A safety early warning method and device for tower assembly construction

Abstract

The present invention discloses a method and device for early warning of safety during tower construction. The method comprises the following steps: collecting tower construction images and processing them to generate a primary model of a tower skeleton; obtaining multiple sets of fitted angle data between the tower axis and the supporting rope based on the primary model of the tower skeleton, and collecting multiple sets of measured angle data between the tower axis and the supporting rope; calculating multiple sets of angle deviation data based on the multiple sets of fitted angle data and the multiple sets of measured angle data, optimizing the primary model of the tower skeleton, and generating a tower skeleton angle recognition model; collecting real-time tower construction images and inputting them into the tower skeleton angle recognition model to obtain real-time angle data, and issuing a safety early warning based on the real-time angle data. The present invention can solve the technical problems of existing methods for obtaining angle data between the supporting rope and the tower axis, such as complex device operation, cumbersome data processing, expensive instruments, data delays / data loss, which affect the progress of tower construction.

Description

Technical Field

[0001] The present invention relates to the technical field of construction monitoring and early warning, and in particular to a method and device for early warning of safety in tower assembly construction. Background Art

[0002] With the development of urbanization, the number of power grids is gradually increasing. As the height of the construction increases, existing cranes, self-propelled cranes and other machinery are unable to reach the construction site. In order to solve the problem of being unable to construct, a large number of poles are assembled by holding and erecting poles.

[0003] By optimizing the holding pole used during the construction process, it can be adapted to the construction terrain of most rural power grid transformation and upgrading projects, effectively improving the stability of the holding pole during construction, ensuring the personal safety of workers, and effectively improving the progress of the project.

[0004] However, the existing pole holding device is not convenient for transportation during use, and its structure is relatively complicated; it is time-consuming to install and dismantle, thereby wasting manpower and material resources, and as the building height increases, the height of the internal suspended pole also increases accordingly; in some areas, the wind is strong, causing the suspended pole to shake easily, and the supporting rope is easily deformed or even broken, eventually causing the pole to fall, resulting in casualties.

[0005] To address these issues, it's necessary to monitor the supporting ropes and mast during construction. For example, the utility model patent with publication number CN215439394U discloses a mast assembly with a force measurement and monitoring system, including a force sensor, an inclination sensor, and a display. The detection values of the force and inclination sensors are displayed on the display. Each supporting rope is connected to a set of force sensors and then to the mast body. Each pull rope is connected to a set of force sensors and then to the mast body. A set of force sensors is connected in series to the lifting rope, and an inclination sensor is installed on the mast body.

[0006] The utility model patent with publication number CN215331840U discloses a monitoring system for a pole tower assembly, which belongs to the technical field of pole tower assembly monitoring systems; the technical problem to be solved is: to provide an improvement in the hardware structure of a pole tower assembly monitoring system; the technical solution adopted to solve the above technical problem is: it includes a pole, and both ends of the pole are respectively provided with a first space sensor for monitoring the inclination angle and direction of the pole, four upper pull wires are provided at the top of the pole, and a supporting rope is provided at the bottom of the pole, the supporting rope is connected to the first winch, and the four upper pull wires are connected to the second winch, and the second space sensor is provided on the upper pull wire; a control cabinet is provided on one side of the pole tower assembly, a control circuit board is provided inside the control cabinet, a central controller is integrated on the control circuit board, and a display screen and an operation panel are provided on the surface of the control cabinet.

[0007] While both of these methods can determine the mast inclination and even the forces acting on the ropes, the resulting data is independent of each other, providing no comprehensive understanding of the overall mast assembly construction status. However, using the inclination and force data to determine the angle between the mast and the supporting ropes, which is used to represent the overall construction status of the mast assembly, presents complex data processing and low angle determination accuracy. Furthermore, each supporting rope corresponds to a force sensor, resulting in high costs and complex operation.

[0008] Therefore, how to design a method that can intuitively reflect the overall construction status of the tower group and has accurate angle measurement and judgment data and simple data processing is a technical problem that needs to be solved urgently in this field. Summary of the Invention

[0009] The present invention provides a safety warning method and device for tower assembly construction. This method addresses the existing drawbacks of acquiring data on the angle between the support rope and the axis of the tower assembly during tower assembly construction, including complex device operation, cumbersome data processing, and expensive instrumentation. Furthermore, the method also suffers from low accuracy in angle measurement and determination, as well as data delays and data loss, which hinder the progress of tower assembly construction.

[0010] In a first aspect, the present invention provides a safety early warning method for tower assembly construction, comprising:

[0011] The image acquisition device is used to collect images of the tower assembly construction;

[0012] Processing the mast tower construction image to generate a primary mast skeleton model;

[0013] Acquire multiple sets of fitted angle data between the pole axis and the supporting rope based on the primary model of the pole skeleton, and collect multiple sets of measured angle data between the pole axis and the supporting rope in practice through an angle sensor device;

[0014] Calculating multiple sets of angle deviation data based on the multiple sets of fitting angle data and the multiple sets of measured angle data;

[0015] Optimizing the primary model of the pole frame according to the angle deviation data to generate a pole frame angle recognition model;

[0016] Real-time pole-mounted tower construction images are collected and input into the pole-mounted frame angle recognition model, real-time angle data is output, and safety warnings are issued based on the real-time angle data.

[0017] Furthermore, based on the primary model of the pole frame, multiple sets of fitting angle data between the pole axis and the supporting rope are obtained, including:

[0018] Based on the primary model of the pole skeleton, a rectangular coordinate system is established with the contact point between the pole and the horizontal plane as the origin, the horizontal plane as the x-axis, and the pole as the y-axis. The angle data between each supporting rope and the pole axis are obtained in the rectangular coordinate system. The obtained angle data between the multiple supporting ropes and the pole axis are integrated to obtain multiple sets of fitting angle data.

[0019] Furthermore, calculating multiple sets of angle deviation data based on the multiple sets of fitting angle data and the multiple sets of measured angle data includes:

[0020] Each fitted angle data is compared with the corresponding measured angle data one by one, and the difference is integrated as the angle deviation data to obtain multiple groups of angle deviation data.

[0021] Furthermore, the mast tower construction image is processed to generate a primary model of the mast skeleton, including:

[0022] Binarizing the tower assembly construction image to obtain a pre-processed tower assembly construction image.

[0023] Perform skeleton recognition based on the pre-processed image of the tower assembly construction to obtain overall skeleton data information;

[0024] Establishing a three-dimensional model framework of a pole tower assembly, and fitting the overall skeleton data information with the three-dimensional model framework of the pole tower assembly to obtain a fitting model;

[0025] The tower skeleton data information in the fitting model is screened and eliminated to obtain the primary model of the pole skeleton.

[0026] Furthermore, a three-dimensional model framework of a pole tower is established, and the overall skeleton data information is fitted with the three-dimensional model framework of the pole tower to obtain a fitting model, including:

[0027] Obtain the target tower model for pole-holding tower construction;

[0028] Traversing a pre-built pole tower database, searching for a target index corresponding to the target tower model from the pole tower database;

[0029] Extracting relevant data blocks from the pole tower database according to the target index, and establishing a three-dimensional model architecture of the pole tower according to the relevant data blocks;

[0030] Extracting skeleton monomer data and skeleton connection nodes from the overall skeleton data information;

[0031] Positioning the skeleton monomers of the three-dimensional model of the tower group according to the skeleton connection nodes;

[0032] The three-dimensional model architecture of the pole group tower is fitted according to the skeleton monomer positioning result and the skeleton monomer data to obtain a fitting model.

[0033] Furthermore, the tower skeleton data information in the fitting model is filtered and eliminated to obtain the primary model of the pole skeleton, including:

[0034] Obtain the target tower model for pole-holding tower construction;

[0035] Collecting multi-angle tower images of the same specifications as the target tower model from big data according to the target tower model;

[0036] Extracting a tower skeleton relative position feature set and a tower skeleton proportion feature set according to the multi-angle tower image;

[0037] Establishing a tower skeleton recognition model based on a BP neural network, and training the tower skeleton recognition model according to the multi-angle tower images, the tower skeleton relative position feature set, and the tower skeleton ratio feature set;

[0038] Inputting the fitting model into the trained tower skeleton recognition model to obtain the tower skeleton identification result of the fitting model;

[0039] The tower skeleton data information in the fitting model is eliminated according to the tower skeleton identification result to obtain the primary model of the pole skeleton.

[0040] Furthermore, the primary model of the pole frame is optimized according to the angle deviation data to generate a pole frame angle recognition model, including:

[0041] The angle deviation data is input into the primary model of the pole frame, and the three-dimensional structure data of the primary model of the pole frame is iteratively adjusted according to the angle deviation data. When a preset number of iterations is reached, the pole frame angle recognition model is generated.

[0042] Furthermore, a safety warning is performed based on the real-time angle data, including:

[0043] generating an angle change curve according to the real-time angle data;

[0044] Calculating the real-time angle change rate according to the angle change curve;

[0045] When the real-time angle changes at a constant rate and within a preset angle range, a first warning message is generated, which is used to remind construction workers to adjust the angle between the holding pole axis and the supporting rope;

[0046] When the rate of change of the real-time angle is accelerated and the range of change is within the preset angle range, a second warning message is generated, and the second warning message is used to remind the construction personnel to suspend the pole holding operation.

[0047] Furthermore, the preset angle range is 30° to 60°.

[0048] In a second aspect, the present invention further provides a safety warning device for tower assembly construction, comprising:

[0049] An acquisition module is used to acquire images of the tower assembly construction process through an image acquisition device;

[0050] A model generation module is used to process the mast tower construction image to generate a primary model of the mast skeleton;

[0051] a measurement module for obtaining multiple sets of fitted angle data between the pole axis and the supporting rope based on the primary model of the pole skeleton, and collecting multiple sets of measured angle data between the pole axis and the supporting rope in practice through an angle sensor;

[0052] A calculation module, configured to calculate multiple sets of angle deviation data based on the multiple sets of fitting angle data and the multiple sets of measured angle data;

[0053] an optimization module, configured to optimize the primary model of the pole skeleton according to the angle deviation data to generate a pole skeleton angle recognition model;

[0054] The early warning module is used to collect real-time pole-holding tower construction images and input them into the pole-holding skeleton angle recognition model, output real-time angle data, and issue safety early warnings based on the real-time angle data.

[0055] The present invention provides a method and device for early warning of safety in tower assembly construction, which has at least the following beneficial effects:

[0056] (1) The present invention collects images of the tower arm assembly construction and generates a primary model of the tower arm skeleton, then calculates the angle deviation using multiple sets of fitted angle data and multiple sets of measured angle data, optimizes the primary model of the tower arm skeleton, generates a tower arm skeleton angle recognition model, and then performs a safety early warning assessment of the tower arm assembly construction based on the real-time angle data output by the tower arm skeleton angle recognition model. This solves the problems in the prior art of using the tower arm assembly for iron tower construction, such as the complex device operation, cumbersome data processing, and expensive instruments in the method for obtaining the angle data between the supporting rope and the axis of the tower arm, as well as the low accuracy of the angle measurement and judgment data, and the data delay / data loss, which affect the progress of the tower arm assembly construction work. The present invention realizes the rational and precise control of the accuracy of non-contact object angle measurement, thereby improving the safety of construction measures.

[0057] (2) By establishing a three-dimensional model framework of the pole tower, extracting the overall skeleton data information to obtain the skeleton monomer data and performing skeleton positioning, the fitting optimization of the three-dimensional model framework of the pole tower is finally achieved, thereby generating a fitting model, achieving the technical effect of accurately positioning the skeleton of the pole tower, improving the model accuracy, and making the output angle recognition result more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 A flowchart of a safety early warning method for tower assembly construction provided by the present invention;

[0059] Figure 2 A flowchart of generating a primary model of a pole skeleton according to an embodiment of the present invention;

[0060] Figure 3 A flowchart of obtaining a fitting model according to an embodiment of the present invention;

[0061] Figure 4 A flowchart of obtaining a primary model of a pole skeleton according to an embodiment of the present invention;

[0062] Figure 5 A flowchart of a security warning according to an embodiment of the present invention;

[0063] Figure 6 A schematic diagram of a safety warning device for tower assembly construction provided by the present invention;

[0064] Figure 7 A schematic diagram of a model generation module according to an embodiment of the present invention;

[0065] Figure 8 A schematic diagram of a fitting model generation module according to an embodiment of the present invention;

[0066] Figure 9A schematic diagram of a primary model acquisition module according to an embodiment of the present invention;

[0067] Figure 10 This is a schematic diagram of an early warning module according to an embodiment of the present invention. DETAILED DESCRIPTION

[0068] In order to better understand the above technical solution, the above technical solution will be described in detail below with reference to the accompanying drawings and specific implementation methods.

[0069] See also Figure 1 As shown, the present invention provides a safety early warning method for tower assembly construction, comprising:

[0070] Step S100: Capturing tower assembly construction images through an image acquisition device;

[0071] Step S200: Processing the mast tower construction image to generate a primary mast skeleton model;

[0072] Step S300: Acquire multiple sets of fitted angle data between the pole axis and the supporting rope based on the primary model of the pole skeleton, and collect multiple sets of measured angle data between the pole axis and the supporting rope in practice through an angle sensor device;

[0073] Step S400, calculating multiple sets of angle deviation data based on multiple sets of fitting angle data and multiple sets of measured angle data;

[0074] Step S500: Optimize the primary model of the pole frame according to the angle deviation data to generate a pole frame angle recognition model;

[0075] Step S600: collect real-time pole-mounted tower construction images and input them into a pole-mounted frame angle recognition model, output real-time angle data, and issue a safety warning based on the real-time angle data.

[0076] This embodiment uses an image acquisition device to capture images of the mast tower construction process. Based on these images, a primary mast skeleton model is generated to obtain multiple sets of fitted angle data. An angle sensing module then acquires multiple sets of measured angle data. The angle deviations of the multiple sets of fitted angle data and the multiple sets of measured angle data are calculated to obtain multiple sets of angle deviation parameters. The primary mast skeleton model is then optimized to generate a mast skeleton angle recognition model. A safety early warning assessment for mast tower construction is then performed based on the real-time angle data. This method addresses the drawbacks of existing methods for acquiring angle data between the support rope and the mast axis during tower construction using mast towers, such as complex device operation, cumbersome data processing, and expensive instrumentation. Furthermore, the system suffers from technical issues such as low accuracy in angle measurement and determination data, and data delays and omissions, which can affect the progress of mast tower construction. This method achieves rational and precise control of the accuracy of non-contact object angle measurement, thereby improving the safety of construction measures.

[0077] In actual application scenarios, the image acquisition device in step S100 can be but is not limited to a camera, a video camera, a still camera, or other devices with a photo-taking function; the image acquisition device collects construction images of the pole-holding tower assembly during construction, thereby obtaining the pole-holding tower assembly construction image, which serves as an important reference for the subsequent pole-holding tower assembly construction safety early warning assessment.

[0078] Step S200 of this embodiment extracts lines from the collected images of the pole-holding tower construction, then filters out lines belonging to the pole based on the characteristics of the pole, filters out lines belonging to the pole-holding support rope based on the characteristics of the pole-holding support rope, and filters out lines belonging to the pole-holding axis based on the characteristics of the pole axis. All the obtained lines are extracted and integrated to generate a primary model of the pole skeleton, thereby ensuring safety early warning assessment of the pole-holding tower construction.

[0079] See also Figure 2 As shown, in step S200, the mast tower construction image is processed to generate a primary mast skeleton model, including:

[0080] Step S210: Binarize the pole-holding tower assembly construction image to obtain a pole-holding tower assembly construction pre-processed image;

[0081] Step S220: Perform skeleton recognition based on the pre-processed image of the tower assembly construction to obtain overall skeleton data information;

[0082] Step S230: Establish a three-dimensional model framework of the pole tower assembly, and fit the overall skeleton data information with the three-dimensional model framework of the pole tower assembly to obtain a fitting model;

[0083] Step S240: Filter and eliminate the tower skeleton data information in the fitting model to obtain a primary model of the pole skeleton.

[0084] The tower mast construction image is binarized, specifically by setting the grayscale values of pixels in the image to 0 or 255. This creates a distinct black and white visual effect, clearly highlighting all lines within the image. Skeleton recognition is then performed based on the pre-processed image. The resulting pre-processed image is then filtered and extracted to create the overall skeleton data, which combines the skeleton data of the mast assembly and the semi-finished tower.

[0085] See also Figure 3 As shown, in step S230, a three-dimensional model framework of a pole tower is established, and the overall skeleton data information is fitted with the three-dimensional model framework of the pole tower to obtain a fitting model, including:

[0086] Step S231: Obtain a target tower model for pole-holding tower assembly construction;

[0087] Step S232: traverse the pre-built pole tower database and search for a target index corresponding to the target tower model from the pole tower database;

[0088] Step S233: extract relevant data blocks from the pole tower database according to the target index, and establish a three-dimensional model architecture of the pole tower according to the relevant data blocks;

[0089] Step S234: extracting skeleton monomer data and skeleton connection nodes from the overall skeleton data information;

[0090] Step S235: Positioning the skeleton of the three-dimensional model of the tower group according to the skeleton connection nodes;

[0091] Step S236: Fit the three-dimensional model of the mast tower assembly according to the skeleton unit positioning result and the skeleton unit data to obtain a fitting model.

[0092] During the traversal of the mast tower database, the block architecture information for the next data block in the database is determined to precisely control IOPS (Input / Output Operations Per Second), improving the user experience. Furthermore, when the current data block reaches the last block, the database traversal ends, and the 3D mast tower model architecture is established. The 3D mast tower model architecture is fitted using the skeleton unit positioning results and skeleton unit data, achieving the technical effect of accurately positioning the mast tower skeleton.

[0093] Specifically, the establishment of the three-dimensional model architecture of the pole tower is achieved by integrating the positioning of fixed points and edges on the pre-processed image of the pole tower construction with the overall skeleton data information collected from the big data, and then using the pole vertices and edges, multiple supporting ropes, and the pole axis to represent the shape, and modifying the points and edges to change the shape of the shape, that is, the constructed form is a simple wireframe diagram, and the relevant data blocks include the mathematical expression related to it, which is the equation of a straight line or curve, the coordinates of the points, and the connection information of the edges and points. The connection information determines which points are the endpoints of which edge, and at which point the edge is adjacent to other edges. The model constructed with the wireframe is called the three-dimensional model architecture of the pole tower. Furthermore, the above-mentioned overall skeleton data information is fitted with the three-dimensional model architecture of the pole structure to better improve the three-dimensional model architecture of the pole tower, thereby generating a fitting model.

[0094] See also Figure 4 As shown, in step S240, the tower skeleton data information in the fitting model is filtered and eliminated to obtain a primary model of the pole skeleton, including:

[0095] Step S241: Obtain a target tower model for pole-holding tower assembly construction;

[0096] Step S242: collecting multi-angle tower images of the same specifications as the target tower model from the big data according to the target tower model;

[0097] Step S243: extracting a tower skeleton relative position feature set and a tower skeleton ratio feature set based on the multi-angle tower image;

[0098] Step S244: establishing a tower skeleton recognition model based on a BP neural network, and training the tower skeleton recognition model based on multi-angle tower images, a tower skeleton relative position feature set, and a tower skeleton ratio feature set;

[0099] Step S245: input the fitting model into the trained tower skeleton recognition model to obtain the tower skeleton identification result of the fitting model;

[0100] Step S246: Eliminate the tower skeleton data information in the fitting model according to the tower skeleton identification result to obtain a primary model of the pole skeleton.

[0101] Step S245 and step S246 are specifically as follows: comparing and screening the tower skeleton data information in the fitting model, that is, overlapping and comparing the tower skeleton data information with the overall skeleton data information. The overlapping part is the tower skeleton data, and the screened tower skeleton data is eliminated in the fitting model, and finally the primary model of the pole skeleton is obtained, further achieving the technical effect of evaluating the safety warning of the pole tower assembly construction based on the primary model of the pole skeleton.

[0102] Specifically, based on the obtained target tower model, the target tower model is traversed through big data collection to obtain multi-angle tower images of the same specifications as the target tower. That is, towers of the same specifications as the target tower model are selected from the big data for collection, and their images from different angles are integrated. Based on the multi-angle tower image set, the tower skeleton relative position feature set and the tower skeleton ratio feature set are respectively extracted. The tower skeleton relative position feature set refers to: recording the features of some specific positions of the tower skeleton in the big data and the target tower model, and then integrating these specific position feature records. The tower skeleton ratio feature set refers to: recording the features of the fixed ratios of the tower skeleton in the big data and the target tower model, and then integrating these fixed ratio feature records.

[0103] In step S244, the tower frame recognition model is a BP neural network model used in machine learning that can continuously perform self-iterative optimization. The tower frame recognition model is trained using a training dataset and a supervised dataset. Each set of training data in the training dataset includes multi-angle tower images, a set of tower frame relative position features, and a set of tower frame proportion features. The supervised dataset is fault type supervisory data that corresponds one-to-one with the training dataset. The tower frame recognition model is constructed by inputting each set of training data in the training dataset into the tower frame recognition model. The output of the tower frame recognition model is supervised and adjusted using the supervisory data corresponding to the training data set. When the output of the tower frame recognition model is consistent with the supervisory data, training for the current set of training data is completed. Once all training data in the training dataset are trained, the tower frame recognition model is trained. To ensure the accuracy of the tower frame recognition model, the tower frame recognition model can be tested using a test dataset. For example, the test accuracy can be set to 90%. When the test accuracy of the test dataset meets 90%, the tower frame recognition model is constructed. Through the high-accuracy tower skeleton recognition model, the technical effect of accurately identifying the tower skeleton can be achieved.

[0104] In step S245 and step S246, after completing the tower skeleton recognition model, the fitting model is input to identify the fitting model, and the tower skeleton identification result with the tower skeleton data information is obtained, wherein the tower skeleton identification result is obtained by feature comparison between the big data and the target tower model, wherein the fitting model includes the tower skeleton, multiple supporting ropes, and the pole axis. Since it is necessary to determine the angles between the multiple pole supporting ropes and the pole axis through the pole skeleton primary model, it is necessary to screen the above-mentioned fitting model through the tower skeleton data obtained according to the tower skeleton identification result and eliminate it from the fitting model, and further generate the pole skeleton primary model, so as to achieve the highest efficiency technical effect of the safety early warning assessment of the pole tower assembly construction in the later stage.

[0105] In step S300, multiple sets of fitting angle data between the pole axis and the supporting rope are obtained based on the primary model of the pole skeleton, including:

[0106] Based on the primary model of the pole skeleton, a rectangular coordinate system is established with the contact point between the pole and the horizontal plane as the origin, the horizontal plane as the x-axis, and the pole as the y-axis. The angle data between each supporting rope and the pole axis are obtained in the rectangular coordinate system. The angle data of multiple supporting ropes and the pole axis are integrated to obtain multiple sets of fitting angle data.

[0107] Integrating the acquired angle data between the multiple supporting ropes and the mast axis to obtain multiple sets of fitted angle data can lay a solid foundation for the safety early warning assessment of the mast tower assembly construction. In addition, in step S300, the angle sensing device collects multiple sets of measured angle data based on the calculation of the level meter to obtain the actual angles between the multiple supporting ropes and the mast axis. The level meter calculation formula can be the height difference method: known elevation + height difference = elevation to be measured, height difference = foresight degree - rear vision reading; or the equal height method: known elevation + known elevation point reading = H, H - reading of the point to be measured = elevation to be measured. The actual angle data between the multiple supporting ropes and the mast axis calculated by the level meter are summarized and integrated to further obtain multiple sets of measured angle data, which has a profound impact on the subsequent safety early warning assessment of the mast tower assembly construction.

[0108] In step S400, multiple sets of angle deviation data are calculated based on multiple sets of fitting angle data and multiple sets of measured angle data, including:

[0109] Each fitted angle data is compared with the corresponding measured angle data one by one, and the difference is integrated as the angle deviation data to obtain multiple groups of angle deviation data.

[0110] Specifically, angle deviation calculations are performed separately based on multiple sets of fitting angle data obtained through the primary model of the pole skeleton and multiple sets of measured angle data calculated through the spirit level. The angle deviation calculation is to calculate and compare the multiple sets of fitting angle data obtained through the primary model of the pole skeleton with the multiple sets of actually measured angle data one by one, that is, first calculate the difference between the multiple sets of fitting angle data and their average, then add the multiple sets of fitting angle data and the above difference calculation result, and finally divide the above addition result by the number of sets of fitting angle data to obtain multiple sets of angle deviation data, so that the multiple sets of fitting angle data that generate the multiple sets of angle deviation data correspond one to one with the multiple sets of measured angle data, where the multiple sets of angle deviation data are a set of results calculated by the angle deviation. On this basis, a more efficient safety warning assessment is performed on the pole tower construction.

[0111] In step S500, the primary model of the pole frame is optimized according to the angle deviation data to generate a pole frame angle recognition model, including:

[0112] The angle deviation data is input into the primary model of the pole frame, and the three-dimensional structure data of the primary model of the pole frame is iteratively adjusted according to the angle deviation data. When a preset number of iterations is reached, the pole frame angle recognition model is generated.

[0113] The pole skeleton angle recognition model is a three-dimensional image model. Angle data can be directly obtained based on this model. The optimization logic of the pole skeleton primary model involves continuously correcting and adjusting the three-dimensional structural data of the pole skeleton primary model based on the angle deviation data. This involves continuously adjusting the pole tower primary model using the angle deviation data. The primary model is optimized based on the angle deviation data to obtain a pole skeleton structural model that faithfully reproduces the actual pole tower structure. Angle data can then be directly derived based on this structural model.

[0114] Specifically, the data from the previously generated primary mast skeleton model is collated and input based on the multiple sets of angle deviation data obtained above. The resulting data, representing the mast skeleton, the lines of the multiple support ropes, and the angles between the support ropes and the mast axis, is iterated using this data to generate the three-dimensional structural data. This model then generates a mast skeleton angle recognition model, which can be used to better assess safety during mast tower construction.

[0115] See also Figure 5 As shown, in step S600, a safety warning is performed based on the real-time angle data, including:

[0116] Step S610: generating an angle change curve according to the real-time angle data;

[0117] Step S620: Calculate the real-time angle change rate according to the angle change curve;

[0118] Step S630: When the real-time angle changes at a uniform rate and within a preset angle range, a first warning message is generated. The first warning message is used to remind the construction personnel to adjust the angle between the holding pole axis and the supporting rope; the preset angle range is 30° to 60°.

[0119] Step S640: When the rate of change of the real-time angle is accelerating and the range of change is within the preset angle range, a second warning message is generated. The second warning message is used to remind the construction personnel to suspend the pole holding operation.

[0120] Specifically, multiple sets of measured angle data obtained by the angle sensing module are used to generate an angle change curve. The angle change curve uses time in seconds as the x-axis and an angle in 5° as the y-axis to construct a rectangular coordinate system. Marking and connecting points are performed in the coordinate system according to the real-time angle data. On this basis, an angle change curve is generated, and the angle change rate of the obtained angle change curve is evaluated. When the angle increases from 45° to 60° or decreases from 45° to 30°, the angle parameter threshold is regarded as the preset angle range. At the same time, when the angle starts to increase or decrease from 45°, the rate of increase or decrease of the angle is calculated, where the angular acceleration formula is α=δω / δt=dω / dt. When the movement is uniformly accelerated, α=w / t, which is regarded as the angle increasing or decreasing at a uniform speed. When the angle increases from 45° to 60° or decreases from 45° to 30°, the angle change parameter is obtained, which is evaluated as a safe value and meets the preset angle range, and the first warning information is output. The first warning information is used to prompt construction personnel to adjust the angles between multiple supporting ropes and the axis of the holding pole within the safe value. When α=δω / δt, it is considered that the angle is increasing or decreasing rapidly. The angle change parameter is obtained when the angle increases from 45° to 60° or decreases from 45° to 30°, which is evaluated as an unsafe value and does not meet the preset angle range. A second early warning information is output, wherein the second early warning information is used to remind the construction personnel to suspend the construction of the pole tower assembly. Finally, the pole tower assembly construction is adjusted in time according to the input early warning information, thereby achieving the technical effect of providing a reference for the safety early warning assessment of the pole tower assembly construction.

[0121] Step S600 of this embodiment inputs the real-time pole-holding tower construction image acquired by the image acquisition device into the pole skeleton angle recognition model to obtain real-time angle data between multiple supporting ropes and the pole axis. On this basis, a real-time angle data set composed of the acquired real-time angle data between multiple supporting ropes and the pole axis is used to perform an early warning assessment of the safety of the pole-holding tower construction. That is, a real-time comparison is performed between the acquired angle data set and the limited safety angle range (preset angle range). If the real-time angle data is within the preset angle range, the pole-holding tower construction is determined to be safe. If the real-time angle data exceeds the limited safety angle range, the pole-holding tower construction is determined to be unsafe, thereby improving the safety of the pole-holding tower construction.

[0122] See also Figure 6 As shown, the present invention also provides a safety warning device for tower assembly construction, comprising:

[0123] Acquisition module 1, used to collect images of the tower assembly construction through an image acquisition device;

[0124] Model generation module 2, used to process the tower assembly construction image to generate a primary model of the tower skeleton;

[0125] The measurement module 3 is used to obtain multiple sets of fitted angle data between the pole axis and the supporting rope based on the primary model of the pole skeleton, and to collect multiple sets of measured angle data between the pole axis and the supporting rope in practice through the angle sensor device;

[0126] Calculation module 4, used for calculating multiple sets of angle deviation data based on multiple sets of fitting angle data and multiple sets of measured angle data;

[0127] Optimization module 5, used to optimize the primary model of the pole frame according to the angle deviation data to generate a pole frame angle recognition model;

[0128] The early warning module 6 is used to collect real-time tower assembly construction images and input them into the tower frame angle recognition model, output real-time angle data, and issue safety early warnings based on the real-time angle data.

[0129] See also Figure 7 As shown, the model generation module 2 may include:

[0130] The image acquisition module 21 is used to perform binarization processing on the tower assembly construction image to obtain a pre-processed image of the tower assembly construction;

[0131] The skeleton data acquisition module 22 is used to perform skeleton recognition based on the pre-processed image of the tower assembly construction to obtain the overall skeleton data information;

[0132] The fitting model generation module 23 is used to establish a three-dimensional model framework of the pole tower group, fit the overall skeleton data information with the three-dimensional model framework of the pole tower group, and obtain a fitting model;

[0133] The primary model acquisition module 24 is used to filter and eliminate the tower skeleton data information in the fitting model to obtain a primary model of the pole skeleton.

[0134] See also Figure 8 As shown, the fitting model generation module 23 may include:

[0135] The first target tower acquisition module 231 is used to obtain a target tower model for the pole-holding tower assembly construction;

[0136] The target index search module 232 is used to traverse the pre-built pole tower database and search the pole tower database for a target index corresponding to the target tower model;

[0137] A three-dimensional model building module 233 is configured to extract relevant data blocks from the pole tower database according to the target index and build a three-dimensional model architecture of the pole tower according to the relevant data blocks;

[0138] Skeleton data extraction module 234, used to extract skeleton monomer data and skeleton connection nodes from the overall skeleton data information;

[0139] The skeleton unit positioning module 235 is used to position the skeleton units of the three-dimensional model of the tower group according to the skeleton connection nodes;

[0140] The fitting model obtaining module 236 is used to fit the three-dimensional model structure of the tower group according to the skeleton monomer positioning result and the skeleton monomer data to obtain a fitting model.

[0141] See also Figure 9 As shown, the primary model acquisition module 24 may include:

[0142] The second target tower acquisition module 241 is used to obtain the target tower model for the pole-holding tower assembly construction;

[0143] The multi-angle image acquisition module 242 is used to collect multi-angle tower images of the same specifications as the target tower model from the big data according to the target tower model;

[0144] A feature set module 243 is used to extract a relative position feature set of the tower skeleton and a ratio feature set of the tower skeleton based on the multi-angle tower image;

[0145] The recognition model training module 244 is used to establish a tower frame recognition model based on the BP neural network, and train the tower frame recognition model based on multi-angle tower images, the tower frame relative position feature set, and the tower frame ratio feature set;

[0146] The identification result obtaining module 245 is used to input the fitting model into the trained tower skeleton recognition model to obtain the tower skeleton identification result of the fitting model;

[0147] The primary model generation module 246 is used to remove the tower skeleton data information in the fitting model according to the tower skeleton identification result to obtain the primary model of the pole skeleton.

[0148] Furthermore, the optimization module 5 is used to input the angle deviation data into the primary model of the pole skeleton, iteratively adjust the three-dimensional structure data of the primary model of the pole skeleton according to the angle deviation data, and generate the pole skeleton angle recognition model when a preset number of iterations is reached.

[0149] See also Figure 10 As shown, the early warning module 6 may include:

[0150] A curve generating module 61 is used to generate an angle change curve according to real-time angle data;

[0151] The parameter change module 62 is used to calculate the real-time angle change rate according to the angle change curve;

[0152] A first warning information output module 63 is configured to generate a first warning information when the rate of change of the real-time angle is uniform and the range of change is within a preset angle range;

[0153] The second warning information output module 64 is configured to generate a second warning information when the rate of change of the real-time angle is accelerating and the range of change is within a preset angle range.

[0154] Although preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they are aware of the basic inventive concepts. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the invention. Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the invention. Thus, the present invention is intended to include such changes and modifications as fall within the scope of the claims and their equivalents.

Claims

1. A safety early warning method for tower assembly construction, characterized in that: include: The image acquisition device is used to collect images of the tower assembly construction; Processing the mast tower construction image to generate a primary mast skeleton model; Acquire multiple sets of fitted angle data between the pole axis and the supporting rope based on the primary model of the pole skeleton, and collect multiple sets of measured angle data between the pole axis and the supporting rope in practice through an angle sensor device; Calculating multiple sets of angle deviation data based on the multiple sets of fitting angle data and the multiple sets of measured angle data; Optimizing the primary model of the pole frame according to the angle deviation data to generate a pole frame angle recognition model; Collect real-time tower assembly construction images and input them into the tower frame angle recognition model, output real-time angle data, and issue safety warnings based on the real-time angle data; The mast tower construction image is processed to generate a primary mast skeleton model, including: Binarizing the tower assembly construction image to obtain a pre-processed tower assembly construction image. Perform skeleton recognition based on the pre-processed image of the tower assembly construction to obtain overall skeleton data information; Establishing a three-dimensional model framework of a pole tower assembly, and fitting the overall skeleton data information with the three-dimensional model framework of the pole tower assembly to obtain a fitting model; Filtering and eliminating the tower skeleton data information in the fitting model to obtain the primary model of the pole skeleton; Establishing a three-dimensional model framework of a pole tower assembly, fitting the overall skeleton data information with the three-dimensional model framework of the pole tower assembly to obtain a fitting model, including: Obtain the target tower model for pole-holding tower construction; Traversing a pre-built pole tower database, searching for a target index corresponding to the target tower model from the pole tower database; Extracting relevant data blocks from the pole tower database according to the target index, and establishing a three-dimensional model architecture of the pole tower according to the relevant data blocks; Extracting skeleton monomer data and skeleton connection nodes from the overall skeleton data information; Positioning the skeleton monomers of the three-dimensional model of the tower group according to the skeleton connection nodes; The three-dimensional model architecture of the pole group tower is fitted according to the skeleton monomer positioning result and the skeleton monomer data to obtain a fitting model.

2. The method according to claim 1, characterized in that Based on the primary model of the pole frame, multiple sets of fitting angle data between the pole axis and the supporting rope are obtained, including: Based on the primary model of the pole skeleton, a rectangular coordinate system is established with the contact point between the pole and the horizontal plane as the origin, the horizontal plane as the x-axis, and the pole as the y-axis. The angle data between each supporting rope and the pole axis are obtained in the rectangular coordinate system. The obtained angle data between the multiple supporting ropes and the pole axis are integrated to obtain multiple sets of fitting angle data.

3. The method according to claim 1, characterized in that Calculating multiple sets of angle deviation data based on the multiple sets of fitting angle data and the multiple sets of measured angle data includes: Each fitted angle data is compared with the corresponding measured angle data one by one, and the difference is integrated as the angle deviation data to obtain multiple groups of angle deviation data.

4. The method according to claim 1, wherein The tower skeleton data information in the fitting model is filtered and eliminated to obtain the primary model of the pole skeleton, including: Obtain the target tower model for pole-holding tower construction; Collecting multi-angle tower images of the same specifications as the target tower model from big data according to the target tower model; Extracting a tower skeleton relative position feature set and a tower skeleton proportion feature set according to the multi-angle tower image; Establishing a tower skeleton recognition model based on a BP neural network, and training the tower skeleton recognition model according to the multi-angle tower images, the tower skeleton relative position feature set, and the tower skeleton ratio feature set; Inputting the fitting model into the trained tower skeleton recognition model to obtain the tower skeleton identification result of the fitting model; The tower skeleton data information in the fitting model is eliminated according to the tower skeleton identification result to obtain the primary model of the pole skeleton.

5. The method according to claim 1, wherein The primary model of the pole frame is optimized according to the angle deviation data to generate a pole frame angle recognition model, including: The angle deviation data is input into the primary model of the pole frame, and the three-dimensional structure data of the primary model of the pole frame is iteratively adjusted according to the angle deviation data. When a preset number of iterations is reached, the pole frame angle recognition model is generated.

6. The method according to claim 1, characterized in that Providing safety warnings based on the real-time angle data includes: generating an angle change curve according to the real-time angle data; Calculating the real-time angle change rate according to the angle change curve; When the real-time angle changes at a constant rate and within a preset angle range, a first warning message is generated, which is used to remind construction workers to adjust the angle between the holding pole axis and the supporting rope; When the rate of change of the real-time angle is accelerated and the range of change is within the preset angle range, a second warning message is generated, and the second warning message is used to remind the construction personnel to suspend the pole holding operation.

7. The method according to claim 6, characterized in that The preset angle range is 30° to 60°.

8. A safety warning device for tower assembly construction, characterized in that: include: An acquisition module is used to acquire images of the tower assembly construction process through an image acquisition device; A model generation module is used to process the mast tower construction image to generate a primary model of the mast skeleton; a measurement module for obtaining multiple sets of fitted angle data between the pole axis and the supporting rope based on the primary model of the pole skeleton, and collecting multiple sets of measured angle data between the pole axis and the supporting rope in practice through an angle sensor; A calculation module, configured to calculate multiple sets of angle deviation data based on the multiple sets of fitting angle data and the multiple sets of measured angle data; an optimization module, configured to optimize the primary model of the pole skeleton according to the angle deviation data to generate a pole skeleton angle recognition model; An early warning module is used to collect real-time images of the tower assembly construction and input them into the tower frame angle recognition model, output real-time angle data, and issue safety warnings based on the real-time angle data; The model generation module includes: An image acquisition module is used to perform binarization processing on the pole-holding tower assembly construction image to obtain a pre-processed pole-holding tower assembly construction image; A skeleton data acquisition module is used to perform skeleton recognition based on the pre-processed image of the tower assembly construction to obtain overall skeleton data information; A fitting model generation module is used to establish a three-dimensional model framework of a pole-mounted tower, and fit the overall skeleton data information with the three-dimensional model framework of the pole-mounted tower to obtain a fitting model; A primary model acquisition module is used to filter and eliminate the tower skeleton data information in the fitting model to obtain the primary model of the pole skeleton; The fitting model generation module includes: The first target tower acquisition module is used to obtain the target tower model for the pole-holding tower assembly construction; A target index search module is used to traverse a pre-built pole tower database and search the pole tower database for a target index corresponding to the target tower model; A three-dimensional model building module is used to extract relevant data blocks from the pole tower database according to the target index, and build a three-dimensional model architecture of the pole tower according to the relevant data blocks; A skeleton data extraction module, configured to extract skeleton monomer data and skeleton connection nodes from the overall skeleton data information; A skeleton unit positioning module is used to position the skeleton units of the three-dimensional model of the mast tower according to the skeleton connection nodes; The fitting model obtaining module is used to fit the three-dimensional model architecture of the pole group tower according to the skeleton monomer positioning result and the skeleton monomer data to obtain a fitting model.

Images