A device and method for replacing a power transmission line bolt without power interruption

Abstract

The application discloses a device and method for replacing power transmission line bolts without power interruption, wherein the device for replacing power transmission line bolts without power interruption comprises an image acquisition module, a bolt damage detection module, a bolt loosening detection module, a bolt fastening assembly and a general control module, the image acquisition module, the bolt damage detection module, the bolt loosening detection module and the bolt fastening assembly are connected with the general control module; the image acquisition module is used for acquiring power transmission line images; the bolt damage detection module is used for judging the positions of damaged bolts according to the power transmission line images; and the bolt loosening detection module is used for judging the positions of loosened bolts according to the power transmission line images. The application can realize damage detection and loosening detection of power transmission line bolts, timely replacement and fastening of the power transmission line bolts, and does not need power interruption during replacement or fastening.

Description

Technical Field

[0001] This invention relates to the field of power line maintenance tools, specifically to a device and method for repairing transmission line bolts without power interruption. Background Technology

[0002] The safe operation of power transmission lines is of great importance. Bolts play a role in fixing the connection between lines in power transmission lines. If they become loose or fall off, it may cause power transmission failure and lead to large-scale power outages, causing great inconvenience to electricity users. Therefore, they need to be inspected frequently. However, the inspection methods in the existing technology often have the problem of low detection accuracy. When repairing or tightening bolts, the existing repair devices are also inconvenient to use. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide an apparatus and method for repairing power transmission line bolts without power interruption, so as to detect the loosening and damage of power transmission line bolts, and repair and tighten them in a timely manner to prevent power outages caused by emergency maintenance.

[0004] To solve the above technical problems, the present invention provides a device for repairing transmission line bolts without power interruption, comprising: an image acquisition module, a bolt damage detection module, a bolt loosening detection module, a bolt fastening assembly, and a main control module, wherein the image acquisition module, the bolt damage detection module, the bolt loosening detection module, and the bolt fastening assembly are connected to the main control module;

[0005] The image acquisition module is used to acquire images of power transmission lines;

[0006] The bolt damage detection module is used to determine the location of damaged bolts based on images of power transmission lines;

[0007] The bolt loosening detection module is used to determine the location of loose bolts based on images of power transmission lines.

[0008] Preferably, the bolt fastening assembly includes an insulating rod, a touch screen, a controller, a camera, and a fastening module. The fastening module is fixedly mounted on the top of the insulating rod, the touch screen is fixedly mounted on the lower side of the insulating rod, and the camera is fixedly mounted on the upper side of the insulating rod. The touch screen, camera, and fastening module are electrically connected to the controller, and the controller is communicatively connected to the main control module via a wireless communication module.

[0009] Preferably, the fastening module includes a rotary motor, a fastening sleeve, and fasteners. The fasteners are disposed inside the fastening sleeve, and the rotary motor is disposed outside the fastening sleeve. The output end of the rotary motor extends into the interior of the fastening sleeve and is fixedly connected to the fasteners. The fasteners drive the bolts to rotate and tighten. The rotary motor is electrically connected to the controller. A control box is also disposed on the side of the insulating rod, and the controller and wireless communication module are disposed inside the control box.

[0010] The present invention also provides a method for repairing transmission line bolts without power interruption, wherein the method is applied to the aforementioned device for repairing transmission line bolts without power interruption, and the method includes:

[0011] Step 1: The image acquisition module acquires images of the power transmission lines in the area to be detected and sends them to the main control module. The main control module then sends the images to the bolt damage detection module and the bolt loosening detection module.

[0012] Step 2: Analyze the transmission line image based on the bolt damage detection module to obtain the location of the damaged bolts;

[0013] Step 3: Analyze the transmission line image based on the bolt loosening detection module to obtain the location of the loosened bolts;

[0014] Step 4: Replace the damaged bolts with the bolt fastening assembly and tighten the loose bolts.

[0015] Preferably, step 2 specifically includes:

[0016] Step 201: Build a damage detection model based on regional attention mechanism and multi-scale feature fusion;

[0017] Step 202: Train the damage detection model based on the preset damaged bolt dataset to obtain the trained damage detection model.

[0018] Step 203: Input the image of the transmission line into the trained damage detection model to obtain the location of the damaged bolt. Send the location of the damaged bolt to the main control module, which then sends it to the controller, which displays it on the touch screen.

[0019] Preferably, step 3 specifically includes:

[0020] Step 301: Obtain an image of the transmission line when it was first established, use it as a reference image, and perform coarse registration on the transmission line image based on the reference image;

[0021] Step 302: Perform bolt center matching on the transmission line image after coarse registration based on the reference image;

[0022] Step 303: Based on the reference image and the bolt outline of the transmission line, determine whether the bolts are loose.

[0023] Preferably, in step 301, coarse registration of the transmission line image based on the reference image is performed as follows:

[0024] The SURF feature detection algorithm is used to detect and match feature points in the reference image and the transmission line image. The RANSAC algorithm is used to remove incorrect matches, resulting in a filtered matching pair.

[0025] Preferably, in step 302, bolt center matching is performed on the transmission line image after coarse registration based on the reference image, specifically as follows:

[0026] An object detector is used to detect the matched reference image and transmission line image to obtain all the positions of the bolts in the reference image and transmission line image. Each bolt is in a separate rectangular frame. The rectangular frame is processed by an edge detection algorithm to obtain a binary image after edge detection. Contours are extracted and filtered using the hierarchical structure and length information of the contours to obtain the boundaries of the bolts. The set center of the bolt head surface is taken as the equivalent center, and the center coordinates of the bolt are calculated. Center point matching is performed based on the center coordinates of the reference image and transmission line image. After matching, the reference image and transmission line image are weighted and binarized to obtain a binary image showing the difference positions.

[0027] Preferably, in step 303, determining whether the bolts are loose based on the reference image and the bolt outline of the transmission line specifically involves:

[0028] The number of pixels M inside each bolt of the transmission line is counted using a summation function and is equivalent to the area of ​​the bolt head surface. The number N of pixels N inside each bolt whose gray value is greater than a first preset threshold is counted and is equivalent to the area of ​​the difference region in the binary image of the difference location. A second preset threshold is set. If the ratio of N to M is greater than the second preset threshold, it is determined that the bolt is loose. The location of the loose bolt is sent to the main control module, which then sends it to the controller, which displays it on the touch screen.

[0029] Preferably, step 4 specifically includes:

[0030] View the location of the damaged bolt by touching the screen, move the fastening module to the corresponding bolt by using the insulating rod, the camera captures the corresponding image and displays it on the touch screen, align the fastener with the bolt, and remove, replace or reinstall the bolt.

[0031] The location of the loose bolt can be viewed on the touch screen. The fastening module is moved to the corresponding bolt using the insulating rod. The camera captures the corresponding image and displays it on the touch screen. The fastener is then aligned with the bolt and tightened.

[0032] The present invention has the following beneficial effects: The present invention acquires images of the power transmission line in the area to be detected through an image acquisition module and sends them to a main control module. The main control module then sends these images to a bolt damage detection module and a bolt loosening detection module. Based on the bolt damage detection module's analysis of the power transmission line images, the locations of damaged bolts are determined. Based on the bolt loosening detection module's analysis of the power transmission line images, the locations of loose bolts are determined. Damaged bolts are replaced using bolt fastening components, and loose bolts are tightened. This enables the detection of damage and loosening of power transmission line bolts, allowing for timely repair and tightening of bolts without power interruption. Attached Figure Description

[0033] 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.

[0034] Figure 1 This is a three-dimensional structural diagram of the bolt fastening assembly of a device for repairing power transmission line bolts without power interruption, according to Embodiment 1 of the present invention.

[0035] Figure 2 This is a top view of the bolt fastening assembly in an embodiment of the present invention.

[0036] Figure 3 This is a flowchart illustrating a method for repairing power transmission line bolts without power interruption, according to Embodiment 2 of the present invention. Detailed Implementation

[0037] The following descriptions of various embodiments are based on the accompanying drawings, illustrating specific embodiments in which the present invention can be implemented. In the description of the present invention, it should be understood that the terms "longitudinal," "length," "circumferential," "front," "rear," "left," "right," "top," "bottom," "inner," and "outer," etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings and are only for the convenience of describing the present invention and simplifying the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention.

[0038] Embodiment 1 of the present invention provides a device for repairing bolts on power transmission lines without power interruption, comprising: an image acquisition module, a bolt damage detection module, a bolt loosening detection module, a bolt fastening assembly, and a main control module, wherein the image acquisition module, the bolt damage detection module, the bolt loosening detection module, and the bolt fastening assembly are connected to the main control module;

[0039] The image acquisition module is used to acquire images of power transmission lines;

[0040] The bolt damage detection module is used to determine the location of damaged bolts based on images of power transmission lines;

[0041] The bolt loosening detection module is used to determine the location of loose bolts based on images of power transmission lines.

[0042] Please refer to the following at the same time Figure 1 and Figure 2 As shown, the bolt fastening assembly includes an insulating rod 1, a touch screen display 6, a controller, a camera 5, and a fastening module. The fastening module is fixedly installed on the top of the insulating rod 1, the touch screen display 6 is fixedly installed on the lower side of the insulating rod 1, and the camera 5 is fixedly installed on the upper side of the insulating rod 1. The touch screen display 6, the camera 5, and the fastening module are electrically connected to the controller, and the controller is communicatively connected to the main control module through a wireless communication module.

[0043] The fastening module includes a rotary motor 3, a fastening sleeve 2, and fasteners. The fastening sleeve 2 is located on the top of the insulating rod 1. The fasteners are located inside the fastening sleeve 2. The rotary motor 3 is located outside the fastening sleeve 2. The output end of the rotary motor 3 extends into the interior of the fastening sleeve 2 and is fixedly connected to the fasteners. It is used to drive the bolts to rotate and tighten through the fasteners. The rotary motor 3 is electrically connected to the controller.

[0044] The fasteners can be conventional fasteners using existing technology, as long as they can enable the removal and installation of bolts.

[0045] The side of the insulating rod 1 is also provided with a control box 4, wherein the controller and wireless communication module are located inside the control box 4.

[0046] The camera 5 of this invention is a 1000TVL CMOS camera, and the touch screen 6 is a 4.3-inch screen.

[0047] like Figure 3 As shown, Embodiment 2 of the present invention also provides a method for repairing transmission line bolts without power interruption, applied to the device for repairing transmission line bolts without power interruption described in Embodiment 1 of the present invention. The method for repairing transmission line bolts without power interruption includes:

[0048] Step 1: The image acquisition module acquires images of the power transmission lines in the area to be detected and sends them to the main control module. The main control module then sends the images to the bolt damage detection module and the bolt loosening detection module.

[0049] Step 2: Analyze the transmission line image based on the bolt damage detection module to obtain the location of the damaged bolts;

[0050] Step 3: Analyze the transmission line image based on the bolt loosening detection module to obtain the location of the loosened bolts;

[0051] Step 4: Replace the damaged bolts with the bolt fastening assembly and tighten the loose bolts.

[0052] In step 2, the transmission line image is analyzed based on the bolt damage detection module to obtain the location of the damaged bolts. This includes the following steps:

[0053] Step 201: Build a damage detection model based on regional attention mechanism and multi-scale feature fusion;

[0054] To elaborate further: The region attention module can be viewed as a plug-and-play network operation unit that can be inserted into any convolution-based detection network. The input of the region attention block is defined as Inputs∈R C×H×W First, two 1×1 convolutional layers W are used. theta and W thi This is transformed into a latent feature space, and then two spatial pooling kernels encode the output of the convolutional layer along the X and Y directions. The encoding method can be represented as follows:

[0055]

[0056]

[0057] Here, the subscript c represents the feature of the c-th channel. Considering the different data distribution in each batch, the data in the X and Y directions are aggregated separately. This can easily lead to a large deviation in the feature weights in the two directions. To alleviate this problem, the encoded results are concatenated and input into a 1×1 convolutional layer, and a batch normalization operation is performed to accelerate the convergence of the weights. After that, the features are again divided into two independent tensors f. h ∈R C×H×1 and f w ∈R C×1×W This method transforms the original two-dimensional features into one-dimensional features encoded in different directions, significantly reducing the computational space of the model. Then, f... h and f w Matrix multiplication is performed to obtain a region-attention-based feature map. This feature map is then input into a sigmoid activation function and encoded on each channel using a spatial pooling kernel. The calculation process is as follows:

[0058]

[0059] Meanwhile, considering that the matrix multiplication of two one-dimensional feature tensors may blur the boundary information of the object, unary units are introduced to model the salient edge information of the object. Therefore, an additional branch is created, in which a convolutional layer with a 1×1 kernel and 1 output channel and a sigmoid activation function are used, as calculated below:

[0060] f unary =Sigmoid(Conv1) c＝1 (Iuput))∈R 1×H×W

[0061] Finally, element-wise summation is performed on `funary` and `fregion`, and the result is multiplied by the feature input. The final output of the region attention module can be represented as:

[0062]

[0063] Multi-Scale Feature Fusion (MS-FPN): Based on FPN, a multi-scale feature fusion network is proposed. By extending the original structure with a bottom-up branch, shallow information is more easily propagated. Simultaneously, shallow features are utilized multiple times to improve the network's ability to focus on small objects. First, Non-Maximum Suppression (NMS) is used to group negative samples into different clusters, and the highest score among all foreground categories is used as the score for the negative sample. Then, the negative samples are sorted in descending order of score. After sorting, the top 1 from each group is placed at the beginning of the final rank, and the top 2 from each group is placed after the previously placed top 1. This process continues until all samples are sorted. Using this method, the network focuses more on learning positive samples and important negative samples in the detection part. PISA also proposes Classification-Aware Regression Loss (CARL) to jointly optimize the classification and regression branches, improving the scores of important samples while suppressing other unimportant samples. The specific expression is as follows:

[0064]

[0065] Where, p i d represents the probability that sample i is the corresponding ground truth class. i The regression offset of the regression branch output, the exponential function will p i Convert to v i Then, the values ​​are rescaled based on the average of all samples. L1 represents the smooth L1 loss commonly used in regression. After using CARL, the classification is supervised by the regression loss, which suppresses the scores of unimportant samples and enhances the attention to important samples.

[0066] Step 202: Train the damage detection model based on the preset damaged bolt dataset to obtain the trained damage detection model. SGD can be used as the model optimizer.

[0067] Step 203: Input the image of the transmission line into the trained damage detection model to obtain the location of the damaged bolt. Send the location of the damaged bolt to the main control module, which then sends it to the controller, which displays it on the touch screen.

[0068] In step 3, the transmission line image is analyzed based on the bolt loosening detection module to obtain the location of the loosened bolts. This includes the following steps:

[0069] Step 301: Obtain an image of the transmission line when it was first established, use it as a reference image, and perform coarse registration on the transmission line image based on the reference image;

[0070] Step 302: Perform bolt center matching on the transmission line image after coarse registration based on the reference image;

[0071] Step 303: Based on the reference image and the bolt outline of the transmission line, determine whether the bolts are loose.

[0072] In step 301, coarse registration is performed on the transmission line image based on the reference image, specifically as follows:

[0073] The SURF feature detection algorithm is used to detect and match feature points in the reference image and the transmission line image. The RANSAC algorithm is used to remove incorrect matches, resulting in filtered matching pairs. The SURF feature detection algorithm is explained in detail below:

[0074] Based on scale space theory, the SURF feature detection algorithm constructs Hessian matrices of different scales at any point (x, y) in the image, and calculates the discriminant of these matrices.

[0075] When the discriminant reaches a local maximum, the current point is determined as the key point. Different scale spatial structures are obtained by changing the size of the Gaussian blur. Different image pyramid layers have different Gaussian blur sizes, and the same layer will also have different image blur sizes. After obtaining the extreme value using the Hessian matrix, non-maximum suppression is performed in the 3×3×3 three-dimensional neighborhood. Only key points that are all greater than or all less than the values ​​of the 26 neighborhoods around the upper scale, the lower scale, and the current scale can be used as candidate feature points. After selecting the feature points, the principal direction of the feature points is assigned and the feature point descriptor is constructed. Both of these steps are implemented based on the Haar wavelet features in the neighborhood of the feature points. Finally, the matching degree is determined by calculating the Euclidean distance between the feature points. In addition, the SURF algorithm will exclude matches with different Hessian matrix trace signs.

[0076] The specific implementation process of the RANSAC algorithm is as follows: randomly select n pairs as samples from the matching pairs obtained by the SURF algorithm, calculate the homography matrix M from the sample coordinates, calculate the consistent set that satisfies matrix M within a reasonable error range, count the number of elements in the consistent set, and determine whether the current consistent set is optimal, i.e., the largest number of elements. If so, update the optimal consistent set and update the current error probability t. If t is greater than the minimum allowed error probability, repeat the above steps to continue iterating until the current error probability t is less than the minimum allowed error probability.

[0077] In step 302, bolt center matching is performed on the transmission line image after coarse registration based on the reference image. Specifically:

[0078] The object detector detects the matched reference image and transmission line image to obtain all the positions of the bolts in the reference image and transmission line image. Each bolt is in a separate rectangular frame. The rectangular frame is processed by the edge detection algorithm to obtain the binary image after edge detection. The contour is extracted and filtered using the hierarchical structure and length information of the contour to obtain the boundary of the bolt. The set center of the bolt head surface is equivalent to the center, and the center coordinates of the bolt are calculated. The center point is matched according to the center coordinates of the reference image and transmission line image. After the matching is completed, the reference image and transmission line image are weighted and binarized to obtain a binary image showing the difference position.

[0079] The object detector was trained using the YOLO (You Only Look Once) deep learning framework.

[0080] In step 303, based on the reference image and the bolt outline of the transmission line, it is determined whether the bolts are loose, specifically as follows:

[0081] The number of pixels M inside each bolt of the transmission line is counted using a summation function and is equivalent to the area of ​​the bolt head surface. The number N of pixels N inside each bolt whose gray value is greater than a first preset threshold is counted and is equivalent to the area of ​​the difference region in the binary image of the difference location. A second preset threshold is set. If the ratio of N to M is greater than the second preset threshold, it is determined that the bolt is loose. The location of the loose bolt is sent to the main control module, which then sends it to the controller, which displays it on the touch screen.

[0082] In step 3, damaged bolts are replaced using the bolt fastening assembly, and loose bolts are tightened. Specifically:

[0083] Operators can view the location of the damaged bolts through the touch screen, move the fastening module to the corresponding bolt using the insulating rod, the camera captures the corresponding image and displays it on the touch screen, and the operator aligns the fastener with the bolt, and removes, replaces or reinstalls the bolt.

[0084] Operators can view the location of loose bolts via a touchscreen display, move the fastening module to the corresponding bolt using an insulating rod, capture the corresponding image with a camera and display it on the touchscreen display, and then align the fastener with the bolt and tighten it.

[0085] As can be seen from the above description, compared with the prior art, the beneficial effects of the present invention are as follows: The present invention acquires images of the power transmission line in the area to be detected through an image acquisition module and sends them to the main control module. The main control module then sends them to the bolt damage detection module and the bolt loosening detection module. Based on the analysis of the power transmission line image by the bolt damage detection module, the location of the damaged bolt is obtained. Based on the analysis of the power transmission line image by the bolt loosening detection module, the location of the loose bolt is obtained. The damaged bolt is replaced and the loose bolt is tightened by the bolt fastening assembly. Thus, it is possible to detect damage and looseness of power transmission line bolts, and to promptly repair and tighten the power transmission line bolts without interrupting the power supply during repair or tightening.

[0086] The above description is merely a preferred embodiment of the present invention and should not be construed as limiting the scope of the invention. Therefore, any equivalent variations made in accordance with the claims of the present invention are still within the scope of the present invention.

Claims

1. A method for repairing transmission line bolts without power interruption, characterized in that, include: Step 1: The image acquisition module acquires images of the power transmission lines in the area to be detected and sends them to the main control module. The main control module then sends the images to the bolt damage detection module and the bolt loosening detection module. Step 2: Analyze the transmission line image based on the bolt damage detection module to obtain the location of the damaged bolts; Step 3: Analyze the transmission line image based on the bolt loosening detection module to obtain the location of the loosened bolts; Step 4: Replace the damaged bolts with the bolt fastening assembly and tighten the loose bolts. Step 2 specifically includes: Step 201: Build a damage detection model based on regional attention mechanism and multi-scale feature fusion; Step 202: Train the damage detection model based on the preset damaged bolt dataset to obtain the trained damage detection model. Step 203: Input the image of the transmission line into the trained damage detection model to obtain the location of the damaged bolt. Send the location of the damaged bolt to the main control module, which then sends it to the controller, which displays it on the touch screen.

2. The method for repairing transmission line bolts without power interruption according to claim 1, characterized in that, Step 3 specifically includes: Step 301: Obtain an image of the transmission line when it was first established, use it as a reference image, and perform coarse registration on the transmission line image based on the reference image; Step 302: Perform bolt center matching on the transmission line image after coarse registration based on the reference image; Step 303: Based on the reference image and the bolt outline of the transmission line, determine whether the bolts are loose.

3. The method for repairing transmission line bolts without power interruption according to claim 2, characterized in that, In step 301, coarse registration of the transmission line image based on the reference image is performed, specifically as follows: The SURF feature detection algorithm is used to detect and match feature points in the reference image and the transmission line image. The RANSAC algorithm is used to remove incorrect matches, resulting in a filtered matching pair.

4. The method for repairing transmission line bolts without power interruption according to claim 2, characterized in that, In step 302, bolt center matching is performed on the transmission line image after coarse registration based on the reference image. Specifically: An object detector is used to detect the matched reference image and transmission line image to obtain all the positions of the bolts in the reference image and transmission line image. Each bolt is in a separate rectangular frame. The rectangular frame is processed by an edge detection algorithm to obtain a binary image after edge detection. Contours are extracted and filtered using the hierarchical structure and length information of the contours to obtain the boundaries of the bolts. The set center of the bolt head surface is taken as the equivalent center, and the center coordinates of the bolt are calculated. Center point matching is performed based on the center coordinates of the reference image and transmission line image. After matching, the reference image and transmission line image are weighted and binarized to obtain a binary image showing the difference positions.

5. The method for repairing transmission line bolts without power interruption according to claim 2, characterized in that, In step 303, based on the reference image and the bolt outline of the transmission line, it is determined whether the bolts are loose, specifically as follows: The number of pixels M inside each bolt of the transmission line is counted using a summation function and is equivalent to the area of ​​the bolt head surface. The number N of pixels N inside each bolt whose gray value is greater than a first preset threshold is counted and is equivalent to the area of ​​the difference region in the binary image of the difference location. A second preset threshold is set. If the ratio of N to M is greater than the second preset threshold, it is determined that the bolt is loose. The location of the loose bolt is sent to the main control module, which then sends it to the controller, which displays it on the touch screen.

6. The method for repairing transmission line bolts without power interruption according to claim 1, characterized in that, Step 4 specifically includes: View the location of the damaged bolt by touching the screen, move the fastening module to the corresponding bolt by using the insulating rod, the camera captures the corresponding image and displays it on the touch screen, align the fastener with the bolt, and remove, replace or reinstall the bolt. The location of the loose bolt can be viewed on the touch screen. The fastening module is moved to the corresponding bolt using the insulating rod. The camera captures the corresponding image and displays it on the touch screen. The fastener is then aligned with the bolt and tightened.

Images