Building engineering measurement system based on segmentation and clustering

Abstract

The invention belongs to the technical field of constructional engineering measurement, and discloses a constructional engineering measurement system based on segmentation and clustering, which comprises a multi-source data acquisition module, a data processing and analysis module and a control and display module. The multi-source data acquisition module is composed of a three-dimensional laser radar, unmanned aerial vehicle oblique photography equipment, a total station and GPS positioning equipment, and is used for acquiring and fusing multi-source data. And the data processing and analysis module realizes automatic and accurate identification and measurement of building components by adopting a process of data preprocessing, point cloud segmentation based on deep learning and combination of DBSCAN clustering and energy function optimization. And the control and display module is used for system control and result three-dimensional visualization. Through combination of multi-source data fusion and an intelligent algorithm, the problems of low efficiency, insufficient precision and poor automation degree of a traditional measurement method are solved, and the intelligent level and the operation efficiency of constructional engineering measurement are improved.

Description

Technical Field

[0001] This invention relates to the field of building engineering surveying technology, and in particular to a building engineering surveying system based on segmentation clustering. Background Technology

[0002] In the field of construction engineering, surveying work spans the entire project lifecycle, including planning and design, construction, and operation and maintenance. Accurate surveying data is a crucial foundation for ensuring project quality, construction safety, and smooth project progress. Traditional construction engineering surveying methods, such as total station surveying and GPS surveying, have many limitations when dealing with complex building structures and diverse construction environments. Traditional surveying methods are inefficient, making it difficult to quickly obtain comprehensive data from large-scale construction sites; measurement accuracy is greatly affected by environmental factors, with significant errors occurring under complex terrain and inclement weather conditions; and for some complex building components and structures, traditional methods struggle to achieve accurate measurement and analysis. With the rapid development of the construction industry and the increasing complexity of building structures, higher demands are placed on the accuracy, efficiency, and intelligence of surveying technology. Therefore, a new construction engineering surveying system is urgently needed to meet the needs of modern construction engineering surveying. Summary of the Invention

[0003] The present invention aims to provide a construction engineering measurement system based on segmentation and clustering. By integrating advanced multi-source data acquisition technology and segmentation and clustering technology that combines deep learning with traditional algorithms, it achieves efficient, accurate, and intelligent measurement and monitoring of construction projects, thereby solving the problems mentioned in the background technology and ensuring the smooth progress of construction projects.

[0004] To achieve the above objectives, the present invention provides the following technical solution: A construction engineering surveying system based on segmentation and clustering includes a multi-source data acquisition module, a data processing and analysis module, and a control and display module. The multi-source data acquisition module includes a 3D LiDAR, a UAV oblique photogrammetry device, a total station, and a GPS positioning device. The 3D LiDAR is used to acquire 3D point cloud data from the site. The UAV oblique photogrammetry device is used to acquire oblique photogrammetric images of the site and generate a 3D model with realistic texture. The total station and GPS positioning device are used to acquire precise positioning data of the measurement target for coordinate calibration. The data processing and analysis module is communicatively connected to the multi-source data acquisition module. The data processing and analysis module includes a data preprocessing unit, a deep learning segmentation unit, and a clustering analysis and optimization unit. The data preprocessing unit performs denoising, filtering, and smoothing on the raw point cloud data. The clustering analysis and optimization unit employs a clustering algorithm and incorporates an optimization algorithm. The control and display module is connected to both the multi-source data acquisition module and the data processing and analysis module. The control and display module controls the operating parameters of the multi-source data acquisition module and displays the processing results of the data processing and analysis module.

[0005] Preferably, the operations performed by the data preprocessing unit include outlier denoising based on statistical analysis, downsampling based on voxel grid filtering, and data smoothing based on curvature filtering.

[0006] Preferably, the deep learning segmentation unit adopts a point cloud segmentation model based on convolutional neural networks and recurrent neural networks.

[0007] Preferably, the point cloud segmentation model adopts the PointNet++ network structure, and the point cloud segmentation model is obtained by supervised learning training through point cloud data samples of building engineering with labeled information.

[0008] Preferably, the clustering algorithm used in the clustering analysis and optimization unit is the density-based DBSCAN algorithm.

[0009] Preferably, the optimization algorithm introduced in the clustering analysis and optimization unit is an energy function-based optimization algorithm, which is used to correct the clustering results according to the physical constraints of the building structure.

[0010] Preferably, the data processing and analysis module has real-time monitoring and intelligent early warning functions.

[0011] Preferably, the control and display module includes a three-dimensional model display unit, which is used to display a three-dimensional model with measurement results of building components and structural deformation information.

[0012] Preferably, the control and display module further includes a construction progress simulation unit, which is used to perform a visual simulation of the construction progress based on a three-dimensional model.

[0013] Preferably, the drone tilt photography device is equipped with a multi-lens camera.

[0014] The beneficial effects of this technical solution compared to existing technologies are as follows: (1) This technical solution integrates multiple devices such as 3D LiDAR and UAV oblique photography equipment by setting up a multi-source data acquisition module, and fuses multi-source data to accurately acquire geometric, texture and positioning information of the construction site, providing a comprehensive and accurate data foundation for measurement; the data processing and analysis module uses data preprocessing, deep learning segmentation and clustering optimization algorithms to effectively remove data noise, accurately identify and segment building components and structures, and improve the accuracy and efficiency of data processing; the control and display module realizes flexible control of the acquisition work, and displays the measurement results and simulates the construction progress with an intuitive 3D model. Compared with traditional measurement methods, it improves measurement accuracy and efficiency, can cope with complex and ever-changing construction environments, enhances project management efficiency, provides reliable and efficient technical support for the full life cycle management of construction projects, and promotes the intelligent development of construction engineering measurement technology.

[0015] (2) Deep learning algorithms are organically combined with traditional clustering algorithms in the data processing and analysis module. The deep learning point cloud segmentation model automatically learns the complex features of the data to achieve accurate segmentation; the DBSCAN clustering algorithm and the energy function-based optimization algorithm further optimize the analysis results. This not only leverages the powerful feature extraction capabilities of deep learning but also utilizes the stability and interpretability of traditional algorithms. It can cope with the complex and ever-changing data and scenarios in building engineering surveying, improve the accuracy and efficiency of data processing, and quickly and accurately identify and analyze various building structures.

[0016] (3) By setting up a data processing and analysis module with real-time monitoring and intelligent early warning functions, data on key parts of the building structure are collected in real time. The data is analyzed in real time using a segmentation and clustering algorithm. Once the structural deformation exceeds the preset threshold, an intelligent early warning is issued immediately to remind relevant personnel to take measures. At the same time, through comprehensive analysis of historical data and real-time monitoring data, the system can also predict the deformation trend of the building structure, discover potential safety hazards in advance, provide a scientific basis and strong guarantee for engineering safety management, avoid safety accidents caused by structural deformation, reduce engineering risks, and ensure the safety of personnel and property and the smooth progress of the project.

[0017] (4) By setting up a 3D model display unit, the measurement results of building components and structural deformation information are presented in an intuitive 3D model form. Managers can view the project information from all angles and in all directions through interactive operation, making the originally abstract data intuitive and easy to understand. The construction progress simulation unit is based on the 3D model and combines the construction plan and actual progress data to perform a visual simulation of the construction progress, which can clearly show the various links and progress of the construction process. This visualization display and construction simulation function helps managers to grasp the project progress more intuitively, discover potential problems and conflicts in the construction process in advance, such as improper connection of procedures and unreasonable allocation of resources, so as to optimize the construction plan in a timely manner, rationally allocate resources, improve construction management efficiency, and ensure that the project is completed on time and with high quality. Attached Figure Description

[0018] Figure 1 This is a system architecture diagram of the present invention; Detailed Implementation The present invention will now be described in further detail with reference to the accompanying drawings and embodiments: like Figure 1 The illustrated building engineering surveying system based on segmentation and clustering includes a multi-source data acquisition module, a data processing and analysis module, and a control and display module. The multi-source data acquisition module integrates a 3D LiDAR, a UAV oblique photogrammetry device, a total station, and a GPS positioning device. The 3D LiDAR utilizes the principle of laser ranging, rapidly acquiring 3D point cloud data of the building site by emitting a laser beam and receiving the time delay of the reflected light. This allows for precise depiction of the building structure's outline, surface details, and spatial relationships, providing rich geometric information for subsequent data analysis. The UAV oblique photogrammetry device, equipped with a multi-lens camera, captures images of the building area from multiple angles. Through image matching and 3D reconstruction techniques, it generates a 3D model with realistic textures, effectively compensating for the LiDAR's limitations in acquiring surface texture information. This system is suitable for measuring the overall layout of the building, its surrounding environment, and complex building exteriors. The total station and GPS positioning device are responsible for acquiring precise positioning data of the measurement target, calibrating the coordinates of the measurement data, ensuring the accuracy and consistency of all measurement data within a unified coordinate system, and improving the reliability of the measurement data.

[0019] The data processing and analysis module, which communicates with the multi-source data acquisition module to achieve data interaction, is the core processing unit of the system. It includes a data preprocessing unit, a deep learning segmentation unit, and a clustering analysis and optimization unit. The data preprocessing unit addresses noise and redundancy issues in the raw point cloud data by performing outlier denoising based on statistical analysis, identifying and removing noise points that do not conform to the data distribution pattern. It employs a voxel grid filtering-based downsampling method to reduce data volume and improve processing efficiency while preserving key data features. Curvature filtering-based data smoothing technology is used to smooth the data, making the point cloud data more regular and laying the foundation for subsequent processing. The deep learning segmentation unit constructs a point cloud segmentation model based on convolutional neural networks and recurrent neural networks, preferably using the PointNet++ network structure, which effectively handles the disorder and irregularity of point cloud data. Supervised learning training is conducted using a large number of labeled architectural engineering point cloud data samples, enabling the model to accurately identify and segment different types of building components, topography, and ancillary facilities, achieving preliminary classification of the point cloud data. The clustering analysis and optimization unit employs the density-based DBSCAN algorithm. Based on the density distribution characteristics of the point cloud data, it clusters points with similar densities and spatial relationships into one class, further identifying different building structural units or objects. Simultaneously, an energy function-based optimization algorithm is introduced to correct and optimize the clustering results according to the physical constraints of the building structure, such as component size limitations and connection relationships, making the analysis results more consistent with the actual structure and physical characteristics of the building project. The data processing and analysis module also features real-time monitoring and intelligent early warning functions, enabling continuous analysis of structural deformation and timely warnings.

[0020] The control and display module is connected to the multi-source data acquisition module and the data processing and analysis module, respectively, to control the entire measurement system and display the results. In terms of control, operators can use this module to set the operating parameters of the multi-source data acquisition module, such as adjusting the scanning range and frequency of the 3D LiDAR, setting the flight path and shooting parameters of the UAV, and flexibly controlling the data acquisition process. In terms of display, the control and display module has a 3D model display unit, which can present the processing results of the data processing and analysis module in an intuitive 3D model format, clearly marking the measurement results of building components, structural deformation information, etc., on the 3D model. In addition, a construction progress simulation unit is also included. Based on the 3D model, combined with the construction plan and actual progress data, it performs a visual simulation of the construction progress, helping managers to intuitively understand the construction progress, identify potential problems in advance, optimize construction plans, and improve construction management efficiency.

[0021] The system comprises a data processing and analysis module and a control and display module. The multi-source data acquisition module includes a 3D LiDAR, a UAV oblique photogrammetry device, a total station, and a GPS positioning device. The 3D LiDAR acquires 3D point cloud data from the site, the UAV oblique photogrammetry device acquires oblique photogrammetric images and generates a 3D model with realistic textures, and the total station and GPS positioning device acquire precise positioning data for the measurement target to achieve coordinate calibration. The data processing and analysis module communicates with the multi-source data acquisition module and includes a data preprocessing unit, a deep learning segmentation unit, and a clustering analysis and optimization unit. The data preprocessing unit performs denoising, filtering, and smoothing on the raw point cloud data, while the clustering analysis and optimization unit employs a clustering algorithm and incorporates an optimization algorithm. The control and display module connects to both the multi-source data acquisition module and the data processing and analysis module, controlling the operating parameters of the multi-source data acquisition module and displaying the processing results of the data processing and analysis module.

[0022] The specific implementation process is as follows: During use, the operating parameters of the multi-source data acquisition module are set through the control and display module according to the measurement task. Then, data acquisition is initiated. A 3D LiDAR scans the construction site to acquire dense 3D point cloud data; a UAV simultaneously performs oblique photography to acquire high-resolution images; and a total station and GPS perform precise coordinate measurements of key control points. All data is transmitted to the data processing and analysis module via wireless or wired connections.

[0023] In the data processing and analysis module, the data preprocessing unit first processes the raw laser point cloud data: statistical analysis is used to remove obvious outlier noise points; voxel grid filtering is used to downsample the massive point cloud, reducing computational load while preserving shape features; curvature filtering is applied to smooth the point cloud, improving data quality. The preprocessed point cloud data is then fed into the deep learning segmentation unit. This unit uses a pre-trained PointNet++ model. This model can automatically identify and classify each point in the point cloud into a predefined category, completing semantic segmentation. The semantic segmentation results are then fed into the clustering analysis and optimization unit. This unit processes each type of point cloud separately. The DBSCAN clustering algorithm is used to segment point clouds of the same category into different independent clusters based on the spatial density distribution of points, with each cluster representing an independent object instance. Subsequently, an optimization algorithm based on an energy function, which defines geometric constraints, is introduced. Through iterative optimization, the algorithm ensures that the clustered model satisfies these physical constraints as much as possible, thereby correcting potential erroneous segmentations and obtaining a more accurate and realistic component model.

[0024] Finally, all processing results are sent to the control and display module. The 3D model display unit overlays the results onto the 3D model for visualization, allowing users to interactively query detailed information for any component. The construction progress simulation unit dynamically demonstrates changes in construction progress based on data from different periods. Simultaneously, the system continues to run, and if the real-time monitoring function detects that the deformation of certain key indicators exceeds a preset threshold, it immediately triggers an intelligent warning, notifying management personnel.

[0025] The above descriptions are merely embodiments of the present invention, and common knowledge such as specific technical solutions and / or characteristics are not described in detail here. It should be noted that those skilled in the art can make various modifications and improvements without departing from the technical solutions of the present invention, and these should also be considered within the scope of protection of the present invention. These modifications and improvements will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A building engineering surveying system based on segmentation and clustering, characterized in that: The system comprises a multi-source data acquisition module, a data processing and analysis module, and a control and display module. The multi-source data acquisition module includes a 3D LiDAR, a UAV oblique photogrammetry device, a total station, and a GPS positioning device. The 3D LiDAR is used to acquire 3D point cloud data from the site. The UAV oblique photogrammetry device is used to acquire oblique photogrammetric images from the site and generate a 3D model with realistic textures. The total station and GPS positioning device are used to acquire precise positioning data of the measurement target for coordinate calibration. The data processing and analysis module is communicatively connected to the multi-source data acquisition module. This module includes a data preprocessing unit, a deep learning segmentation unit, and a clustering analysis and optimization unit. The data preprocessing unit performs denoising, filtering, and smoothing on the raw point cloud data. The clustering analysis and optimization unit employs a clustering algorithm and incorporates an optimization algorithm. The control and display module is connected to both the multi-source data acquisition module and the data processing and analysis module. This module controls the operating parameters of the multi-source data acquisition module and displays the processing results of the data processing and analysis module.

2. The building engineering surveying system based on segmentation clustering as described in claim 1, characterized in that: The data preprocessing unit performs operations including outlier denoising based on statistical analysis, downsampling based on voxel grid filtering, and data smoothing based on curvature filtering.

3. The building engineering surveying system based on segmentation clustering as described in claim 1, characterized in that: The deep learning segmentation unit employs a point cloud segmentation model built upon convolutional neural networks and recurrent neural networks.

4. The building engineering surveying system based on segmentation clustering as described in claim 3, characterized in that: The point cloud segmentation model adopts the PointNet++ network structure and is trained through supervised learning using point cloud data samples of building engineering with labeled information.

5. A building engineering surveying system based on segmentation clustering as described in claim 1, characterized in that: The clustering algorithm used in the clustering analysis and optimization unit is the density-based DBSCAN algorithm.

6. The building engineering surveying system based on segmentation clustering as described in claim 1, characterized in that: The optimization algorithm introduced in the clustering analysis and optimization unit is an energy function-based optimization algorithm, which is used to correct the clustering results according to the physical constraints of the building structure.

7. The building engineering surveying system based on segmentation clustering as described in claim 1, characterized in that: The data processing and analysis module has real-time monitoring and intelligent early warning functions.

8. A building engineering surveying system based on segmentation clustering as described in claim 1, characterized in that: The control and display module includes a three-dimensional model display unit, which is used to display a three-dimensional model with measurement results of building components and structural deformation information.

9. A building engineering surveying system based on segmentation clustering as described in claim 8, characterized in that: The control and display module also includes a construction progress simulation unit, which is used to perform visual simulation of construction progress based on a three-dimensional model.

10. A building engineering surveying system based on segmentation clustering as described in claim 1, characterized in that: The drone tilt photography equipment is equipped with a multi-lens camera.

Images