Steel structure positioning system and method based on computer vision and ai

Abstract

The application provides a steel structure positioning system and method based on computer vision and AI, which comprises a preset module, a shooting module, an image processing module, a sensor module and an adjustment module. The preset module is used for setting the preset rectangular structure form of the steel structure corridor and the accurate poses of the corresponding four corner points. The shooting module is used for shooting the first image of the steel structure corridor in real time. The image processing module is used for obtaining the real-time deviation angle of the steel structure corridor. The sensor module is used for obtaining the real-time poses of the four corner points of the steel structure corridor and the real-time pose difference of the four corner points. The adjustment module is used for adjusting the steel structure corridor in real time until the real-time deviation angle is less than the preset deviation angle threshold and the real-time pose difference of the four corner points is less than the preset pose difference threshold. The application can ensure that all node qualities are qualified in one installation process, so as to ensure the consistency and stability of the steel structure corridor positioning in a long-distance and high-complexity environment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of steel structure positioning, and specifically relates to a steel structure positioning system and method based on computer vision and AI. BACKGROUND

[0002] In modern construction engineering, steel structures are widely used in the construction of large buildings due to their high strength, light weight, durability, and recyclability. In particular, during the installation process of long-distance steel structure corridors, accurate positioning is the key to ensuring engineering quality. However, traditional theodolite measurement methods are difficult to ensure high-precision positioning consistency due to human factors and equipment limitations. Existing solutions mainly achieve steel structure positioning through manual measurement and adjustment. Although this approach can achieve certain positioning functions, it has problems such as low precision, low efficiency, and high cost in operation. In addition, due to the strong subjectivity of manual operation, the consistency and stability of the positioning results may be insufficient. Existing steel structure positioning methods mainly rely on manual operation, which not only consumes time and effort, but also may have insufficient precision and stability due to human factors. In the face of long-distance and high-complexity environments, existing methods are difficult to ensure the consistency and stability of positioning. When dealing with large amounts of image data and performing complex calculations, these methods may have slow processing speed and low precision.

[0003] Chinese invention patent CN111677294A discloses a steel structure intelligent installation method based on digital twin technology. Before steel structure construction, Revit is used to model the steel structure to obtain a steel structure model. The steel structure model is imported into Ansys for mechanical performance analysis. After calculation, a construction animation is made for the steel structure node, and the construction process is technically disclosed to the steel structure process through VR glasses. During the construction process, the model established by Revit is strictly followed, and a three-dimensional scanner is used to record each steel structure installation and construction step in time to form point cloud data inverse modeling compared with the design model, and the steel structure entity installation is adjusted to ensure the consistency of construction and design. Light sensors are arranged at each connection node after the steel structure entity assembly is completed, and the data images of the sensors are used to timely investigate and rectify construction quality problems. From design to construction, virtual reality is realized in each stage to achieve digital twin, improving the quality and efficiency of steel structure installation and construction.

[0004] In summary, the steel structure intelligent installation method based on the digital twin technology disclosed mainly aims at the problems of poor installation construction quality and complex process of the existing steel structure beams and columns. After the steel structure is assembled, light sensors are arranged at the connection nodes to timely investigate and rectify the construction quality problems through the data images of the sensors, instead of ensuring the quality of all nodes in one installation process. Moreover, the related technology cannot meet the consistency and stability of the steel structure corridor installation positioning in a long-distance and high-complexity environment. SUMMARY

[0005] In view of the deficiencies of the prior art, the purpose of the present application is to provide a steel structure positioning system and method based on computer vision and AI, which can ensure the quality of all nodes in one installation process, and guarantee the consistency and stability of the steel structure corridor positioning in a long-distance and high-complexity environment.

[0006] In order to solve the above technical problems, the technical scheme adopted by the present application is as follows:

[0007] The steel structure positioning system based on computer vision and AI comprises:

[0008] A preset module is configured to set the preset rectangular structure form of the steel structure corridor and the accurate poses of the four corner points corresponding to the preset rectangular structure form;

[0009] A shooting module is configured to shoot a first image of the steel structure corridor in real time;

[0010] An image processing module is configured to pre-process the first image to obtain a second image, identify the real-time rectangular structure form of the steel structure corridor based on the second image, and obtain the real-time offset angle of the steel structure corridor based on the real-time rectangular structure form and the preset rectangular structure form;

[0011] A sensor module is configured to obtain the real-time poses of the four corner points of the steel structure corridor, and obtain the real-time pose difference of the four corner points based on the accurate poses and the real-time poses of the four corner points;

[0012] An adjustment module is configured to adjust the steel structure corridor in real time until the real-time offset angle is less than a preset offset angle threshold and the real-time pose difference of the four corner points is less than a preset pose difference threshold.

[0013] Further, the pose comprises attitude information and position information, the sensor module comprises an inertial measurement unit and a global positioning system, the inertial measurement unit is configured to measure the attitude information of the four corner points of the steel structure corridor when moving, and the attitude information comprises angular velocity and acceleration information, and the global positioning system is configured to provide the position information of the four corner points of the steel structure corridor.

[0014] Further, the adjusting module comprises a magnetic suction plate, a support, a handle and a pushing piece, the magnetic suction plate is used for adsorbing the steel structure corridor, the support is rotatably connected to the magnetic suction plate, the pushing piece comprises a pushing seat and a pushing shaft, one side of the support is fixed with the handle, and the other side of the support is fixed with the pushing seat, the handle is used for hooking the steel structure corridor, and the pushing shaft is fixed on the steel structure support frame.

[0015] Further, the handle comprises a first U-shaped arm and a second U-shaped arm, the first U-shaped arm is fixed with the support, the first U-shaped arm comprises a first end and a second end, the second U-shaped arm comprises a third end and a fourth end, the first end and the third end are connected through a first straight rod, the second end and the fourth end are connected through a second straight rod, the included angle between the first straight rod and the first end is 110-150 degrees, the included angle between the third end and the first straight rod is 110-150 degrees, the included angle between the second straight rod and the second end is 110-150 degrees, and the included angle between the fourth end and the second straight rod is 110-150 degrees.

[0016] The steel structure positioning method based on computer vision and AI is applied to the steel structure positioning system based on computer vision and AI, and comprises the following steps:

[0017] The preset rectangular structure form of the steel structure corridor and the accurate poses of the four corner points corresponding to the preset rectangular structure form are set;

[0018] The first image of the steel structure corridor is captured in real time;

[0019] The first image is preprocessed to obtain a second image, the real-time rectangular structure form of the steel structure corridor is recognized based on the second image, and the real-time offset angle of the steel structure corridor is obtained based on the real-time rectangular structure form and the preset rectangular structure form;

[0020] The real-time poses of the four corner points of the steel structure corridor are obtained, and the real-time pose difference of the four corner points is obtained based on the accurate poses and the real-time poses of the four corner points;

[0021] The steel structure corridor is adjusted in real time until the real-time offset angle is less than the preset offset angle threshold and the real-time pose difference of the four corner points is less than the preset pose difference threshold.

[0022] Further, the method for preprocessing the first image to obtain the second image is:

[0023] Gray scale transformation and median filtering denoising processing.

[0024] Further, based on the second image, the real-time rectangular structure form of the steel structure corridor is identified, and the real-time offset angle of the steel structure corridor is obtained based on the real-time rectangular structure form and the preset rectangular structure form.

[0025] The edge of the steel structure corridor is identified by using a Canny edge detection algorithm, and morphological operations are used to extract the structural form features of the edge of the steel structure corridor, including dilation and corrosion.

[0026] SIFT and SURF algorithms are used to extract key feature points of the structural form features, which are four corner points of the steel structure corridor, and a real-time rectangular structure form is established based on the extracted four corner points of the steel structure corridor.

[0027] Four edges of the steel structure corridor are obtained based on the preset rectangular structure form, and the actual four edges of the steel structure corridor are obtained based on the real-time rectangular structure form, and the values of the four included angles are averaged to obtain the real-time offset angle of the steel structure corridor based on the real-time rectangular structure form and the preset rectangular structure form.

[0028] Further, a deep learning model is established, and the specific method is as follows:

[0029] A certain amount of image data of the installation process of the steel structure corridor is collected to establish an image set;

[0030] The image set is manually labeled to label the steel structure corridor;

[0031] A VGG-16 convolutional neural network is used for model training;

[0032] The trained convolutional neural network is used for real-time rectangular structure form identification of the steel structure corridor;

[0033] The real-time offset angle of the steel structure corridor is obtained based on the real-time rectangular structure form and the preset rectangular structure form.

[0034] Further, according to the real-time offset angle and the real-time pose error, the direction and amplitude that need to be adjusted are determined, and the proportional, integral and differential parts of the PID controller are combined to generate accurate correction instructions for real-time adjustment of the steel structure corridor.

[0035] Further, a virtual model of the steel structure corridor is established in 3ds Max, and a virtual body is added to each corner point of the virtual model of the steel structure corridor, and the real-time pose of the four corner points of the steel structure corridor is obtained by tracking the displacement changes of the four virtual bodies.

[0036] Compared with the prior art, the present application has the following advantages and beneficial effects:

[0037] The present application realizes high-precision positioning and correction of steel structure corridors in long-distance and high-complexity environments. The system realizes automatic positioning and correction of steel structures by combining a shooting module, an image processing module, a sensor module, and an adjustment module. System integration and optimization improve system operation efficiency and positioning accuracy, enhance system robustness and anti-interference ability through multiple tests and calibration. Through the above technical means, the present application realizes high-precision positioning and correction of steel structures, improves work efficiency and accuracy, reduces the error of manual operation, adapts to complex construction environment and reduces cost. The technology has wide application prospect and market demand in the fields of building engineering, mechanical manufacturing and intelligent robots. BRIEF DESCRIPTION OF DRAWINGS

[0038] The drawings described herein are used to provide further understanding of the present application and form a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation of the present application. In the drawings:

[0039] Figure 1 is a schematic diagram of the steel structure positioning system based on computer vision and AI of the present application;

[0040] Figure 2 is a structural schematic diagram of the present application applied to steel structure corridor field adjustment;

[0041] Figure 3 is Figure 2 is an enlarged schematic diagram of A;

[0042] Figure 4 is a schematic diagram of the rapid alignment system of the present application;

[0043] Figure 5 is a flowchart of the steel structure positioning method based on computer vision and AI of the present application.

[0044] 1, a shooting module; 2, a sensor module; 3, an adjustment module; 31, a magnetic chuck; 32, a support; 33, a handle; 331, a first U-shaped arm; 3311, a first end; 3312, a second end; 332, a second U-shaped wall; 3321, a third end; 3322, a fourth end; 333, a first straight rod; 334, a second straight rod; 34, a pusher; 341, a push shaft; 342, a push seat; 35, a hydraulic station; 4, a steel structure corridor; 5, a steel structure support frame; DETAILED DESCRIPTION

[0045] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application.

[0046] In the following description, numerous specific details are set forth to provide a more thorough understanding of the present application. However, it will be apparent to one of skill in the art that the present application can be practiced without one or more of these specific details. In other instances, well-known features have not been described in order to avoid obscuring the present application.

[0047] It should be understood that the present application can be practiced with modification and alteration, and that the present application should not be construed or limited to the examples described herein. Accordingly, the description is to be regarded as illustrative in nature and not a limitation on the scope of the present application.

[0048] Embodiment 1

[0049] Embodiment 1 provides a steel structure positioning system based on computer vision and AI, as shown in Figure 1 and Figure 2 , comprising:

[0050] a preset module for setting a preset rectangular structure form of the steel structure corridor 4 and accurate poses of four corner points corresponding to the preset rectangular structure form;

[0051] a shooting module 1 for shooting a first image of the steel structure corridor 4 in real time;

[0052] an image processing module for pre-processing the first image to obtain a second image, recognizing a real-time rectangular structure form of the steel structure corridor 4 based on the second image, and obtaining a real-time offset angle of the steel structure corridor 4 based on the real-time rectangular structure form and the preset rectangular structure form;

[0053] a sensor module 2 for obtaining real-time poses of the four corner points of the steel structure corridor 4, and obtaining real-time pose differences of the four corner points based on the accurate poses and the real-time poses of the four corner points;

[0054] an adjustment module 3 for adjusting the steel structure corridor 4 in real time until the real-time offset angle is less than a preset offset angle threshold and the real-time pose differences of the four corner points are all less than a preset pose difference threshold.

[0055] The computer vision and AI-based steel structure positioning system provided in Embodiment 1 is used to achieve high-precision positioning and correction of the steel structure corridor 4 in a long-distance and high-complexity environment. The system realizes automatic positioning and correction of the steel structure by combining the shooting module 1, the image processing module, the sensor module 2, and the adjustment module 3. System integration and optimization improve system operation efficiency and positioning accuracy, and enhance the robustness and anti-interference ability of the system. Through the above technical means, the present application realizes high-precision positioning and correction of the steel structure, improves work efficiency and accuracy, reduces the error of manual operation, adapts to complex construction environment and reduces cost. The technology has wide application prospect and market demand in the fields of building engineering, mechanical manufacturing and intelligent robots.

[0056] In the shooting module 1 of the present embodiment, a plurality of high-definition cameras are installed on the steel structure corridor 4 and / or the steel structure support frame 5. These cameras should have high resolution, high frame rate and low delay, etc. in order to capture real-time images. At the same time, high-precision servo motors or hydraulic systems are equipped to perform correction and adjustment operations of the steel structure.

[0057] In the sensor module 2 of the present embodiment, the pose includes attitude information and position information. The sensor module 2 includes an inertial measurement unit (IMU) and a global positioning system (GPS sensor) to realize accurate angle and position measurement. The inertial measurement unit is used to measure the attitude information of the four corner points of the steel structure corridor 4 when moving, and the attitude information includes angular velocity and acceleration information. The global positioning system is used to provide the position information of the four corner points of the steel structure corridor 4.

[0058] Specifically, the inertial measurement unit (IMU) includes a three-axis gyroscope, a three-axis accelerometer and a magnetometer to measure angular velocity and acceleration. The global positioning system (GPS sensor) provides high-precision position data. After fusion of IMU and GPS data, more accurate position and attitude information is obtained. Before actual application, IMU and GPS need to be calibrated to ensure accurate measurement data.

[0059] In the adjustment module 3 of the present embodiment, as shown in Figures 2-4 , the adjustment module 3 includes a magnetic chuck 31, a bracket 32, a handle 33 and a pushing piece 34. The magnetic chuck 31 is used to adsorb the steel structure corridor 4. The bracket 32 is rotatably connected to the magnetic chuck 31. The pushing piece 34 includes a pushing seat 342 and a pushing shaft 341. One side of the bracket 32 is fixed with the handle 33, and the other side is fixed with the pushing seat 342. The handle 33 is used to hook the steel structure corridor 4. The pushing shaft 341 is fixed on the steel structure support frame 5.

[0060] Specifically, the magnetic chuck 31 adopts an electrically controlled permanent magnetic chuck, and the generation or disappearance of the magnetic field is realized through the electric control principle, so as to flexibly adsorb or release the load. The electrically controlled permanent magnetic chuck is operated through a control unit, and the control unit is integrated on the handle 33, facilitating operation.

[0061] The pushing member 34 is a hydraulic cylinder, which is powered by a hydraulic station 35, and provides necessary supporting force and position adjustment to ensure the smooth and accurate alignment of the steel structure corridor 4. The adjustment module 3 is installed at the closest position of the steel structure corridor 4 and the steel structure support frame 5, the steel structure corridor 4 is fixed by the electrically controlled permanent magnetic chuck, and the hydraulic cylinder is adjusted accurately to realize rapid alignment. When disassembling, the state of the electrically controlled permanent magnetic chuck is switched through the control unit, and the steel structure corridor 4 is quickly disassembled by the lever action.

[0062] As shown in Figure 3 The handle 33 includes a first U-shaped arm 331 and a second U-shaped arm 332, the first U-shaped arm 331 is fixed with the support 32, the first U-shaped arm 331 includes a first end 3311 and a second end 3312, the second U-shaped arm 332 includes a third end 3321 and a fourth end 3322, the first end 3311 and the third end 3321 are connected by a first straight rod 333, the second end 3312 and the fourth end 3322 are connected by a second straight rod 334, the included angle between the first straight rod 333 and the first end 3311 is 110-150 degrees, the included angle between the third end 3321 and the first straight rod 333 is 110-150 degrees, the included angle between the second straight rod 334 and the second end 3312 is 110-150 degrees, and the included angle between the fourth end 3322 and the second straight rod 334 is 110-150 degrees. By arranging the first U-shaped arm 331, the second U-shaped arm 332, the first straight rod 333 and the second straight rod 334, the steel structure corridor 4 can be hooked, and the real-time offset angle and the real-time pose difference of the steel structure corridor 4 can be adjusted to ensure that the real-time offset angle is less than the preset offset angle threshold and the real-time pose difference of the four corner points is less than the preset pose difference threshold.

[0063] Specifically, one end of the first straight rod 333 is fixed with the first end 3311, and the other end is fixed with the third end 3321, one end of the second straight rod 333 is fixed with the second end 3312, and the other end is fixed with the fourth end 3322. The angle of the support 32 is adjusted by the extension and retraction of the pushing shaft 341, and then the angle of the first U-shaped arm 331 is adjusted, so that the angle of the second U-shaped arm 332 and the steel structure corridor 4 is adjusted, and the real-time offset angle and the real-time pose difference of the steel structure corridor 4 are adjusted to ensure that the real-time offset angle is less than the preset offset angle threshold and the real-time pose difference of the four corner points is less than the preset pose difference threshold.

[0064] Embodiment 2

[0065] Embodiment 2 provides a steel structure positioning method based on computer vision and AI, which is applied to the computer vision and AI-based steel structure positioning system as described above, as shown in Figure 5 and includes the following steps:

[0066] Step S1: Set the preset rectangular structure form of the steel structure corridor 4 and the accurate poses of the four corner points corresponding to the preset rectangular structure form.

[0067] Step S2: Real-time shooting of the first image of the steel structure corridor 4.

[0068] Step S3: Preprocessing of the first image to obtain the second image, real-time rectangular structure form of the steel structure corridor 4 based on the second image, and real-time offset angle of the steel structure corridor 4 based on the real-time rectangular structure form and the preset rectangular structure form.

[0069] Step S4: Obtain the real-time poses of the four corner points of the steel structure corridor 4, and obtain the real-time pose difference of the four corner points based on the accurate poses and the real-time poses of the four corner points.

[0070] Step S5: Real-time adjustment of the steel structure corridor 4 until the real-time offset angle is less than the preset offset angle threshold and the real-time pose difference of the four corner points is less than the preset pose difference threshold.

[0071] In step S1 of this embodiment, a virtual model of the steel structure corridor 4 is established in 3ds Max, and then the preset rectangular structure form of the steel structure corridor 4 and the accurate poses of the four corner points corresponding to the preset rectangular structure form are set.

[0072] Specifically, virtual bodies are added to each corner point of the virtual model, which will be used to track the displacement changes of the corner points. The virtual bodies are bound to the corner points using the binding tool of 3ds Max, ensuring that the virtual bodies move with the corner points. The displacement data of the virtual bodies during the entire hoisting process is recorded using the animation and trajectory recording function of 3ds Max, and the real-time poses of the four corner points of the steel structure corridor 4 are obtained by tracking the displacement changes of the four corner points of the steel structure corridor 4 through the four virtual bodies.

[0073] In terms of installation and monitoring in the actual environment, IMU (Inertial Measurement Unit) and GPS (Global Positioning System) sensors are installed on the corridor to achieve precise angle and coordinate measurement, ensuring that the coordinate system of the recording module is consistent with the coordinate system in 3ds Max. During hoisting, the displacement data of the corner points is monitored in real time and compared with the preset data in 3ds Max, and the hoisting position is adjusted to ensure that the final position is aligned with the design coordinates.

[0074] In step S2 of this embodiment, image quality and feature extraction accuracy are improved through processing steps such as grayscale transformation, median filter denoising, and Canny edge detection. The Canny edge detection algorithm can accurately capture edge details in the image, enhancing the accuracy of subsequent analysis and adjustment.

[0075] Specifically, the method for pre-processing the first image to obtain the second image includes grayscale transformation and median filter denoising.

[0076] In step S2 of this embodiment, the real-time rectangular structure morphology of the steel structure corridor 4 is identified based on the second image. The real-time offset angle of the steel structure corridor 4 is obtained based on the real-time rectangular structure morphology and the preset rectangular structure morphology.

[0077] The edge of the steel structure corridor 4 is identified using the Canny edge detection algorithm, and morphological operations such as dilation and erosion are used to extract the structural morphology features of the edge of the steel structure corridor 4.

[0078] SIFT and SURF algorithms are used to extract key feature points of the structural morphology features. The key feature points are the four corner points of the steel structure corridor 4. The real-time rectangular structure morphology is established based on the extracted four corner points of the steel structure corridor 4.

[0079] The Canny edge detection algorithm is a classic image processing technique. This algorithm achieves edge detection through five main steps: first, Gaussian smoothing is performed on the image to reduce noise effects; second, the gradient strength and direction of each pixel point in the image are calculated; third, non-maximum suppression is used to thin the edges, retaining only the local maximum values; fourth, high and low threshold values are applied for detection, dividing the gradient strength into strong edges, weak edges, and non-edges; and finally, strong and weak edges are connected to form the final edge detection result. The Canny algorithm is known for its high detection rate, high positioning accuracy, and low response redundancy, effectively balancing noise suppression and edge positioning, and is suitable for various application scenarios such as object recognition, image segmentation, and contour extraction.

[0080] The Canny edge detection algorithm is used to identify the edge of the steel structure, and morphological operations such as dilation and erosion are applied to further extract structural morphology features. Meanwhile, SIFT and SURF algorithms are used to extract key feature points, which will be helpful for subsequent matching and positioning work.

[0081] SIFT (Scale-Invariant Feature Transform) and SURF (Speeded-Up Robust Features) are two commonly used image feature extraction algorithms, widely used in image recognition and matching tasks. SIFT detects and describes key points in images, achieving scale and rotation invariance, with high robustness and matching accuracy, but with high computational complexity. SURF, on the other hand, maintains high matching accuracy while improving computational speed, making it suitable for real-time applications. The main difference between the two is computational efficiency and feature descriptor dimension: SIFT descriptor is 128-dimensional, and SURF is 64-dimensional. SURF is more efficient but less robust than SIFT in complex scenarios.

[0082] Based on the preset rectangular structure morphology, the four sides of the steel structure corridor 4 are obtained, and based on the real-time rectangular structure morphology, the actual four sides of the steel structure corridor 4 are obtained. According to the included angle of the four sides and the actual four sides corresponding to each other, the values of the four included angles are averaged to obtain the real-time offset angle of the steel structure corridor 4.

[0083] In step S2 of this embodiment, a deep learning model is also established, and the specific method is:

[0084] A certain amount of image data of the steel structure corridor 4 installation process is collected to establish an image set;

[0085] The image set is manually annotated to annotate the steel structure corridor 4;

[0086] A VGG-16 convolutional neural network is used for model training;

[0087] The trained convolutional neural network is used for real-time rectangular structure morphology recognition of the steel structure corridor 4;

[0088] Based on the real-time rectangular structure morphology and the preset rectangular structure morphology, the real-time offset angle of the steel structure corridor 4 is obtained.

[0089] In this embodiment, a large amount of image data of the steel structure corridor 4 installation process is collected and manually annotated to form a training data set. The VGG-16 convolutional neural network architecture is selected for model training. Through data enhancement techniques such as rotation, flipping and scaling, more variant samples are generated to enhance the generalization ability of the model.

[0090] VGG-16 is a deep convolutional neural network model proposed by the Visual Geometry Group (VGG) at the University of Oxford, widely used for its outstanding performance in image recognition tasks. The model consists of 13 convolutional layers and 3 fully connected layers, totaling 16 layers deep. Convolutional layers use 3x3 convolution kernels with a stride of 1 and padding of 1, ensuring that the spatial dimensions of the input and output remain unchanged, and are usually followed by a ReLU activation function. The pooling layer after the convolutional layer uses 2x2 max pooling with a stride of 2, reducing the spatial dimension and reducing the number of parameters and computational complexity. The fully connected layer integrates the features extracted by the convolutional layer and outputs the classification result. Although the structure is simple, due to its depth and approximately 138 million parameters, VGG-16 performs well on large image datasets. The model is commonly used for image classification, object detection, image segmentation, and other tasks, requiring a large amount of computational resources for training. Its advantages lie in high performance and simple structure, but also because the model is large, it requires high computational resources. VGG-16 can process image data captured by cameras in a steel structure positioning and real-time adjustment system, generate corresponding correction instructions by analyzing real-time position and angle, and improve system automation and accuracy.

[0091] In step S4 of the embodiment, the real-time pose difference of the four corner points is obtained by subtracting the accurate poses of the four corner points from the real-time poses.

[0092] In step S5 of the embodiment, during the installation of the steel structure corridor 4, real-time camera image data is input and processed. By combining the trained VGG-16 model with the real-time poses of the four corner points of the steel structure corridor 4, the real-time offset angle and real-time pose error of the current section of the steel structure corridor 4 are predicted. According to the real-time offset angle and real-time pose error, the direction and amplitude of the adjustment required are determined, and the steel structure corridor 4 is adjusted in real time. Combined with the proportional, integral, and derivative parts of the PID controller, accurate correction instructions are generated to control the hydraulic cylinder to perform correction operations and adjust the pose and angle of the steel structure corridor 4. After each correction operation, image data is reacquired, and the analysis and correction steps are repeated to form a closed-loop control mechanism.

[0093] A PID controller consists of three parts: proportional (P), integral (I), and derivative (D). The proportional part quickly adjusts the system output based on the size of the error, but may cause the system to have a certain steady-state error. The integral part eliminates steady-state error by accumulating error to ensure that the system eventually reaches the target value, but may slow down the system response. The derivative part makes an early correction by predicting the trend of error change to reduce oscillation and overshoot, but is sensitive to noise. By adjusting the weights of the three parts reasonably, accurate control of the system can be achieved.

[0094] In this embodiment, multi-segment linkage correction: through the cooperative work of multiple cameras, the relative position and angle information of each segment of the steel structure corridor 4 are captured. Real-time rectangular structure form recognition is performed using a deep learning model, and the real-time offset angle of the steel structure corridor 4 is obtained based on the real-time rectangular structure form and the preset rectangular structure form, and the overall adjustment is performed through the adjustment system to ensure that all segments are on the same straight line.

[0095] In this embodiment, an integrated system is also constructed, which includes a preset module, an image processing module, a shooting module 1, a sensor module 2 and an adjustment module 3. Through multiple tests and calibrations, the cooperative work of each module is optimized, and the system operation efficiency and positioning accuracy are improved. The robustness and anti-interference ability of the system are enhanced to adapt to complex construction environments. In addition, the data is exported and analyzed in the later stage, the displacement data of the virtual model and the actually recorded displacement data are compared, the errors and deviations in the hoisting process are analyzed, and a report containing displacement data, error analysis and adjustment suggestions is generated to ensure that the installation process meets the design requirements.

[0096] In summary, the present application has the following advantages:

[0097] (1) The present application adopts computer vision technology, deep learning algorithm, PID controller, high-precision sensor (IMU and GPS) and virtual reality technology (3ds Max), solves the problem of accurate positioning and adjustment in the process of steel structure installation, ensures the accurate installation of steel structure and the consistency of overall structure. Realize the automatic and high-precision positioning and correction of steel structure, significantly improve the installation efficiency and quality, and reduce the error and risk of manual operation.

[0098] (2) The key features of the steel structure are extracted by using edge detection algorithm. SIFT and SURF algorithms are used to further extract key feature points, ensuring high precision of image feature extraction. Through these technical means, the edge and form features of the steel structure can be accurately recognized and positioned, providing high-quality data for subsequent deep learning model training.

[0099] (3) The present application collects a large amount of image data in the process of steel structure installation and carries out artificial annotation to form training data set. VGG-16 convolutional neural network architecture is selected, and more variant samples are generated through data enhancement techniques (such as rotation, flipping and scaling) to enhance the generalization ability of the model. The trained model can accurately predict the position, angle and deviation of the steel structure in real-time image analysis.

[0100] (4) Combined with the proportional, integral and differential parts of the PID controller, accurate correction instructions are generated to control high-precision servo motors or hydraulic systems to perform correction operations. Through the feedback of real-time image data, repeated analysis and correction steps are formed to form a closed-loop control mechanism to ensure the accurate positioning and adjustment of the steel structure.

[0101] (5) The application installs multiple high-definition cameras on the steel structure corridor 4 to capture the relative position and angle information of each segment of steel structure. Through the cooperative work of multiple cameras and the calculation of deep learning model, the deviation of each segment of steel structure can be adjusted as a whole to ensure that all segments are on the same straight line, realizing high-precision laying of continuous steel structure.

[0102] (6) A virtual body is established on the 3ds Max corridor model to track the displacement change of the corner point, record the displacement data of the virtual body during the whole hoisting process, and monitor the information of IMU (Inertial Measurement Unit) and GPS (Global Positioning System) sensors in the actual installation process synchronously to ensure that the coordinate system of the recording module is consistent with that in 3ds Max. During hoisting, the displacement data of the corner point is monitored in real time and compared with the preset data in 3ds Max to adjust the hoisting position and ensure that the final position is aligned with the design coordinates.

[0103] (7) An integrated system including preset module, image processing module, shooting module 1, sensor module 2 and adjustment module 3 is constructed. Through multiple tests and calibration, the cooperative work of each module is optimized to improve the system operation efficiency and positioning accuracy. The robustness and anti-interference ability of the system are enhanced to ensure its stable operation in complex construction environment.

[0104] Through the above innovations, the application can provide a high-precision and high-efficiency steel structure positioning and correction system, effectively improving the precision and stability of steel structure corridor 4 installation, reducing the error of manual operation, improving the construction efficiency and quality, adapting to complex construction environment and reducing the cost. The technology has wide application prospect and huge market demand in the fields of building engineering, mechanical manufacturing, intelligent robots, etc.

[0105] Obviously, those skilled in the art can make various modifications and variations to the application without departing from the spirit and scope of the application. Therefore, if these modifications and variations of the application fall within the scope of the claims of the application and their equivalent technologies, the application also intends to include these modifications and variations.

Claims

1. A steel structure positioning system based on computer vision and AI, characterized in that, include: The preset module is used to set the preset rectangular structural shape of the steel structure corridor and the accurate pose of the four corner points corresponding to the preset rectangular structural shape. The camera module is used to capture the first image of the steel structure corridor in real time. The image processing module is used to preprocess the first image to obtain the second image, identify the real-time rectangular structure shape of the steel structure corridor based on the second image, and obtain the real-time offset angle of the steel structure corridor based on the real-time rectangular structure shape and the preset rectangular structure shape. The sensor module is used to acquire the real-time pose of the four corner points of the steel structure corridor. Based on the accurate pose and real-time pose of the four corner points, the real-time pose difference of the four corner points is acquired. The adjustment module is used to adjust the steel structure corridor in real time until the real-time offset angle is less than the preset offset angle threshold and the real-time pose difference of the four corner points is less than the preset pose difference threshold. The adjustment module includes a magnetic chuck, a bracket, a handle, and a pusher. The magnetic chuck is used to attract the steel structure corridor. The bracket is rotatably connected to the magnetic chuck. The pusher includes a pusher base and a pusher shaft. One side of the bracket is fixed to the handle, and the other side is fixed to the pusher base. The handle is used to hook the steel structure corridor, and the pusher shaft is fixed to the steel structure support frame. The handle includes a first U-shaped arm and a second U-shaped arm. The first U-shaped arm is fixed to the bracket. The first U-shaped arm includes a first end and a second end. The second U-shaped arm includes a third end and a fourth end. The first end and the third end are connected by a first straight rod. The second end and the fourth end are connected by a second straight rod. The angle between the first straight rod and the first end is 110-150 degrees. The angle between the third end and the first straight rod is 110-150 degrees. The angle between the second straight rod and the second end is 110-150 degrees. The angle between the fourth end and the second straight rod is 110-150 degrees.

2. The steel structure positioning system based on computer vision and AI according to claim 1, characterized in that: The pose includes attitude information and position information. The sensor module includes an inertial measurement unit and a global positioning system. The inertial measurement unit is used to measure the attitude information of the four corner points of the steel structure corridor when they move. The attitude information includes angular velocity and acceleration information. The global positioning system is used to provide the position information of the four corner points of the steel structure corridor.

3. A steel structure positioning method based on computer vision and AI, applied to the steel structure positioning system based on computer vision and AI as described in any one of claims 1-2, characterized in that, Includes the following steps: The preset rectangular structural form of the steel structure corridor and the accurate pose of the four corner points corresponding to the preset rectangular structural form are set. The first image of the steel structure corridor was captured in real time. The first image is preprocessed to obtain the second image. The real-time rectangular structure shape of the steel structure corridor is identified based on the second image. The real-time offset angle of the steel structure corridor is obtained based on the real-time rectangular structure shape and the preset rectangular structure shape. The real-time poses of the four corner points of the steel structure corridor are obtained. Based on the accurate poses and real-time poses of the four corner points, the real-time pose difference of the four corner points is obtained. The steel structure corridor is adjusted in real time until the real-time offset angle is less than the preset offset angle threshold and the real-time pose difference of the four corner points is less than the preset pose difference threshold.

4. The steel structure positioning method based on computer vision and AI according to claim 3, characterized in that, The method for preprocessing the first image to obtain the second image is as follows: Grayscale transformation and median filtering for noise reduction.

5. The steel structure positioning method based on computer vision and AI according to claim 3, characterized in that, The method for obtaining the real-time offset angle of the steel structure corridor based on the real-time rectangular structure shape and the preset rectangular structure shape is as follows: The edges of the steel structure corridor are identified using the Canny edge detection algorithm, and the structural morphological features of the edges of the steel structure corridor are extracted using morphological operations, including expansion and corrosion. The SIFT and SURF algorithms are used to extract key feature points of the structural morphology. The key feature points are the four corner points of the steel structure corridor. Based on the extracted four corner points of the steel structure corridor, a real-time rectangular structural morphology is established. The four sides of the steel structure corridor are obtained based on the preset rectangular structure shape. The four sides of the steel structure corridor are obtained based on the real-time rectangular structure shape. The angles between the preset four sides and the actual four sides that correspond one-to-one are averaged to obtain the real-time offset angle of the steel structure corridor obtained from the real-time rectangular structure shape and the preset rectangular structure shape.

6. The steel structure positioning method based on computer vision and AI according to claim 5, characterized in that: This also includes building deep learning models, the specific methods of which are as follows: Collect a certain amount of image data of the steel structure corridor installation process to create an image set; The image set was manually annotated to mark the steel structure corridor; The model was trained using a VGG-16 convolutional neural network. The trained convolutional neural network was used for real-time rectangular structural morphology recognition of the steel structure corridor. The real-time offset angle of the steel structure corridor is obtained based on the real-time rectangular structure shape and the preset rectangular structure shape.

7. The steel structure positioning method based on computer vision and AI according to claim 5, characterized in that: Based on the real-time offset angle and real-time pose error, the direction and magnitude of adjustment required are determined. Combining the proportional, integral, and derivative parts of the PID controller, precise correction commands are generated to adjust the steel structure corridor in real time.

8. The steel structure positioning method based on computer vision and AI according to claim 3, characterized in that: A virtual model of the steel structure corridor is created in 3ds Max. Virtual bodies are added to each corner of the steel structure corridor in the virtual model. The real-time pose of the four corners of the steel structure corridor is obtained by tracking the displacement changes of the four corners one by one through the four virtual bodies.

Images