Building component lightweight detection method and device based on OneFormer3D and storage medium

Abstract

The invention discloses a building component lightweight detection method and device based on OneFormer3D and a storage medium. A ground three-dimensional laser scanner is used for collecting three-dimensional point cloud data in a building. And performing denoising processing on the point cloud data in the building, and performing point cloud labeling to complete data preparation. And adopting a deep learning model based on OneFomer3D network structure optimization to obtain instance information of the building component. The building component with instance information is detected in a machine learning mode and compared with related specifications, and light-weight detection is achieved. Through the deep learning method based on the point cloud data, the detection task of the building component can be rapidly and accurately completed. Meanwhile, the method has the advantages of being high in automation degree, low in cost and the like, is suitable for quality control and maintenance of large-scale construction projects, and has wide application prospects.

Description

Technical Field

[0001] The present invention relates to a OneFormer3D-based building component lightweight detection method, equipment and storage medium, and belongs to the technical field of civil engineering and artificial intelligence interaction. Background Art

[0002] At present, the detection of construction dimension deviation in engineering projects mostly relies on manual measurement and uses traditional tools such as steel tape measure, ruler, feeler gauge, theodolite and level. Although automated equipment such as laser rangefinder and laser level have improved measurement efficiency to a certain extent, its point-by-point measurement mode still relies on manual operation, which limits further efficiency improvement.

[0003] In order to improve the efficiency and accuracy of construction quality inspection, 3D laser scanning technology is widely used because it can quickly and accurately obtain the surface shape of objects and convert them into point cloud data. Existing studies have realized intelligent dimensional quality inspection for precast concrete components, steel structures, and cast-in-place concrete components. However, the existing point cloud-based inspection technology still does not meet the needs of engineering applications in terms of effect. At present, although some research teams have made progress in the flatness inspection of precast concrete components and the size and position inspection of steel bars, the inspection accuracy still does not meet the engineering requirements.

[0004] In terms of component recognition algorithms, deep learning algorithms have the advantage of wide applicability. PointNet++, PointConv and other networks have been used in component classification and recognition, but the accuracy rate is generally less than 40%. Although some studies have achieved good detection results, existing studies mainly focus on planar components such as floor slabs and wall panels, and most of them do not consider the complex environment of the construction site. Deep learning algorithms have the problem of uneven classification accuracy, which needs further improvement to improve the actual application effect.

[0005] Therefore, in order to further improve the efficiency and accuracy of construction quality inspection, it is urgent to study more intelligent and efficient construction dimension quality inspection technology. Summary of the invention

[0006] Purpose: In order to overcome the deficiencies in the prior art, the present invention provides a lightweight detection method, device and storage medium for building components based on OneFormer3D, which can realize automatic identification of building components, reduce labor costs and greatly improve engineering efficiency.

[0007] Technical solution: To solve the above technical problems, the technical solution adopted by the present invention is:

[0008] In the first aspect, a lightweight detection method for building components based on OneFormer3D specifically includes:

[0009] Step S1: Acquire point cloud data inside the building.

[0010] Step S2: De-noise and simplify the point cloud data inside the building to obtain pre-processed point cloud data.

[0011] Step S3: Classify the preprocessed point cloud data by region, and annotate the building component types of the point cloud data in each region, so as to obtain the building components in each region and the annotations of each building component as the point cloud data set.

[0012] Step S4: Use the deep learning model based on the Onefomer3D network structure optimization to train the point cloud data set, and perform multiple trainings by setting different training hyperparameter modules to obtain the best training hyperparameter module, and configure the deep learning model with the best training hyperparameter module for training to obtain a trained deep learning model; input the point cloud data to be predicted into the trained deep learning model to obtain the instance information point cloud data inside the building.

[0013] Step S5: Post-process and optimize the instance information point cloud data to obtain optimized instance information point cloud data.

[0014] Step S6: Using a multi-plane fitting method based on random consistency and a directed envelope box algorithm based on principal component analysis on the optimized instance information point cloud data, the size information and spatial position information of the building components corresponding to the optimized instance information point cloud data are obtained.

[0015] Step S7: Compare the size information and position information of the building component with the design specifications to complete the detection of the building component.

[0016] As a preferred solution, the point cloud data inside the building in step S1 meets the following requirements: the data density of the point cloud data contains at least 100 points per square meter, and shall not be less than 50 points per square meter under special circumstances or requirements; the spatial range covers the entire target area, including the area within 30 meters outside the edge wire; the file storage format is the E57 file format, the scanning resolution is 8000 points per meter, and the ranging error is ≤±1mm at 25 meters; the point cloud data contains RGB color information.

[0017] As a preferred solution, the denoising process uses Gaussian filtering, and the simplification uses an Octree tree structure to perform 12-level downsampling simplification on the point cloud data.

[0018] As a preferred solution, the annotated objects of the point cloud data set also include on-site objects.

[0019] As a preferred solution, the deep learning model specifically includes: the output end of the downsampling module is respectively connected to the second input end of the mask module, the second input end of the instance module and the input end of the superpoint extraction module, and the output end of the superpoint extraction module is respectively connected to the first input end of the mask module, the first input end of the decoder and the first input end of the instance module. The output end of the instance module is connected to the second input end of the decoder, the output end of the mask module is connected to the first input end of the prediction head, and the output end of the decoder is connected to the second input end of the prediction head.

[0020] The downsampling module includes a first fully connected layer with a dimension of 6x32, a farthest point sampling layer, a K nearest neighbor layer, a second fully connected layer with a dimension of 32x16, and a maximum pooling layer, which are connected in sequence.

[0021] The farthest point sampling layer is used to iteratively calculate the distance from each point in the current point set to the center of the existing point cloud model, and select the N points with the farthest distance as a new subset.

[0022] The K nearest neighbor layer is used to capture the local structure information of the point set.

[0023] The super-point extraction module is used to voxelize the input point cloud, convert it into a sparse three-dimensional convolution format, and extract point-by-point features.

[0024] The mask module is used to integrate the residual mask feature vector output by the downsampling module and the feature vector output by the superpoint extraction module to perceive the global feature.

[0025] The instance module is used to obtain a query vector with instance features from the downsampling module and the superpoint extraction module.

[0026] The decoder is used to convert a high-dimensional query vector into a low-dimensional feature representation.

[0027] The prediction head is used to convert the output mask feature perception information and instance mask into a final prediction result and evaluate the quality of each prediction.

[0028] As a preferred solution, the training hyperparameter module configuration includes: number of threads, learning rate, training batch size, number of iterations, weight decay, optimizer and learning rate scheduler.

[0029] As a preferred solution, the post-processing optimization adopts a statistical filtering algorithm and a cluster analysis method.

[0030] As a preferred solution, step S6 specifically includes:

[0031] S6-1: Building components are divided into three-dimensional components and planar components according to their point cloud scanning features.

[0032] S6-2: For a three-dimensional component with multiple planes, a multi-plane fitting method based on random consistency is used to fit each plane and calculate the distance between parallel planes to obtain the size information of the three-dimensional component.

[0033] S6-3: For a planar component having only a single plane, the size information of the planar component is obtained by measuring the size and center coordinates of the directed bounding box.

[0034] In a second aspect, a computer-readable storage medium stores a computer program, which, when executed by a processor, implements a lightweight detection method for building components based on OneFormer3D as described in any one of the first aspects.

[0035] According to a third aspect, a computer device includes:

[0036] Memory, used to store instructions.

[0037] The processor is used to execute the instructions so that the computer device performs an operation based on the OneFormer3D building component lightweight detection method as described in any one of the first aspects.

[0038] Beneficial effects: The invention provides a lightweight detection method, device and storage medium for building components based on OneFormer3D. The invention can quickly and accurately complete the detection task of building components through a deep learning method based on point cloud data. At the same time, the method has the advantages of high automation and low cost, and is suitable for quality control and maintenance of large-scale construction projects, and has broad application prospects.

[0039] Compared with the existing technology, its advantages are as follows:

[0040] 1. The present invention supplements the blank of point cloud dataset in the existing building completion field.

[0041] 2. The present invention uses 3D laser scanners and deep learning technologies such as the OneFormer3D model to quickly capture and process point cloud data inside buildings. Compared with current measurement methods, this automated process greatly shortens the time for data collection and analysis, and substantially improves the efficiency of construction dimension detection.

[0042] 3. The present invention enhances the recognition capability of complex structures through precise size detection and efficient data processing, ensuring higher data accuracy and consistency of detection results.

[0043] 4. The present invention helps the construction party and the supervision party to make decisions quickly during the acceptance stage, and also improves the quality of the entire project. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a flow chart of the lightweight detection method of building components based on OneFormer3D.

[0045] Figure 2 This is a schematic diagram of the Faro Focus Premium 150 terrestrial 3D laser scanner used for acquisition in an example.

[0046] Figure 3 The present invention is a schematic diagram of an on-site picture of collecting point clouds of building components using a ground-based three-dimensional laser scanner in an example.

[0047] Figure 4 This is a schematic diagram of a building point cloud on a certain floor after denoising in an example.

[0048] Figure 5 This is a schematic diagram of the preprocessed point cloud after annotation in an example. Figure 5 The middle picture (a) is an unlabeled point cloud. Figure 5 The middle figure (b) is the annotated point cloud.

[0049] Figure 6 This is a network diagram of a deep learning model based on Onefomer3D network structure optimization.

[0050] Figure 7 It is a schematic diagram of instance segmentation of point cloud area under an instance. Figure 7 The middle picture (a) is the point cloud before segmentation. Figure 7 The middle figure (b) is the point cloud after instance segmentation.

[0051] Figure 8 This is a schematic diagram of model hyperparameter configuration under an embodiment.

[0052] Fig. 9 It is a schematic diagram of the fitting plane of the beams and columns in the point cloud area after multi-plane fitting in an example.

[0053] Fig.10 This is a schematic diagram of the ceiling, floor, and wall of the point cloud area after being processed by a directed envelope box in an example. DETAILED DESCRIPTION

[0054] The following is a clear and complete description of the technical solutions in the examples of the present invention in conjunction with the accompanying drawings in the examples of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the protection scope of the present invention.

[0055] The present invention will be further described below in conjunction with specific embodiments.

[0056] Embodiment 1:

[0057] This embodiment introduces a lightweight detection method for building components based on OneFormer3D, which relies on a deep learning neural network framework and point cloud processing technology to achieve lightweight detection of building components, and specifically includes the following steps:

[0058] Step S1: Use a ground-based 3D laser scanner to collect point cloud data inside the building.

[0059] Step S2: De-noise and simplify the point cloud data inside the building to obtain pre-processed point cloud data.

[0060] Step S3: Classify the preprocessed point cloud data by region, and annotate the building component types of the point cloud data in each region, obtain the building components in each region and the annotations of each building component, and use them as a point cloud data set. The building components include but are not limited to walls, beams, columns, plates, doors or windows.

[0061] Step S4: Build a deep learning model based on Onefomer3D network structure optimization, train the point cloud data set obtained in step S3, perform multiple trainings by setting different training hyperparameter modules, obtain the best training hyperparameter module, configure the deep learning model with the best training hyperparameter module for training, obtain a trained deep learning model, and save the best training hyperparameters; input the point cloud data to be predicted into the trained deep learning model to obtain the instance information point cloud data inside the building.

[0062] Step S5: Post-processing and optimizing the instance information point cloud data of step S5 to obtain optimized instance information point cloud data, thereby further improving data quality.

[0063] Step S6: For the optimized instance information point cloud data in step S5, a multi-plane fitting method based on random consensus (RANSAC) and an oriented bounding box (OBB) algorithm based on principal component analysis are used to obtain the size information and spatial position information of the building components corresponding to the optimized instance information point cloud data.

[0064] Step S7: By combining the size information and position information of the building components obtained in step S6 and comparing with the design specifications, the lightweight detection of the building components can be completed.

[0065] Furthermore, the point cloud data collected in step S1 must ensure that the data density contains at least 100 points per square meter, and shall not be less than 50 points per square meter in special environments or under demand conditions; the spatial range must cover the entire target area, including the 80-meter range outside the edge line; the file storage format is the E57 file format, the scanning resolution is 8000 points per meter (Pt / m), and the ranging error is ≤±1mm at 25 meters. The point cloud data needs to contain RGB color information to provide richer visual details and the possibility of subsequent analysis.

[0066] Furthermore, in step S2, Gaussian filtering is used to effectively remove noise. After denoising, the point cloud data is simplified by 12-level downsampling using an Octree tree structure, aiming to reduce the amount of data while keeping its structural features unchanged.

[0067] Furthermore, the point cloud dataset produced in step S3 is divided into three parts: training set, validation set, and test set in a ratio of 8:1:1 or 7:2:1. The annotated objects are building components and on-site items, such as stacked steel bars, masonry and other building materials, discarded construction waste, construction tools, etc. The point cloud dataset is saved in the format of the S3DIS instance segmentation dataset.

[0068] Furthermore, the number of point clouds in each region in step S3 does not exceed five million points.

[0069] Furthermore, the deep learning model based on Onefomer3D network structure optimization described in step S4 uses sparse 3D convolution as the backbone, uses superpoint pooling, does not use query selection, and uses a separate instance query. The backbone is pre-trained on the S3DIS dataset, the Structured3D dataset, and the ScanNet dataset.

[0070] Furthermore, the deep learning model optimization part in step S4 includes the introduction of a downsampling module and a residual connection mechanism. These improvements not only reduce the computational overhead, but also ensure the information integrity of the deep neural network, effectively prevent training anomalies caused by the vanishing gradient phenomenon, and improve training efficiency and accuracy.

[0071] The working principle of the entire network structure is described as follows:

[0072] Input the Nx6 point cloud dataset into the downsampling module.

[0073] Among them, N represents the number of input point cloud data, 6 represents the dimension of point cloud data, including: x, y, z, R, G, B. Among them, x represents the value of the x-axis in the spatial coordinate system, y represents the value of the y-axis in the spatial coordinate system, z represents the value of the z-axis in the spatial coordinate system, R represents the red value of the point cloud data, G represents the green value of the point cloud data, and B represents the blue value of the point cloud data.

[0074] The downsampling module includes a first fully connected layer with a dimension of 6x32, a farthest point sampling layer, a K nearest neighbor layer, a second fully connected layer with a dimension of 32x16, and a maximum pooling layer, which are connected in sequence.

[0075] The farthest point sampling layer is used to iteratively calculate the distance from each point in the current point set to the center of the existing point cloud model, and select the N points with the farthest distance as a new subset.

[0076] The K nearest neighbor layer is used to capture the local structural information of the point set, thereby enhancing the expressiveness and robustness of the features.

[0077] The downsampling module extracts deep features through two fully connected layers, selects representative subsets using farthest point sampling, reduces the number of points while retaining global spatial information. Subsequently, local features are aggregated using K nearest neighbors and then mapped to a low-dimensional space through a fully connected layer to reduce computational overhead and keep the context point cloud information intact.

[0078] The output end of the downsampling module is connected to the second input end of the mask module, the second input end of the instance module and the input end of the super point extraction module respectively, and the output end of the super point extraction module is connected to the first input end of the mask module, the first input end of the decoder and the first input end of the instance module respectively. The output end of the instance module is connected to the second input end of the decoder, the output end of the mask module is connected to the first input end of the prediction head, and the output end of the decoder is connected to the second input end of the prediction head.

[0079] The superpoint extraction module is used to voxelize the input point cloud, convert it into a sparse 3D convolutional format, and extract point-by-point features. The module consists of a U-net-like backbone composed of sparse 3D convolutions and an average pooling layer. After processing by this module, the output feature matrix undergoes different linear transformations to generate three key vectors: query vector (Query), key vector (Key) and value vector (Value). By converting the input point cloud of millions of points into only hundreds of superpoints, the computational cost of subsequent processing is significantly reduced.

[0080] The mask module is used to integrate the residual mask feature vector output by the downsampling module and the feature vector output by the superpoint extraction module to perceive the global features. This module uses a multi-layer perceptron (MLP) consisting of four fully connected layers to process the outputs of the two modules. In this way, the generated mask features can efficiently fuse and transform the input features, generate richer and more detailed mask feature perception information, and have better generalization performance.

[0081] The instance module is used to obtain the query vector with instance features from the downsampling module and the superpoint extraction module. The module is composed of multiple linear transformation matrices, position encoding matrices and nonlinear activation functions. The feature vectors from different modules are multiplied by the corresponding matrices to obtain the query vector; the instance query vector obtained after processing dynamically calculates and integrates the relevant information from the instance residual, so that the model can more accurately capture the complex structure and pattern in the point cloud data. During the training process, if only the query vector from the superpoint is used, some feature information may be lost. The introduction of instance residuals can also ensure the integrity of the information and effectively prevent the gradient disappearance phenomenon.

[0082] The decoder is used to convert the high-dimensional query vector into a low-dimensional feature representation. This module consists of six sequential converter decoder layers, which accept the query vector from the instance module and the keys and values ​​from the super-point features for cross-attention input and convert them into feature convolution kernels. These kernels are then convolved with the super-point features, and instance masks and residual masks are generated through the convolution operation. The instance mask is used to predict different instances in the point cloud, and the residual mask is used to capture the difference between the predicted result and the actual target. By analyzing the difference between the predicted mask and the true mask, the model can adjust its parameters more finely to better fit the training data, which helps to accelerate convergence and improve the generalization ability of the model.

[0083] The prediction head is used to convert the output mask feature perception information and instance mask into the final prediction result and evaluate the quality of each prediction. This module bypasses the traditional bipartite graph matching method based on the Hungarian algorithm by decoupling the matching. According to the mask feature perception information, the corresponding relationship from the instance mask to the real instance is directly established, which significantly improves the efficiency of the matching process. After matching, the instance loss can be finally calculated. Classification errors are penalized by cross entropy loss.

[0084] The training hyperparameter module configuration of the deep learning model described in step S4 includes the following training hyperparameters that need to be adjusted: number of threads num_workers, learning rate lr, training batch size batch_size, number of iterations iteration, weight decay weight_decay, optimizer optimizer, and learning rate scheduler param_scheduler.

[0085] Furthermore, the post-processing optimization in step S5 uses a combination of statistical filtering algorithm and cluster analysis method to process the instance information point cloud data to obtain optimized instance information point cloud data, aiming to not only remove outliers but also accurately identify and remove abnormal instance labels.

[0086] Furthermore, the specific process of obtaining the size, spatial position and other information of the building components in step S6 is as follows:

[0087] S6-1: First, the building components are divided into three-dimensional components (such as columns and beams) and planar components (such as walls, panels, doors and windows) according to their point cloud scanning features;

[0088] S6-2: For three-dimensional components with multiple planes, a multi-plane fitting method based on random consensus (RANSAC) is used to fit each plane and calculate the distance between parallel planes. Single-plane RANSAC fitting is inefficient and ineffective when facing point cloud data with uneven density acquired by laser scanners. Multi-plane RANSAC fitting finds the optimal plane set through multiple random sampling and model estimation, and fits multiple planes at the same time, which improves the fitting accuracy;

[0089] S6-3: For planar components with only a single plane, the oriented bounding box (OBB) method can be used to more accurately represent its boundaries. The size information of the planar component can be obtained by measuring the size and center coordinates of the oriented bounding box.

[0090] Embodiment 2:

[0091] This embodiment introduces a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, a lightweight detection method for building components based on OneFormer3D as described in any one of Embodiments 1 is implemented.

[0092] Embodiment 3:

[0093] A computer device comprising:

[0094] Memory, used to store instructions.

[0095] The processor is used to execute the instructions so that the computer device performs an operation based on the OneFormer3D building component lightweight detection method as described in any one of Example 1.

[0096] Embodiment 4:

[0097] This embodiment introduces an application process of the OneFormer3D building component lightweight detection method according to the present invention, which specifically includes the following steps:

[0098] S1: Point cloud data collection: using Figure 2 The terrestrial laser scanner shown collects point cloud data in multiple building scenes, including concrete structure and decoration and renovation stages, such as Figure 3 As shown, the scanning resolution is 8000 points per meter (Pt / m). In the future, it is planned to expand to more construction sites.

[0099] S2: De-noise and simplify the point cloud data obtained in step S1. The processed point cloud data is divided into different areas, each area does not exceed 5 million point clouds, such as Figure 4 shown.

[0100] S3: The point cloud data processed in step S2 is labeled as wall, column, beam, plate, door, window and other instances according to the building component categories to obtain a point cloud data set that can be used for quality inspection (the total number of point clouds is about 350 million), such as Figure 5 As shown in Figure (a) and Figure (b), the point cloud dataset in this implementation adopts a 7:2:1 ratio, and the dataset is saved in S3DIS format.

[0101] S4: Use the deep learning model based on Onefomer3D network structure optimization to train the point cloud dataset. The backbone is pre-trained on the S3DIS dataset, Structured3D dataset, and ScanNet dataset. On this basis, the downsampling module and residual connection module are added to reduce the computational overhead, ensure the information integrity of the deep neural network, ensure the lightweight model, and improve the training efficiency and accuracy. Figure 7 (a) shows the original input point cloud dataset. Figure 7 (b) is a schematic diagram after deep network instance segmentation. By setting different training hyperparameter modules for multiple trainings, the best hyperparameter module is obtained, and the training is performed with the hyperparameter module configuration, and the training parameters are saved. The model training results of the point cloud training set selected in this example are shown in Table 1; the unlabeled point cloud can obtain the corresponding instance information after being processed by the deep learning model, as shown in Table 1.

[0102] Table 1 is a specific data diagram of model training under an example.

[0103]

[0104] The hyperparameter module configuration of the deep learning model based on Onefomer3D network structure optimization is as follows:

[0105] Use LightGBM to adjust hyperparameters and use Bayesian optimization to automatically find the optimal parameter configuration. By setting the maximum depth from 5 to 50 and sampling with a step size of 5, multiple hyperparameter combinations are generated. Each combination is evaluated by k-fold cross validation to ensure that the selected parameters have good generalization ability.

[0106] The hyperparameters that need to be adjusted include: number of threads num_workers, learning rate lr, training batch size batch_size, number of iterations iteration, weight decay weight_decay, optimizer optimizer, and learning rate scheduler param_scheduler.

[0107] In this embodiment, the training is performed on a server with a CPU model of Intel(R) Xeon(R) Platinum 8336C CPU@2.30GHz, a GPU model of NVIDIA GeForce RTX 4090 (24G), and an operating system of Ubuntu 20.04. The optimal configuration is as follows Figure 8 as shown in .

[0108] S5: In order to optimize the point cloud data exported by S4 with some erroneous instance labels, advanced data processing technology combining statistical filtering and cluster analysis is used for post-processing. This can not only effectively remove outliers, but also accurately identify and remove abnormal instance labels, thereby further improving the reliability and accuracy of the data.

[0109] S6: The optimized data is divided into three-dimensional components (such as columns and beams) and plane components (such as walls, panels, doors, and windows). For three-dimensional components with multiple planes, the multi-plane fitting technology based on random consensus (RANSAC) is used to fit each plane and calculate the parallel plane distance, such as Fig. 9 As shown. Traditional single-plane RANSAC fitting has strong assumptions about data distribution and can only fit a single model. It is ineffective and inefficient when facing point cloud data with uneven density obtained by laser scanners. Multi-plane RANSAC fitting finds the optimal plane set through multiple random sampling and model estimation, and fits multiple planes at the same time, which improves the fitting accuracy, thereby obtaining information such as the size, spatial position and flatness of the three-dimensional component. For planar components that usually have only a single plane, the use of an directed bounding box (OBB) can more accurately represent its boundary. The size information of the planar component can be obtained by measuring the size and center coordinates of the envelope box, such as Fig.10 shown.

[0110] S7: Combine the position information of all components in S6 with the specifications to achieve lightweight component detection.

[0111] The present invention proposes a lightweight detection method for building components based on OneFormer3D. The method combines point cloud processing technology and deep neural network model, and can be effectively applied to the detection of building components in concrete engineering and decoration engineering. The method includes: collecting point cloud data, downsampling, denoising and annotating the data to produce a training data set, optimizing the deep neural network model, and using the produced data set for model training, inputting the point cloud data to be predicted into the deep learning model, obtaining instance information inside the building, and further processing and optimizing the results, using the multi-plane fitting technology based on random consistency (RANSAC) and the directed envelope box (OBB) algorithm to obtain the size, spatial position and other information of the component, and compare with the relevant specifications to complete the lightweight detection. The present invention can efficiently and accurately realize the detection of building components, and has broad application prospects.

[0112] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A lightweight detection method for building components based on OneFormer3D, characterized in that: Specifically include: Step S1: Obtaining point cloud data inside the building; Step S2: De-noising and simplifying the point cloud data inside the building to obtain pre-processed point cloud data; Step S3: classifying the preprocessed point cloud data by region, and marking the building component types of the point cloud data in each region, obtaining each building component in each region and the marking of each building component as a point cloud data set; Step S4: using the deep learning model based on the OneFormer3D network structure optimization to train the point cloud data set, and performing multiple trainings by setting different training hyperparameter modules to obtain the best training hyperparameter module, and configuring the deep learning model with the best training hyperparameter module for training to obtain a trained deep learning model; inputting the point cloud data to be predicted into the trained deep learning model to obtain the instance information point cloud data inside the building; Step S5: performing post-processing optimization on the instance information point cloud data to obtain optimized instance information point cloud data; Step S6: using a multi-plane fitting method based on random consistency and a directed envelope box algorithm based on principal component analysis on the optimized instance information point cloud data to obtain the size information and spatial position information of the building components corresponding to the optimized instance information point cloud data; Step S7: Compare the size information and position information of the building component with the design specifications to complete the detection of the building component.

2. The method for detecting lightweight building components based on OneFormer3D according to claim 1, characterized in that: The point cloud data inside the building in step S1 meets the following requirements: the data density of the point cloud data contains at least 100 points per square meter, and shall not be less than 50 points per square meter under special circumstances or requirements; the spatial range covers the entire target area, including the area within 30 meters outside the edge wire; the file storage format is the E57 file format, the scanning resolution is 8000 points per meter, and the ranging error is ≤±1mm at 25 meters; the point cloud data contains RGB color information.

3. The method for detecting lightweight building components based on OneFormer3D according to claim 1, characterized in that: The denoising process uses Gaussian filtering, and the simplification uses an Octree tree structure to perform 12-level downsampling simplification on the point cloud data.

4. The method for detecting lightweight building components based on OneFormer3D according to claim 1, characterized in that: The deep learning model specifically includes: an output end of the downsampling module is respectively connected to the second input end of the mask module, the second input end of the instance module and the input end of the superpoint extraction module, and the output end of the superpoint extraction module is respectively connected to the first input end of the mask module, the first input end of the decoder and the first input end of the instance module; the output end of the instance module is connected to the second input end of the decoder, the output end of the mask module is connected to the first input end of the prediction head, and the output end of the decoder is connected to the second input end of the prediction head; The downsampling module includes a first fully connected layer with a dimension of 6x32, a farthest point sampling layer, a K nearest neighbor layer, a second fully connected layer with a dimension of 32x16, and a maximum pooling layer, which are connected in sequence; The farthest point sampling layer is used to iteratively calculate the distance from each point in the current point set to the center of the existing point cloud model, and select the N points with the farthest distance as the new subset; The K nearest neighbor layer is used to capture the local structural information of the point set; The superpoint extraction module is used to voxelize the input point cloud, convert it into a sparse three-dimensional convolutional format, and extract point-by-point features; The mask module is used to integrate the residual mask feature vector output by the downsampling module and the feature vector output by the superpoint extraction module to perceive the global feature; The instance module is used to obtain a query vector with instance features from the downsampling module and the superpoint extraction module; The decoder is used to convert the high-dimensional query vector into a low-dimensional feature representation; The prediction head is used to convert the output mask feature perception information and instance mask into a final prediction result and evaluate the quality of each prediction.

5. The method for detecting lightweight building components based on OneFormer3D according to claim 1, characterized in that: The training hyperparameter module configuration includes: number of threads, learning rate, training batch size, number of iterations, weight decay, optimizer and learning rate scheduler.

6. The method for detecting lightweight building components based on OneFormer3D according to claim 1, characterized in that: The post-processing optimization adopts a statistical filtering algorithm and a cluster analysis method.

7. The method for detecting lightweight building components based on OneFormer3D according to claim 1, characterized in that: The annotated objects of the point cloud data set also include on-site objects.

8. The method for detecting lightweight building components based on OneFormer3D according to claim 1, characterized in that: The step S6 specifically includes: S6-1: Building components are divided into three-dimensional components and planar components according to their point cloud scanning features; S6-2: For a three-dimensional component with multiple planes, a multi-plane fitting method based on random consistency is used to fit each plane and calculate the distance between parallel planes to obtain the size information of the three-dimensional component; S6-3: For a planar component having only a single plane, the size information of the planar component is obtained by measuring the size and center coordinates of the directed bounding box.

9. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by a processor, a lightweight detection method for building components based on OneFormer3D is implemented as described in any one of claims 1 to 8.

10. A computer device, characterized in that: include: A memory for storing instructions; The processor is used to execute the instructions so that the computer device performs the operation of the OneFormer3D building component lightweight detection method according to any one of claims 1 to 8.

Images