Screen cabinet adding method based on BIM and computer vision technology

Abstract

The invention discloses a screen cabinet adding method based on BIM and a computer vision technology. The method comprises the following steps: 1) establishing a BIM model template; 2) carrying out 2D visual detection and ROI extraction; 3) acquiring and downsampling 3D point cloud data; 4) processing the downsampled 3D point cloud data by using point cloud correction filtering, including removing point cloud outliers by using an improved DBSCAN algorithm and performing point cloud completion based on a grid method, thereby improving the quality of the point cloud; and 5) registering the point cloud and the BIM model template, and verifying whether the installation space meets the installation requirements of the screen cabinet or not. According to the method, 2D and 3D visual technologies are combined, efficient and accurate space information acquisition and screen cabinet adding feasibility evaluation are achieved, the labor cost is reduced, the design accuracy is improved, and technical support is provided for efficient transformation of the urban power distribution room.

Description

Technical Field

[0001] The present invention belongs to the technical field of equipment management and transformation of power distribution rooms, and in particular relates to a screen cabinet installation method based on BIM and computer vision technology. Background Art

[0002] Urban power distribution rooms are widely distributed in cities and are scattered in locations. As electricity demand changes, it is often necessary to renovate the power distribution rooms, usually by installing new cabinets to meet demand. Traditional methods rely on on-site surveys to confirm equipment space, cable trench layout, and cabinet installation locations. This method not only consumes a lot of manpower, but is also prone to inaccurate or missing data. Summary of the invention

[0003] The purpose of the present invention is to address the deficiencies of the prior art. The present invention proposes a screen cabinet installation method based on BIM and computer vision technology, which combines 2D and 3D vision technology to achieve efficient and accurate spatial information acquisition and screen cabinet installation feasibility assessment, which not only reduces labor costs, but also improves the accuracy of the design, providing technical support for the efficient transformation of urban distribution rooms.

[0004] In order to achieve the above object, the present invention adopts the following technical solution.

[0005] A screen cabinet installation method based on BIM and computer vision technology includes the following steps: Step S1: Create a BIM model template Use BIM software to build a BIM model template for the power distribution room. The BIM model template contains the key features of the cable trench and cabinets in the power distribution room, which serves as a reference framework for subsequent visual inspection and point cloud registration to ensure the spatial consistency between the point cloud and the BIM model template. Through selective input, only the feature data related to the installation of the cabinets is input into the BIM model template, reducing unnecessary data processing and optimizing the resource utilization efficiency of the system. Step S2: 2D visual detection and ROI extraction The 2D camera installed on the 3D camera is used to obtain the RGB image of the distribution room. The target detection algorithm Nanodet identifies the key feature areas of cable trenches and cabinets, and marks the region of interest (ROI) in the RGB image, providing accurate spatial position reference for 3D point cloud data processing; Step S3: Acquire and downsample 3D point cloud data In the region of interest ROI marked in step S2, a 3D camera is used to acquire point cloud data, the point cloud is divided into spherical regions by a uniform sampling method, and the center point of each spherical region is selected as a sampling point to reduce the data density while retaining the main geometric features of the spherical region; Step S4, using point cloud correction filtering to process the downsampled 3D point cloud data to improve the point cloud quality; Step S5: Registration of point cloud and BIM model template By extracting key contour features from the BIM model template and point cloud data, calculating the normal vector and Euclidean distance of the feature points, performing weighted calculation on the similarity between the two, and using the KD tree structure to implement nearest neighbor search and matching, the point cloud data is accurately aligned with the BIM model template, and evaluating whether the additional space meets the requirements for cabinet installation.

[0006] Specifically, the algorithm network structure of the pre-trained target detection algorithm Nanodet in step S2 includes: Step S21, using a lightweight ShuffleNetV2 structure, removing the last convolution layer of the algorithm network, retaining the 16x and 32x downsampling layers, and improving the inference speed of the model in detection; Step S22: Use a feature fusion module in the Neck part. The feature fusion module includes a 3*3 convolution and two dilated convolutions with different dilation rates. The Prelu function is used to enhance the nonlinear ability of the model and avoid the problem of neuron death. Then, the Concat operation is used to fuse the feature information, and the dimension is reduced to the dimension of the input feature through a 1*1 convolution. Then, two convolution layers are used to extract richer deep features, and a residual connection operation is performed with the input features to avoid the loss of shallow information during transmission.

[0007] Specifically, the step S4 of processing the downsampled 3D point cloud data using the point cloud correction filter includes the following steps: Step S41: Use the improved DBSCAN algorithm to remove outliers from the point cloud According to the geometric properties of the cable or cabinet, the scanning radius R and the minimum number of points M are set as algorithm parameters, and the improved DBSCAN algorithm is used to mark the core points that meet the neighborhood density conditions and recursively expand them to form a reasonable clustering cluster; for the points that do not meet the neighborhood density conditions, they are marked as noise points and removed from the point cloud data, thereby obtaining the optimized point cloud data; Step S42: Point cloud completion based on grid method Furthermore, the improved DBSCAN algorithm in step S41 specifically includes the following contents: Step S411, setting the scanning radius R and the minimum number of points M according to the geometric properties of the cable and the cabinet; Step S412: Substitute R and M, traverse the point cloud data, check the number of points in the neighborhood of each point, divide the core point and use it as the starting point of the cluster, divide the points not in the cluster as noise, and recursively expand each core point until all point cloud data are traversed; Step S413, calculate the center point coordinates (x_mean, y_mean) of the region of interest ROI, obtain the center value Z_mean of the object in the Z direction through clustering in step S411 and step S412, and then obtain the centroid coordinates P (x_mean, y_mean, z_mean) of the object; Step S414, calculate the distance and direction from each core point and noise point to P, and set a threshold to automatically remove core points with unreasonable ranges and misjudged noise points based on the prior characteristics of the object's geometric properties, so as to obtain a more reasonable clustering result.

[0008] Furthermore, the point cloud completion based on the grid method in step S42 specifically includes the following steps: Step S421, dividing the denoised point cloud data into three-dimensional grids; Step S422: fill in missing points in the divided three-dimensional grid using an interpolation algorithm.

[0009] Specifically, the registration of the point cloud and the BIM model template in step S5 specifically includes the following steps: Step S51, extracting contour features from the BIM model template and point cloud data and performing interval sampling to reduce the computational burden; Step S52, calculating the normal vector of each feature point, and performing weighted summation after normalization based on the Euclidean distance and normal vector direction difference between the feature points to obtain a similarity score; Step S53: By constructing a KD tree structure and performing a nearest neighbor search, the points with the highest similarity in the point cloud are matched with the feature points in the BIM model template, thereby achieving accurate alignment of the point cloud and the BIM model template.

[0010] Compared with the prior art, the present invention has the following beneficial effects: 1. The method of the present invention reduces manual survey and measurement errors: by obtaining complete spatial information of the distribution room through a three-dimensional scan, the problem of multiple on-site surveys and heavy workload required for traditional manual surveys is avoided, and the measurement errors caused by environmental limitations in manual surveys are reduced.

[0011] 2. The method of the present invention can effectively improve the design efficiency. By combining BIM and computer vision technology, it can automatically calculate the installation space of the panel cabinet and quickly evaluate whether the available length of the cable trench meets the demand, thereby greatly shortening the design time.

[0012] 3. The BIM model template of the method of the present invention can automatically analyze the installation space and determine whether the size and position of the screen cabinet meet the installation requirements, thereby improving the accuracy and effectiveness of the installation plan.

[0013] 4. The method of the present invention supports remote planning, performs remote design and feasibility analysis through a digital three-dimensional model, realizes the intelligence and efficiency of the transformation of the distribution room, and reduces manpower and time costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present disclosure and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying creative work.

[0015] Figure 1 It is a flow chart of a screen cabinet installation method based on BIM and computer vision technology of the present invention; Figure 2 It is a schematic diagram of the network structure of the Nanodet algorithm after pre-training of the present invention; Figure 3 It is a schematic diagram of the feature fusion module architecture of the present invention; Figure 4 is a flow chart of point cloud correction filtering of the present invention; Figure 5 It is a schematic diagram of the BIM model of the power distribution room in the example of the present invention; Figure 6 It is a schematic diagram of the cabinet position prediction and ROI generation in the example of the present invention; Figure 7 It is a schematic diagram of the point cloud data processing result of the screen cabinet in the example of the present invention; Figure 8 It is a schematic diagram of the registration of the BIM model and the screen cabinet point cloud in the example of the present invention. DETAILED DESCRIPTION

[0016] In order to facilitate those of ordinary skill in the art to understand and implement the present invention, the following detailed description is given to each step of the method proposed by the present invention, and it should be understood that these embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention. In addition, it should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms also fall within the scope defined by the appended claims of the present application.

[0017] Example like Figure 1 As shown, the present invention discloses a screen cabinet installation method based on BIM and computer vision technology, comprising the following steps: Step S1: Create a BIM model template Use BIM software to build a BIM model template for the power distribution room. The BIM model template contains the key features of the cable trench and cabinets in the power distribution room, which serves as a reference framework for subsequent visual inspection and point cloud registration to ensure the spatial consistency between the point cloud and the BIM model template. Through selective input, only the feature data related to the installation of the cabinets is input into the BIM model template, reducing unnecessary data processing and optimizing the resource utilization efficiency of the system. Step S2: 2D visual detection and ROI extraction The 2D camera installed on the 3D camera is used to obtain the RGB image of the distribution room. The target detection algorithm Nanodet identifies the key feature areas of cable trenches and cabinets, and marks the region of interest (ROI) in the RGB image, providing accurate spatial position reference for 3D point cloud data processing; Step S3: Acquire and downsample 3D point cloud data In the region of interest ROI marked in step S2, a 3D camera is used to acquire point cloud data, the point cloud is divided into spherical regions by a uniform sampling method, and the center point of each spherical region is selected as a sampling point to reduce the data density while retaining the main geometric features of the spherical region; Step S4, using point cloud correction filtering to process the downsampled 3D point cloud data to improve the point cloud quality; Step S5: Registration of point cloud and BIM model template By extracting key contour features from the BIM model template and point cloud data, calculating the normal vector and Euclidean distance of the feature points, performing weighted calculation on the similarity between the two, and using the KD tree structure to implement nearest neighbor search and matching, the point cloud data is accurately aligned with the BIM model template, and evaluating whether the additional space meets the requirements for cabinet installation.

[0018] Specifically, Figure 2 As shown, the algorithm network structure of the pre-trained target detection algorithm Nanodet in step S2 includes: Step S21, using a lightweight ShuffleNetV2 structure, removing the last convolution layer of the algorithm network, retaining the 16x and 32x downsampling layers, and improving the inference speed of the model in detection; Step S22: Use feature fusion module in the Neck part, such as Figure 3As shown in the figure, the feature fusion module includes a 3*3 convolution and two dilated convolutions with different dilation rates. The Prelu function is used to enhance the nonlinear ability of the model and avoid the problem of neuron death. The Concat operation is then used to fuse the feature information, and the dimension is reduced to the dimension of the input feature through a 1*1 convolution. Then, two convolution layers are used to extract richer deep features, and a residual connection operation is performed with the input features to avoid the loss of shallow information during transmission.

[0019] Specifically, Figure 4 As shown, the step S4 uses the point cloud correction filter to process the downsampled 3D point cloud data, including the following steps: Step S41: Use the improved DBSCAN algorithm to remove outliers from the point cloud According to the geometric properties of the cable or cabinet, the scanning radius R and the minimum number of points M are set as algorithm parameters, and the improved DBSCAN algorithm is used to mark the core points that meet the neighborhood density conditions and recursively expand them to form a reasonable clustering cluster; for the points that do not meet the neighborhood density conditions, they are marked as noise points and removed from the point cloud data, thereby obtaining the optimized point cloud data; Step S42: Point cloud completion based on grid method Furthermore, the improved DBSCAN algorithm in step S41 specifically includes the following contents: Step S411, setting the scanning radius R and the minimum number of points M according to the geometric properties of the cable and the cabinet; Step S412: Substitute R and M, traverse the point cloud data, check the number of points in the neighborhood of each point, divide the core point and use it as the starting point of the cluster, divide the points not in the cluster as noise, and recursively expand each core point until all point cloud data are traversed; Step S413, calculate the center point coordinates (x_mean, y_mean) of the region of interest ROI, obtain the center value Z_mean of the object in the Z direction through clustering in step S411 and step S412, and then obtain the centroid coordinates P (x_mean, y_mean, z_mean) of the object; Step S414, calculate the distance and direction from each core point and noise point to P, and set a threshold to automatically remove core points with unreasonable ranges and misjudged noise points based on the prior characteristics of the object's geometric properties, so as to obtain a more reasonable clustering result.

[0020] Furthermore, the point cloud completion based on the grid method in step S42 specifically includes the following steps: Step S421, dividing the denoised point cloud data into three-dimensional grids; Step S422: fill in missing points in the divided three-dimensional grid using an interpolation algorithm.

[0021] Specifically, the registration of the point cloud and the BIM model template in step S5 specifically includes the following steps: Step S51, extracting contour features from the BIM model template and point cloud data and performing interval sampling to reduce the computational burden; Step S52, calculating the normal vector of each feature point, and performing weighted summation after normalization based on the Euclidean distance and normal vector direction difference between the feature points to obtain a similarity score; Step S53: By constructing a KD tree structure and performing a nearest neighbor search, the points with the highest similarity in the point cloud are matched with the feature points in the BIM model template, thereby achieving accurate alignment of the point cloud and the BIM model template.

[0022] The feasibility of the method of the present invention is further verified by a specific installation example.

[0023] The traditional process of judging the feasibility of manual installation is as follows: 1. Before each installation, an infrared rangefinder must be used to accurately measure the wall dimensions on site.

[0024] 2. Draw a preliminary sketch by hand on site.

[0025] 3. Use professional software to convert sketches into CAD drawings to address on-site judgment problems caused by the irregular shapes of some distribution rooms.

[0026] 4. Finally, decide whether the installation can be carried out based on the CAD drawings.

[0027] The installation feasibility judgment process of the method of the present invention is as follows: 1. Obtain BIM model: Based on a systematic survey, establish the BIM model of the power distribution room in the urban area; Figure 5 As shown in the figure, find the corresponding model according to the distribution room that needs to be renovated.

[0028] 2. Generate ROI: Figure 6 As shown, the improved Nanodet algorithm is used to predict the position and confidence of the screen cabinet, and the predicted area is defined as ROI, that is, only the point cloud data needs to be processed within this area.

[0029] 3. Point cloud processing: Extract the point cloud data of the screen cabinet from the ROI, and obtain the point cloud data that can be used for registration through point cloud post-processing, such as Figure 7 Shown is the point cloud processing result of a screen cabinet.

[0030] 4. Accurately align the BIM model and the cabinet point cloud, such as Figure 8 As shown, the BIM model is accurately aligned with the screen cabinet point cloud to draw a conclusion on whether the installation can be performed.

[0031] In summary, by comparing the method of the present invention with the traditional method for judging the feasibility of retrofitting, it can be clearly seen that: first, the method of the present invention can effectively reduce manual survey and measurement errors: the complete spatial information of the distribution room is obtained through a three-dimensional scan, which avoids the problem of multiple on-site surveys and heavy workload required for traditional manual surveys, and reduces the measurement errors caused by environmental limitations in manual surveys; secondly, the method of the present invention can effectively improve the design efficiency, and automatically calculate the installation space of the panel cabinet by combining BIM and computer vision technology, and quickly evaluate whether the available length of the cable trench meets the demand, which greatly shortens the design time and improves the accuracy and effectiveness of the retrofit plan; finally, the method of the present invention supports remote planning, and performs remote design and feasibility analysis through digital three-dimensional models, thereby realizing the intelligence and efficiency of the transformation of the distribution room and reducing manpower and time costs.

[0032] The above is only a preferred embodiment of the present invention, and does not limit the present invention in other forms. Any technician familiar with the profession may use the above disclosed technical content to change or modify it into an equivalent embodiment with equivalent changes. However, any simple modification, equivalent change and modification made to the above embodiment according to the technical essence of the present invention without departing from the technical solution of the present invention still belongs to the protection scope of the technical solution of the present invention.

Claims

1. A screen cabinet installation method based on BIM and computer vision technology, characterized in that: The following steps are involved: Step S1: Create a BIM model template Use BIM software to build a BIM model template for the power distribution room. The BIM model template contains the key features of the cable trench and cabinets in the power distribution room, which serves as a reference framework for subsequent visual inspection and point cloud registration to ensure the spatial consistency between the point cloud and the BIM model template. Through selective input, only the feature data related to the installation of the cabinets is input into the BIM model template; Step S2: 2D visual detection and ROI extraction The 2D camera installed on the 3D camera is used to obtain the RGB image of the distribution room. The target detection algorithm Nanodet identifies the key feature areas of cable trenches and cabinets, and marks the region of interest (ROI) in the RGB image, providing accurate spatial position reference for 3D point cloud data processing; Step S3: Acquire and downsample 3D point cloud data In the region of interest ROI marked in step S2, a 3D camera is used to acquire point cloud data, the point cloud is divided into spherical regions by a uniform sampling method, and the center point of each spherical region is selected as a sampling point to reduce the data density while retaining the main geometric features of the spherical region; Step S4, using point cloud correction filtering to process the downsampled 3D point cloud data to improve the point cloud quality; Step S5: Registration of point cloud and BIM model template By extracting key contour features from the BIM model template and point cloud data, calculating the normal vector and Euclidean distance of the feature points, performing weighted calculation on the similarity between the two, and using the KD tree structure to implement nearest neighbor search and matching, the point cloud data is accurately aligned with the BIM model template, and evaluating whether the additional space meets the requirements for cabinet installation.

2. According to the method of installing a screen cabinet based on BIM and computer vision technology in claim 1, it is characterized in that: The algorithm network structure of the pre-trained target detection algorithm Nanodet in step S2 includes: Step S21, using a lightweight ShuffleNetV2 structure, removing the last convolution layer of the algorithm network, retaining the 16x and 32x downsampling layers, and improving the inference speed of the model in detection; Step S22: Use a feature fusion module in the Neck part. The feature fusion module includes a 3*3 convolution and two dilated convolutions with different dilation rates. The Prelu function is used to enhance the nonlinear ability of the model and avoid the problem of neuron death. Then, the Concat operation is used to fuse the feature information, and the dimension is reduced to the dimension of the input feature through a 1*1 convolution. Then, two convolution layers are used to extract richer deep features, and a residual connection operation is performed with the input features to avoid the loss of shallow information during transmission.

3. According to the method of installing a screen cabinet based on BIM and computer vision technology in claim 1, it is characterized in that: The step S4 uses point cloud correction filtering to process the downsampled 3D point cloud data, including the following steps: Step S41: Use the improved DBSCAN algorithm to remove outliers from the point cloud According to the geometric properties of the cable or cabinet, the scanning radius R and the minimum number of points M are set as algorithm parameters, and the improved DBSCAN algorithm is used to mark the core points that meet the neighborhood density conditions and recursively expand them to form a reasonable clustering cluster; for the points that do not meet the neighborhood density conditions, they are marked as noise points and removed from the point cloud data, thereby obtaining the optimized point cloud data; Step S42: point cloud completion based on the grid method.

4. According to the method of installing a screen cabinet based on BIM and computer vision technology in claim 3, it is characterized in that: The improved DBSCAN algorithm described in step S41 specifically includes the following contents: Step S411, setting the scanning radius R and the minimum number of points M according to the geometric properties of the cable and the cabinet; Step S412: Substitute R and M, traverse the point cloud data, check the number of points in the neighborhood of each point, divide the core point and use it as the starting point of the cluster, divide the points not in the cluster as noise, and recursively expand each core point until all point cloud data are traversed; Step S413, calculate the center point coordinates (x_mean, y_mean) of the region of interest ROI, obtain the center value Z_mean of the object in the Z direction through clustering in step S411 and step S412, and then obtain the centroid coordinates P (x_mean, y_mean, z_mean) of the object; Step S414, calculate the distance and direction from each core point and noise point to P, and set a threshold to automatically remove core points with unreasonable ranges and misjudged noise points based on the prior characteristics of the object's geometric properties, so as to obtain a more reasonable clustering result.

5. According to the method of installing a screen cabinet based on BIM and computer vision technology in claim 3, it is characterized in that: The point cloud completion based on the grid method in step S42 specifically includes the following steps: Step S421, dividing the denoised point cloud data into three-dimensional grids; Step S422: fill in missing points in the divided three-dimensional grid using an interpolation algorithm.

6. The method for installing a screen cabinet based on BIM and computer vision technology according to claim 1 is characterized in that: The registration of the point cloud and the BIM model template in step S5 specifically includes the following steps: Step S51, extracting contour features from the BIM model template and point cloud data and performing interval sampling to reduce the computational burden; Step S52, calculating the normal vector of each feature point, and performing weighted summation after normalization based on the Euclidean distance and normal vector direction difference between the feature points to obtain a similarity score; Step S53: By constructing a KD tree structure and performing a nearest neighbor search, the points with the highest similarity in the point cloud are matched with the feature points in the BIM model template, thereby achieving accurate alignment of the point cloud and the BIM model template.

Images