenvalue is the modal frequency.

[0023] Further, in step S6, the transfer function of the bandpass filter is: Where: f L =max(0, f0-Δf) is the lower limit frequency of the filter, which must be a positive value; f H =f0+Δf is the upper limit frequency of the filter; f0 is the target frequency, and Δf is the frequency bandwidth.

[0024] Further, in step S7, a Find peaks function is used to detect peaks in the power spectrum.

[0025] Compared with the prior art, the present invention has the following advantages:

[0026] The method of the present invention extracts displacement data of multiple target cables in the bridge in real time, obtains acceleration data using the central difference method, picks up the fundamental frequency of the cable using the enhanced frequency domain decomposition method, and combines the cable force calculation model to achieve efficient and accurate identification of multiple target cable forces. This method not only avoids the limitations brought by traditional sensor installation, but also can achieve synchronous monitoring of multiple target cables, and has a high application prospect. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The drawings constituting part of the present invention are used to provide a further understanding of the present invention. The exemplary embodiments and descriptions of the present invention are used to explain the present invention and do not constitute an improper limitation on the present invention. In the drawings:

[0028] Figure 1 A schematic diagram of the process flow of the invention; DETAILED DESCRIPTION

[0029] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0030] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In addition, the terms "first", "second", etc. are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, features defined as "first", "second", etc. may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "multiple" means two or more.

[0031] In the description of the invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installation", "connection" and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the invention can be understood according to specific circumstances.

[0032] The present invention will be described in detail below with reference to the accompanying drawings and in combination with embodiments.

[0033] A method for identifying multi-target cable forces based on digital image technology mainly includes the steps of digital image recognition, data preprocessing, modal identification fundamental frequency, and target cable force calculation. Figure 1 As shown, the specific steps include:

[0034] S1. Select a target cable, use a camera to shoot a video of multiple target cables of the bridge, and collect image data of the bridge cables;

[0035] S2. Detect and identify the positions and contours of multiple target cables in the acquired image data, and obtain the target cable coordinates X(t x , t y ), that is, the displacement vibration data is obtained as X;

[0036] When the displacement data is obtained, the YOLOv8 model is used to detect and identify the positions and contours of multiple target lines in the collected image data. The YOLOv8 model identifies the target bounding box, and the prediction formula for the center coordinates (x, y), width w, and height h of each bounding box is:

[0037] t x =σ(b x )+c x , t y =σ(b y )+c y ; where t x and t y is the center position of the marking point, b x and b y is the offset, c x and c y is the coordinate of the current grid, and σ is the Sigmoid function, which is used to limit the value to the range [0, 1].

[0038] t w , t h is the bounding box width and height of the marker point, p w , p his the preset reference size, b w , b h is the predicted scaling factor

[0039] S3, performing the second-order derivative processing of the central difference method on the original displacement data, and calculating the acceleration data of the target cable;

[0040] S4, calculating the power spectrum density, dividing the acceleration data into several segments, each segment is N in length, performing window function processing and fast Fourier transform on each segment signal, averaging the fast Fourier transform results of all segments to obtain the average power spectrum density;

[0041] S5. Identify the modal frequencies by eigenvalue decomposition of the power spectral density matrix;

[0042] During eigenvalue decomposition, the enhanced frequency domain decomposition method identifies the modal frequencies by decomposing the power spectral density (PSD) matrix. The formula for the power spectral density is: yy (f) = E[Y(f) Y * (f)] where: S yy (f) is the power spectral density at frequency f; Y(f) is the frequency component Y after Fourier transformation * (f) is the conjugate complex number of Y(f). E is the expectation operation, which represents the average of multiple response signals.

[0043] In EFDD, eigenvalue decomposition is performed: S yy (f)=V(f) Λ(f) V * (f)] Where: V(f) is the modal matrix (eigenvector matrix). Λ(f) is a diagonal matrix containing eigenvalues, corresponding to the modal frequencies. The frequency corresponding to the maximum eigenvalue is the modal frequency.

[0044] S6. Extract the time domain signal related to the baseband: In order to extract the time domain signal related to the baseband, a bandpass filter is used. The transfer function of the bandpass filter can be expressed as: Where: f L =max(0, f0-Δf) is the lower limit frequency of the filter, which must be a positive value; f H =f0+Δf is the upper frequency limit of the filter; f0 is the target frequency, Δf is the frequency bandwidth;

[0045] S7, detecting the peak in the power spectrum and picking up the fundamental frequency f of the cable structure;

[0046] S8. According to the identified fundamental frequency f, the cable force F of the target cable is calculated using the following formula: Wherein, F is the cable force, L is the length of the inclined cable, m is the mass per unit length of the inclined cable, f is the fundamental frequency of the cable, EI is the bending stiffness of the inclined cable, and π is pi.

[0047] Through the above steps, all the collected target cables can be processed and the cable force F of each cable of the bridge can be calculated.

[0048] In step S1, the camera acquisition frequency is higher than 32 Hz and the resolution is higher than 1080P.

[0049] In step S2, YOLOv8 uses an anchor-free detection strategy, and uses the loss function to weight the bounding box loss, classification loss, and confidence loss, and improves the accuracy of the detected multi-scale structure through the feature pyramid network (FPN) and path aggregation network (PANet).

[0050] The YOLOv8 model is used to detect and identify the location and contour of the target rope in the acquired image data.

[0051] In step S3, the formula for calculating the target cable acceleration data is as follows: Where Δt is the time interval between adjacent data points.

[0052] In step S4, the power spectral density is calculated using the Welch method. The fast Fourier transform calculation formula is: Where N is the total number of data points, A(n) is the nth acceleration data point, k is the frequency point, and j is the imaginary unit.

[0053] In step S7, the Find peaks function is used to detect peaks in the power spectrum.

[0054] The method of the present invention avoids the limitations of traditional sensor layout through non-contact measurement and is suitable for multi-target, complex environment and high-risk areas. In addition, the image recognition YOLOv8 model and enhanced frequency domain decomposition (EFDD) are used to achieve high-precision and high-efficiency data processing, with a high degree of automation and reduced human errors.

[0055] Moreover, this method is easy to operate, has strong anti-interference ability, can work stably in a noisy environment, is widely applicable to different types of bridges, has good operability and sustainability, and provides a new solution for the cable force calculation of various types of bridges.

[0056] 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 spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for identifying multi-target cable forces based on digital image technology, characterized in that: The steps include: S1. Select a target cable, use a camera to shoot a video of multiple target cables of the bridge, and collect image data of the bridge cables; S2, detecting and identifying the positions and contours of multiple target cables in the acquired image data, determining the target cable coordinates; and obtaining the original displacement data of the target cables; S3, performing the second-order derivative processing of the central difference method on the original displacement data, and calculating the acceleration data of the target cable; S4, dividing the acceleration data into several segments, performing window function processing and fast Fourier transform on each segment signal, averaging the fast Fourier transform results of all segments to obtain an average power spectral density; S5. Identify the modal frequencies by eigenvalue decomposition of the power spectral density matrix; S6, filtering using a filter to extract a time domain signal related to the base frequency; S7, detecting the peak in the power spectrum and picking up the fundamental frequency f of the cable structure; S8. According to the identified fundamental frequency f, the cable force F of the target cable is calculated using the following formula: Wherein, F is the cable force, L is the length of the inclined cable, m is the mass per unit length of the inclined cable, f is the fundamental frequency of the cable, EI is the bending stiffness of the inclined cable, and π is pi.

2. The method for identifying multi-target cable forces based on digital image technology according to claim 1, characterized in that: The YOLOv8 model is used to detect and identify the location and contour of the target rope in the acquired image data.

3. The method for identifying multi-target cable forces based on digital image technology according to claim 1, characterized in that: The power spectral density was calculated using the Welch method.

4. The method for identifying multi-target cable forces based on digital image technology according to claim 1, characterized in that: In step S1, the camera acquisition frequency is higher than 32 Hz and the resolution is higher than 1080P.

5. The method for identifying multi-target cable forces based on digital image technology according to claim 1, characterized in that: In step S2, YOLOv8 uses an anchor-free detection strategy and uses a loss function to weight the bounding box loss, classification loss, and confidence loss.

6. The method for identifying multi-target cable forces based on digital image technology according to claim 1, characterized in that: In step S2, the target index coordinate X(t x , t y ), and obtain the displacement vibration data as X.

7. The method for identifying multi-target cable forces based on digital image technology according to claim 1, characterized in that: In step S3, the formula for calculating the target cable acceleration data is as follows: Where Δt is the time interval between adjacent data points.

8. The method for identifying multi-target cable forces based on digital image technology according to claim 1, characterized in that: In step S4, the fast Fourier transform calculation formula is: Where N is the total number of data points, A(n) is the nth acceleration data point, k is the frequency point, and j is the imaginary unit.

9. The method for identifying multi-target cable forces based on digital image technology according to claim 1, characterized in that: In step S5, during eigenvalue decomposition, the enhanced frequency domain decomposition method identifies the modal frequency by decomposing the power spectral density (PSD) matrix through eigenvalue decomposition. The formula of the power spectral density is: yy (f) = E[Y(f)Y * (f)] where: S yy (f) is the power spectral density at frequency f; Y(f) is the frequency component Y after Fourier transformation * (f) is the conjugate complex number of Y(f); E is the expectation operation, which represents the average of multiple response signals.

10. The method for identifying multi-target cable forces based on digital image technology according to claim 1, characterized in that: In step S5, in EFDD, eigenvalue decomposition is performed: S yy (f)=V(f)Λ(f)V * (f)] where: V(f) is the modal matrix (eigenvector matrix); Λ(f) is a diagonal matrix, in which the frequency corresponding to the maximum eigenvalue is the modal frequency.

11. The method for identifying multi-target cable forces based on digital image technology according to claim 1, characterized in that: In step S6, the transfer function of the bandpass filter is: Where: f L =max(0, f0-Δf) is the lower limit frequency of the filter, which must be a positive value; f H =f0+Δf is the upper limit frequency of the filter; f0 is the target frequency, and Δf is the frequency bandwidth.

12. The method for identifying multi-target cable forces based on digital image technology according to claim 1, characterized in that: In step S7, the Find peaks function is used to detect peaks in the power spectrum.

Images