A smart construction operation, maintenance, management and control method based on BIM technology

Abstract

The application provides a smart construction operation and maintenance management method based on BIM technology, and relates to the technical field of construction management. The method comprises the following steps: obtaining a multi-angle projection image of a target component BIM initial model, comparing the shape with a two-dimensional drawing input image, and obtaining missing component two-dimensional design parameters; sorting the modeling complexity of the missing component two-dimensional design parameters from simple to complex to obtain a design parameter sequence; constructing a component index binary tree according to the design parameter sequence; based on the first-level node design parameter of the index binary tree, searching for a first-level matching BIM modeling parameter from a modeling material library, when the first-level matching BIM modeling parameter is not empty, searching for a second-level matching BIM modeling parameter, and searching for N-level matching BIM modeling parameters until the N-level matching BIM modeling parameters are obtained; and sending the first-level matching BIM modeling parameter, the second-level matching BIM modeling parameter, and the N-level matching BIM modeling parameter to a construction management end for construction operation and maintenance management. The technical problem of untimely and inaccurate construction operation and maintenance management in the prior art is solved.

Description

Technical Field

[0001] This invention relates to the field of construction management and control, and in particular to a smart construction operation and maintenance management and control method based on BIM technology. Background Technology

[0002] In complex building construction, when faced with problems such as missing or inaccurate representation of BIM model components (especially irregular structures, curved components, or special nodes), traditional solutions mainly rely on human experience and manual technical means, such as relying on construction personnel to manually supplement the model. However, for complex missing nodes, relying on experience to imagine the three-dimensional shape can easily lead to misunderstandings and low efficiency, which in turn leads to untimely and inaccurate construction operation and maintenance management. Summary of the Invention

[0003] This invention addresses the technical problems of untimely and inaccurate construction operation and maintenance management in existing technologies by providing a smart construction operation and maintenance management method based on BIM technology.

[0004] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0005] This invention provides a smart construction operation and maintenance management method based on BIM technology, including:

[0006] Obtain multi-angle projection images of the initial BIM model of the target component, compare them with the images entered in the two-dimensional drawings, and obtain the two-dimensional design parameters of the missing component.

[0007] The modeling complexity of the two-dimensional design parameters of the missing component is sorted from simple to complex to obtain a sequence of design parameters.

[0008] Component index binary tree according to the design parameter sequence;

[0009] Based on the design parameters of the first-level nodes of the indexed binary tree, first-level matching BIM modeling parameters are retrieved from the modeling material library. When the first-level matching BIM modeling parameters are not empty, second-level matching BIM modeling parameters are retrieved from the modeling material library based on the design parameters of the second-level nodes of the indexed binary tree, until N-level matching BIM modeling parameters are obtained, where N represents the number of design parameters.

[0010] The first-level matching BIM modeling parameters, the second-level matching BIM modeling parameters, and so on up to the Nth-level matching BIM modeling parameters are sent to the construction management terminal for construction operation and maintenance control.

[0011] The beneficial effects of this invention are:

[0012] Compared to existing technologies, this application first obtains multi-angle projection images of the initial BIM model of the target component, compares their shapes with the images entered in 2D drawings, obtains the 2D design parameters of the missing component, and performs multi-view union analysis through deviation region union analysis to accurately locate the shape deviation region, providing a reliable data foundation for subsequent operation and maintenance management. Secondly, the 2D design parameters of the missing component are sorted from simple to complex in terms of modeling complexity to obtain a design parameter sequence, generating a processing sequence from simple to complex, optimizing the BIM modeling process and improving efficiency. Thirdly, an indexed binary tree is constructed according to the design parameter sequence to quickly match parameters in the modeling material library through hierarchical retrieval, improving model retrieval efficiency. Furthermore, based on the design parameters of the first-level nodes of the indexed binary tree, first-level matching BIM modeling parameters are retrieved from the modeling material library. When the first-level matching BIM modeling parameters are not empty, second-level matching BIM modeling parameters are retrieved from the modeling material library based on the design parameters of the second-level nodes of the indexed binary tree, until N-level matching BIM modeling parameters are obtained. This allows for rapid location of the required parameters, reducing retrieval time and improving efficiency. Finally, the first-level matching BIM modeling parameters, the second-level matching BIM modeling parameters, and so on up to the Nth-level matching BIM modeling parameters are sent to the construction management terminal for construction operation and maintenance control. The construction management terminal can quickly obtain lightweight data and carry out construction operation and maintenance control accordingly, thereby improving the efficiency and accuracy of operation and maintenance control.

[0013] Through the above technical solution, this application determines the deviation area by comparing images from five views. Then, the design parameters for the deviation area are sorted based on modeling complexity. Based on this, existing modeling parameters are retrieved from the material library and sent to the construction management terminal. In this way, accurate BIM data support is provided for the construction site, realizing intelligent operation and maintenance management and avoiding the inefficiency caused by global comparison and the errors caused by human experience. Thus, the efficiency and accuracy of construction operation and maintenance are improved. Attached Figure Description

[0014] Figure 1 A flowchart illustrating a smart construction operation and maintenance management method based on BIM technology provided by this invention;

[0015] Figure 2 This is a schematic diagram illustrating the process of constructing an indexed binary tree in a smart construction operation and maintenance management method based on BIM technology provided by the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0018] In the description of this invention, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed herein.

[0019] Example 1, as Figure 1 As shown, this embodiment of the invention provides a smart construction operation and maintenance management method based on BIM technology, including:

[0020] S10: Obtain multi-angle projection images of the initial BIM model of the target component, compare the shape with the image entered in the two-dimensional drawing, and obtain the two-dimensional design parameters of the missing component.

[0021] In existing technologies, when dealing with the problem of missing components in BIM models during complex building construction, manual comparison is usually used. However, manual comparison is prone to misjudging irregular components (such as inclined steel beams) as multiple independent missing components from multiple perspectives, and manual comparison is inefficient.

[0022] To address the aforementioned issues, this application obtains multi-angle projection images of the initial BIM model of the target component, compares the image shapes with the images entered in the two-dimensional drawings from the same perspective, identifies the shape deviation areas, and retrieves the two-dimensional design parameters of the missing component accordingly.

[0023] Specifically, step S10 in the method includes:

[0024] Images are entered from the two-dimensional drawings, and the images of the front view, rear view, left view, right view, and top view are extracted.

[0025] Based on the construction positioning parameters, after locating the initial BIM model of the target component, frontal projection image, rear projection image, left projection image, right projection image and top projection image are acquired.

[0026] The front view paper image, the rear view paper image, the left view paper image, the right view paper image, and the top view paper image are compared with the front view projection image, the rear view projection image, the left view projection image, the right view projection image, and the top view projection image from the same viewing angle to obtain the shape deviation area.

[0027] Based on the shape deviation area and the image entered from the two-dimensional drawing, the two-dimensional design parameters of the missing component are retrieved.

[0028] In this embodiment, images are first entered from the two-dimensional drawing, and then the front view, rear view, left view, right view, and top view images are extracted. For example, based on the two-dimensional drawing, the views can be automatically segmented using view identifiers (such as the "front view" label) or layout features (such as fixed area divisions) to extract the images from the five perspectives: front, rear, left, right, and top.

[0029] Secondly, based on construction positioning parameters (such as construction coordinates), the initial BIM model of the target component is positioned, and then frontal projection images, rear projection images, left projection images, right projection images, and top projection images are acquired. For example, the initial BIM model can be spatially positioned according to construction positioning parameters (such as construction coordinates (X=100m, Y=50m, Z=0)), and then five standard view projection images of the positioned model can be generated using a BIM software API (such as the Revit API).

[0030] Next, the front view, rear view, left view, right view, and top view images are compared with the front projection, rear projection, left projection, right projection, and top projection images from the same viewing angle to obtain shape deviation areas. Specifically, the front view images obtained from the 2D drawing and the front projection images obtained from the BIM initial model are compared with the front projection images from the same viewing angle to obtain the front view deviation areas. This process is repeated for other views to obtain the rear view deviation areas, left view deviation areas, right view deviation areas, and top view deviation areas. The same-view image shape comparison process is based on... For example, comparing the top view image with the top projection image: the top view image has a 2500mm × 2500mm square area in the upper left corner, while the top projection image has no corresponding component; therefore, the top view deviation area is determined to be the upper left corner area.

[0031] Finally, based on the shape deviation area and in conjunction with the entered 2D drawing image, the 2D design parameters of the missing component are retrieved. Specifically, within the shape deviation area, the 2D design parameters (such as geometric dimensions, elevation, angles, etc.) of the missing component are extracted from the entered 2D drawing image. For example, based on the deviation area of ​​the top view drawing (such as the upper left corner area), the design parameters are retrieved from the entered 2D drawing image. For instance, component type: cooling tower foundation, dimensions: 2500mm × 2500mm, elevation: +23.50m.

[0032] Further, the step of "comparing the front view image, the rear view image, the left view image, the right view image, and the top view image with the front projection image, the rear projection image, the left projection image, the right projection image, and the top projection image from the same viewing angle to obtain the shape deviation area" includes:

[0033] The front view paper image, the rear view paper image, the left view paper image, the right view paper image, and the top view paper image are compared with the front view projection image, the rear view projection image, the left view projection image, the right view projection image, and the top view projection image from the same viewing angle to obtain the front view paper deviation area, the rear view paper deviation area, the left view paper deviation area, the right view paper deviation area, and the top view paper deviation area.

[0034] By using a pre-trained deviation region union analysis model, the deviation regions of the front view paper, the rear view paper, the left view paper, the right view paper, and the top view paper are processed to obtain the shape deviation regions. Among them, deviation regions that are marked with the same color in different drawings are the same deviation region.

[0035] In this embodiment, the deviation regions of five views are first obtained. Specifically, the front, rear, left, right, and top view paper-entry images are compared with the front, rear, left, right, and top view projected images from the same viewpoint to obtain the deviation regions of the front, rear, left, right, and top view paper. Further, the same-viewpoint image shape comparison process can be performed using a pre-trained image shape comparison model. This model is based on a convolutional neural network and automatically extracts image features through machine learning to achieve high-precision difference detection between the paper-entry images and the BIM projection images, thereby accurately locating the deviation regions of the same-viewpoint images.

[0036] For example, a Siamese network architecture image shape comparison model can be adopted, mainly consisting of an input layer, a shared convolutional encoder, a feature fusion and comparison layer, and an output layer. The input layer simultaneously takes in two images from the same viewpoint (e.g., an image entered from a frontal view paper and an image projected from the frontal view), both pre-processed with normalization. The shared convolutional encoder consists of multiple convolutional and pooling layers used to extract image shape features. The convolutional layers scan the image using convolutional kernels of different sizes (e.g., 3×3, 5×5). The lower convolutional layers extract low-level features such as edges, contours, and geometric shapes, while the deeper convolutional layers capture high-level semantic features such as holes, chamfers, and concave / convex structures. The pooling layers use max pooling or average pooling to reduce the feature map size, decrease computation, and improve translation invariance. For example, the lower convolutional layers output edge feature maps to identify basic shapes such as straight lines and arcs; the higher convolutional layers output semantic feature maps to identify complex structures such as "rectangular grooves" and "circular through holes." The feature fusion and comparison layer generates a difference feature map (e.g., displaying pixel-level shape deviations) by element-wise subtraction or concatenation of the feature maps output from the two encoders (e.g., 16×16×256 dimensions). It can also introduce an attention mechanism to focus on key areas in the image prone to design deviations (e.g., component connections, hole locations). The output layer restores the size through convolutional transposition or upsampling, outputting a deviation mask map of the same size as the input image. Different colors or grayscale values ​​are used to identify shape deviation areas (e.g., white areas indicate deviations).

[0037] For example, the training process of the image shape comparison model can be implemented through the following technical path: 1. Data preparation: Construct a dataset containing positive samples (image pairs where the image entered on paper and the projected image are completely identical) and negative samples (image pairs where the image entered on paper and the projected image are not completely identical), and divide it into training set, validation set, and test set according to a ratio of 7:1.5:1.5, using this as training data. 2. Model training: Calculate the error between the predicted result and the true label using loss functions such as binary cross-entropy loss and Dice loss. Update the network parameters using the backpropagation algorithm and adjust the model weights to make the model output closer to the true deviation. If the model overfits, it can be solved by adding regularization terms and adjusting the Dropout ratio. After multiple rounds of iterative training and optimization, if the accuracy reaches >95% on the validation set, the model is considered converged and training is stopped. Finally, use the test set to evaluate the model's generalization ability, enabling the model to achieve high accuracy and stability in detecting shape deviations between the image entered on paper and the projected image, meeting the actual construction operation and maintenance management needs.

[0038] Secondly, by using a pre-trained deviation region union analysis model, the deviation regions of the front view paper, rear view paper, left view paper, right view paper, and top view paper are processed to obtain shape deviation regions. Among them, deviation regions marked with the same color in different drawings are the same deviation regions. In this way, the same deviation in different views can be identified, eliminating redundancy and ambiguity.

[0039] Specifically, the construction process of the "deviation region union analysis model" is as follows:

[0040] Extract the front view recording image, rear view recording image, left view recording image, right view recording image and top view recording image of BIM modeling, use the first color to randomly select local areas, and obtain the first front view marker image, the first rear view marker image, the first left view marker image, the first right view marker image and the first top view marker image;

[0041] Based on the BIM modeling, the first color-selected area is clustered at the same location to obtain the first cluster of selected areas up to the Nth cluster of selected areas;

[0042] Based on the first cluster selection area up to the Nth cluster selection area, N colors different from the first color are used to select the first front view sign image, the first rear view sign image, the first left view sign image, the first right view sign image and the first top view sign image respectively, to obtain the second front view sign image, the second rear view sign image, the second left view sign image, the second right view sign image and the second top view sign image;

[0043] Using the second front-view sign image, the second rear-view sign image, the second left-view sign image, the second right-view sign image, and the second top-view sign image as supervision, and the first front-view sign image, the first rear-view sign image, the first left-view sign image, the first right-view sign image, and the first top-view sign image as input, a deviation region union analysis model is trained.

[0044] In this embodiment, the front view, rear view, left view, right view, and top view images from the BIM model are first extracted and then randomly selected using a first color to obtain a first front view marker image, a first rear view marker image, a first left view marker image, a first right view marker image, and a first top view marker image. This random selection is used to simulate the uncertainty of missing components during actual construction, thereby improving the model's generalization ability. For example, based on historical data, multiple local areas are randomly selected using a first color (e.g., red) from the front view, rear view, left view, right view, and top view images from the BIM model to obtain the first front view, rear view, left view, right view, and top view marker images.

[0045] Secondly, based on the BIM modeling, the selected areas with the first color outline are clustered at the same location to obtain the first cluster of selected areas up to the Nth cluster. Specifically, the selected areas can be clustered using Euclidean distance based on the spatial coordinates of the BIM model (for example, setting a threshold ≤ 0.5m, which can be adjusted according to the actual situation by those skilled in the art) to obtain the first cluster of selected areas up to the Nth cluster. For example, in the BIM model, if the coordinates of a selected area in the first top-view sign image are (100, 200, 50), and the coordinates of a selected area in the first front-view sign image are also (100, 200, 50), then these two areas are at the same location and are therefore clustered into cluster 1. This process is repeated for each of the first sign images to output the first cluster of selected areas up to the Nth cluster.

[0046] Next, based on the first cluster selection area up to the Nth cluster selection area, N colors different from the first color are used to select the first front view marker image, the first rear view marker image, the first left view marker image, the first right view marker image, and the first top view marker image, respectively, to obtain the second front view marker image, the second rear view marker image, the second left view marker image, the second right view marker image, and the second top view marker image. For example, the clustered first cluster selection area up to the Nth cluster selection area is re-labeled with N colors different from the first color (e.g., blue for cluster 1, green for cluster 2), generating the second front view, rear view, left view, right view, and top view marker images, where images labeled with the same color represent the same deviation area.

[0047] Finally, using the second front-view sign image, the second rear-view sign image, the second left-view sign image, the second right-view sign image, and the second top-view sign image as supervision, and the first front-view sign image, the first rear-view sign image, the first left-view sign image, the first right-view sign image, and the first top-view sign image as input, a deviation region union analysis model is trained. This deviation region union analysis model is built upon a convolutional neural network. The pre-trained model can achieve a mapping from the first sign image to the second sign image; that is, based on the local deviation of the single-color bounding box, it infers the true position of the bounding box in three-dimensional space and identifies the same positional deviation in different views.

[0048] For example, the deviation region union analysis model consists of five parts: an input layer, a feature extraction module, a union modeling module, and an output layer. Each part is connected hierarchically to achieve end-to-end computation from image input to deviation union result. The input layer receives the first frontal, back, left, right, and top-view labeled images and performs standardization preprocessing (such as normalizing pixel values ​​to [-1, 1], adjusting the size to a fixed resolution, etc.) to ensure a uniform network input format. The feature extraction module performs multi-level semantic feature capture. Based on the CNN encoder architecture (such as ResNet, UNet encoder, or custom convolutional blocks), it performs low-level, mid-level, and high-level feature extraction layer by layer. The low-level feature extraction can use small-sized convolutional kernels (such as 3×3) and convolutional layers with a stride of 1, combined with the ReLU activation function, to capture basic geometric features such as edges, line segments, and corners in the image. A max pooling layer (pooling kernel 2×2, stride 2) is then superimposed to reduce the feature map resolution, expand the receptive field, and reduce the computational load. Mid-level feature extraction can increase the number of channels in convolutional layers (e.g., gradually increasing from 64 channels to 256 channels) to capture more complex structural features (such as component outlines and hole shapes) through dilated or dilated convolutions. High-level semantic extraction can employ global average pooling or adaptive pooling to compress feature maps into fixed-length semantic vectors, representing the overall structure in the image (such as the spatial layout of components like walls and pipes). The union modeling module identifies identical locations in different views through fully connected layers or attention mechanisms. The output layer outputs shape deviation regions with different color features.

[0049] For example, the supervised training process of the deviation region union analysis model can be implemented through the following technical path: taking the first frontal, rear, left, right, and top view marker images as input, and the second frontal, rear, left, right, and top view marker images as supervision, and performing supervised training based on the mapping rule that overlapping spatial coordinates must be the same entity.

[0050] In summary, compared to existing technologies, this application obtains multi-angle projection images of the initial BIM model of the target component and compares the shapes of these images with the images entered in the 2D drawings from the same viewpoint to identify the shape deviation areas. Based on this, the missing 2D design parameters of the component are retrieved. Thus, through deviation area union analysis, multi-view union analysis is performed, accurately locating the shape deviation areas and providing a reliable data foundation for subsequent operation and maintenance management.

[0051] S20: The modeling complexity of the two-dimensional design parameters of the missing component is sorted from simple to complex to obtain a sequence of design parameters;

[0052] The longer a design parameter takes to model, the higher its modeling complexity tends to be (e.g., curve parameters usually take longer than straight line parameters). Therefore, based on historical records, the modeling processing sequence can be obtained by establishing a correlation between different design parameters and modeling complexity.

[0053] To address the aforementioned issues, this application quantifies the modeling complexity of missing component two-dimensional design parameters using historical modeling duration data, and sorts them from simple to complex to obtain a sequence of design parameters.

[0054] Specifically, step S20 in the method includes:

[0055] Extract the first two-dimensional design parameters from the two-dimensional design parameters of the missing component;

[0056] Based on the first two-dimensional design parameters, retrieve a set of BIM modeling time records for similar two-dimensional design parameters;

[0057] Box plot analysis was performed on the set of BIM modeling time records to obtain non-discrete BIM modeling time records.

[0058] The mode of the recorded non-discrete BIM modeling time is statistically analyzed to obtain the modeling complexity of the first two-dimensional design parameters, and then added to the modeling complexity set.

[0059] The two-dimensional design parameters of the missing components are sorted in ascending order of the modeling complexity set to obtain the design parameter sequence.

[0060] In this embodiment, a first two-dimensional design parameter is first extracted from the two-dimensional design parameters of the missing component. For example, one parameter is selected from the two-dimensional design parameters of the missing component (such as length L=5m, angle θ=30°, arc radius R=2m) as the first two-dimensional design parameter (e.g., length L=5m). Similarly, by traversing all the two-dimensional design parameters of the missing component, each parameter will be analyzed as the first two-dimensional design parameter in turn.

[0061] Secondly, based on the first two-dimensional design parameters, a set of BIM modeling time records for similar two-dimensional design parameters is retrieved. Specifically, based on the parameter type of the first two-dimensional design parameters (such as length, angle, radius, etc.), a set of modeling time records for the same type of parameters is retrieved from the historical BIM modeling database. For example, based on "length L=5m", the modeling time of all "length" type parameters in the historical records is retrieved to form a set of modeling time records: [20min, 25min, 18min, 45min, 20min].

[0062] Next, box plot analysis is performed on the BIM modeling time record set to obtain non-discrete BIM modeling time record values. Box plot analysis is an intuitive and effective method for graphically representing statistical data, visually presenting the distribution characteristics and dispersion of data to help identify outliers. In the box plot, the box is defined by the lower quartile (Q1), median (Q2), and upper quartile (Q3), and the range of the box is the interquartile range (IQR), reflecting the middle 50% dispersion of the data; the upper and lower bands extend to the minimum and maximum non-outlier values, respectively, and data points outside the band range are identified as outliers. This application identifies and removes abnormal modeling time data by calculating key statistics such as IQR, avoiding interference from extreme values ​​caused by human error or model anomalies, and retaining time data that truly reflects normal modeling conditions. For example, the set of modeling time records, after being sorted, is [18min, 20min, 20min, 25min, 45min]. Then, Q1 (lower quartile, i.e., 25%) is 20min, Q3 (upper quartile, i.e., 75%) is 25min, IQR (interquartile range) is Q3-Q1=5min, and the discrete value threshold is Q1-1.5×IQR=20-1.5*5=12.5min or Q3+1.5×IQR=25+1.5*5=32.5min. The effective range is 12.5min~32.5min. In the example, 45min exceeds 32.5min and is judged as a discrete value. After removing it, the remaining values ​​are [18min, 20min, 20min, 25min], which are used as non-discrete BIM modeling time records.

[0063] Furthermore, the mode of the recorded non-discrete BIM modeling time is statistically analyzed to obtain the modeling complexity of the first two-dimensional design parameter, which is then added to the modeling complexity set. For example, the mode of the recorded non-discrete BIM modeling time [18min, 20min, 20min, 25min] is statistically analyzed (the mode is 20min) to obtain the modeling complexity of the first two-dimensional design parameter. That is, the time corresponding to the mode is used as the modeling complexity of that parameter and added to the modeling complexity set (e.g., "length L=5m" corresponds to 20min). The mode reflects the modeling time in most cases, avoiding the mean being affected by extreme data, and is more representative.

[0064] Finally, the two-dimensional design parameters of the missing components are sorted from smallest to largest according to the set of modeling complexity to obtain the design parameter sequence. Specifically, the modeling complexity of all two-dimensional design parameters of the missing components is sorted from smallest to largest to obtain the design parameter sequence. For example, if the modeling complexity of "angle θ=30°" is 15 minutes, the modeling complexity of "arc radius R=2m" is 30 minutes, and the modeling complexity of "length L=5m" is 20 minutes, then the sorting is: angle θ=30° (15 minutes), length L=5m (20 minutes), arc radius R=2m (30 minutes), obtaining the design parameter sequence. In this way, parameters are processed according to the principle of modeling from simple to complex, reducing blockages in the modeling process and improving efficiency.

[0065] In summary, compared with existing technologies, this application quantifies the modeling complexity of missing component 2D design parameters by using historical modeling time data, and sorts them from simple to complex to obtain a sequence of design parameters. In this way, a processing sequence from simple to complex is generated, which optimizes the BIM modeling process and improves efficiency.

[0066] S30: Construct an index binary tree according to the design parameter sequence;

[0067] During the model building process, simple design parameters should be prioritized, such as the basic dimensions (length, width, height) and basic shapes (rectangle, circle) of components. Following a logic of progressing from easy to difficult in a step-by-step manner can effectively reduce modeling complexity and improve overall work efficiency.

[0068] To address the aforementioned issues, this application constructs an index binary tree from easy to difficult according to the design parameter sequence.

[0069] Specifically, such as Figure 2 As shown, step S30 in the method includes:

[0070] Extract the first sequence design parameter from the design parameter sequence and construct the first-level node of the binary tree;

[0071] Until the Nth design parameter is extracted from the design parameter sequence, an N-level binary tree node is constructed;

[0072] The indexed binary tree is constructed based on the first-level nodes of the binary tree up to the Nth-level nodes.

[0073] In the embodiments of this application, such as Figure 2 As shown, utilizing the hierarchical indexing characteristics of binary trees, simple parameters (low complexity) are used as first-level nodes, and complex parameters (high complexity) are used as N-level nodes, forming a "from easy to difficult" retrieval path. For example, first determine the basic dimensions of the component, such as length and width (first-level nodes), and then retrieve complex parameters such as material and chamfer (second-level or N-level nodes). Specifically:

[0074] First, extract the first-order design parameter (the parameter with the lowest modeling complexity) from the design parameter sequence and construct the first-level node of the binary tree (i.e., the root node of the binary tree). For example, if the design parameter sequence is: angle θ=30° (15min), length L=5m (20min), and arc radius R=2m (30min), then angle θ=30° is used as the first-level node and length L=5m is used as the second-level node.

[0075] Next, the design parameter with the Nth index is extracted from the design parameter sequence to construct an N-level node of the binary tree. Specifically, the i-th design parameter (i=3,...,N) is used as the i-th level node of the binary tree according to the design parameter sequence until an N-level node of the binary tree is constructed. Each node belongs to only one level, and the level number is equal to its index in the design parameter sequence (e.g., the 3rd parameter is a level 3 node).

[0076] Finally, based on the first-level nodes of the binary tree up to the N-level nodes, the index binary tree is constructed, thus forming a binary tree of depth N. Each node corresponds to a design parameter, and the hierarchical relationship directly reflects the order of the parameters in the design parameter sequence (from simple to complex). The first-level node (root node) is the most basic parameter. The higher the level, the higher the modeling complexity. The path from the root node to any leaf node corresponds to the complete modeling order from basic parameters to complex parameters.

[0077] In summary, compared with existing technologies, this application constructs an index binary tree according to the sequence of design parameters, from easy to difficult, so as to quickly match the parameters in the modeling material library through hierarchical retrieval, thereby improving the efficiency of model retrieval.

[0078] S40: Based on the design parameters of the first-level nodes of the indexed binary tree, retrieve the first-level matching BIM modeling parameters from the modeling material library. When the first-level matching BIM modeling parameters are not empty, retrieve the second-level matching BIM modeling parameters from the modeling material library based on the design parameters of the second-level nodes of the indexed binary tree, until N-level matching BIM modeling parameters are obtained, where N represents the number of design parameters.

[0079] In this embodiment, the index binary tree is constructed from simple to complex according to the design parameters, with each node corresponding to one design parameter. The first-level nodes store the most basic and simplest design parameters. Therefore, the retrieval process is guided by the hierarchical structure of the binary tree, starting from the first-level nodes and proceeding sequentially layer by layer. This approach ensures that basic parameters are obtained first, followed by more complex parameters. Specifically:

[0080] Based on the design parameters of the first-level nodes of the indexed binary tree, first-level matching BIM modeling parameters are retrieved from the modeling material library. For example, if the first-level node parameter is "a wall is 5 meters long", the library is searched for wall model parameters containing this length information. When the first-level matching BIM modeling parameter is not empty (i.e., a corresponding parameter is found, indicating that the material library contains parameters that meet the basic requirements), the library retrieves second-level matching BIM modeling parameters based on the design parameters of the second-level nodes of the indexed binary tree, until N levels of matching BIM modeling parameters are obtained, where N represents the number of design parameters.

[0081] For example, based on the design parameters of the first-level nodes of the indexed binary tree, after retrieving the first-level matching BIM modeling parameters (such as wall length), and after indexing the design parameters of the second-level nodes (such as thickness), more complex parameters such as wall surface texture and the location of door and window openings are retrieved.

[0082] In summary, compared to existing technologies, this application avoids blind traversal of the modeling resource library through hierarchical retrieval. The ordered structure of a binary tree allows for rapid location of required parameters, reducing retrieval time and improving efficiency. This is particularly beneficial in large and complex projects with a vast number of parameters, significantly reducing retrieval complexity. Furthermore, strictly adhering to a parameter retrieval order from basic to complex ensures the completeness and logical consistency of parameters during model construction, preventing model errors caused by parameter omissions or mismatches, and improving the quality of the BIM model.

[0083] S50: Send the first-level matching BIM modeling parameters, the second-level matching BIM modeling parameters, and so on up to the N-level matching BIM modeling parameters to the construction management terminal for construction operation and maintenance control.

[0084] The aforementioned steps match BIM modeling parameters using an indexed binary tree, and this data can be sent to the construction management terminal for construction operation and maintenance control. Furthermore, the BIM model of a large-scale engineering project may contain tens of thousands to hundreds of thousands of parameters, and the index matching process may encounter situations where there are no model combinations in the resource library that provide complete parameter matching.

[0085] To address the aforementioned issues, this application sends first-level matching BIM modeling parameters, second-level matching BIM modeling parameters, and up to N-level matching BIM modeling parameters to the construction management terminal for construction operation and maintenance control. If there is no complete parameter matching model combination in the material library, the fitness is analyzed through partial matching, and compensatory matching modeling is performed.

[0086] Specifically, step S50 in the method includes:

[0087] Obtain a first BIM material model combination up to the Qth BIM material model combination that has all the first-level matching BIM modeling parameters, the second-level matching BIM modeling parameters, up to the Nth-level matching BIM modeling parameters;

[0088] Extract the first BIM material model combination until the minimum number of combinations of the Qth BIM material model combination is reached, set as the target BIM material model combination, and send it to the construction management terminal for construction operation and maintenance control.

[0089] In this embodiment, a first BIM material model combination up to the Qth BIM material model combination, which has all the first-level matching BIM modeling parameters, the second-level matching BIM modeling parameters, and so on up to the Nth-level matching BIM modeling parameters, is first obtained. Specifically, a BIM material model combination that simultaneously contains all matching BIM modeling parameters is retrieved from the modeling material library to obtain the first to the Qth combinations. Each of these combinations can meet the design parameter requirements, but the number of sub-models it contains varies. For example, if the modeling parameters are "length L + material M + surface texture T", the fully matching combination may include: combination 1 (containing a single model integrating L, M, and T), and combination 2 (containing three independent models corresponding to L, M, and T respectively).

[0090] Secondly, the minimum number of combinations from the first BIM material model combination to the Qth BIM material model combination is extracted and set as the target BIM material model combination, which is then sent to the construction management terminal for construction operation and maintenance control. Specifically, the number of sub-models contained in each of the first to Qth BIM material model combinations is calculated, and the combination with the fewest sub-models is selected as the target combination. This is because a smaller number of combinations means a more compact and simpler model structure, lower data redundancy, reduced assembly complexity between models, avoids errors caused by multi-model collaboration, and facilitates rapid loading and parsing by the construction management terminal. For example, combination 1 (containing a single model integrating L, M, and T) and combination 2 (containing three independent models corresponding to L, M, and T respectively) are both superior. Therefore, combination 1 (containing 1 model) is better than combination 2 (containing 3 models), indicating that combination 1 does not require assembly compared to combination 2. Therefore, combination 1 is selected as the target BIM material model combination and sent to the construction management terminal for construction operation and maintenance control.

[0091] Specifically, step S50 of the method further includes:

[0092] When the number of BIM material model combinations having all the first-level matching BIM modeling parameters, the second-level matching BIM modeling parameters up to the Nth-level matching BIM modeling parameters is equal to 0, a number of BIM material model combinations having some of the first-level matching BIM modeling parameters, the second-level matching BIM modeling parameters up to the Nth-level matching BIM modeling parameters are obtained.

[0093] The modeling parameters of the several BIM material model combinations are summed in series to obtain the fitness of the several BIM material model combinations.

[0094] Extract the maximum value of the fitness of the several BIM material model combinations to obtain the target BIM material model combination to be compensated, and send it to the construction management terminal for construction operation and maintenance control in combination with the BIM modeling parameters to be compensated.

[0095] In this embodiment of the application, when there are no model combinations with completely matching parameters in the material library (i.e., the number of model combinations is equal to 0), the solution closest to the design requirements is selected through fitness evaluation of partially matching combinations, and then compensated for by matching BIM modeling parameters. Specifically:

[0096] First, when the number of BIM material model combinations having all the first-level matching BIM modeling parameters, the second-level matching BIM modeling parameters, up to the Nth-level matching BIM modeling parameters is equal to 0, several BIM material model combinations having some of the first-level matching BIM modeling parameters, the second-level matching BIM modeling parameters, up to the Nth-level matching BIM modeling parameters are obtained. For example, a material library is searched for model combinations containing at least one matching BIM modeling parameter, forming several partially matching BIM material model combinations. For instance, if the modeling parameter is "length L + angle θ + radius R", the partially matching BIM material model combinations might be: combination A (e.g., matching length L + angle θ), combination B (e.g., matching length L + radius R).

[0097] Secondly, the modeling parameters of the several BIM material model combinations are summed hierarchically to obtain the fitness of the BIM material model combinations. Specifically, the hierarchy of parameters in the indexed binary tree is used to sum the modeling parameters of several BIM material model combinations, and this summation is used as the fitness of the BIM material model combination. The higher the fitness, the better the combination. For example, if the indexed binary tree obtains: combination A (e.g., matching length L + angle θ) has length at level one and angle at level two, and combination B (e.g., matching length L + radius R) has length at level one and radius at level three, then the fitness of the BIM material model combination of combination A = 1 + 2 = 3, and the fitness of the BIM material model combination of combination B = 1 + 3 = 4, indicating that combination B is better.

[0098] Finally, the maximum fitness value of the several BIM material model combinations is extracted to obtain the target BIM material model combination to be compensated. This combination, along with the BIM modeling parameters to be compensated, is then sent to the construction management terminal for construction operation and maintenance control. Extracting the maximum fitness value of the several BIM material model combinations to obtain the target BIM material model combination to be compensated is to ensure that high-complexity parameters (such as the third-level radius R) are preferentially matched through hierarchical weighting, reducing the risk of deviations in critical structures and lowering the difficulty of compensation modeling. For example, the combination corresponding to the maximum fitness value of the BIM material model combination is extracted as the target BIM material model combination to be compensated (e.g., combination B). For unmatched parameters (such as the missing angle θ in combination B), supplementation is done through manual modeling, temporary parameter substitution, or design change processes to form the BIM modeling parameters to be compensated, which are then sent to the construction management terminal for construction operation and maintenance control.

[0099] In summary, compared with existing technologies, this application sends first-level matching BIM modeling parameters, second-level matching BIM modeling parameters, and up to N-level matching BIM modeling parameters to the construction management end for construction operation and maintenance control. If there is no model combination with complete parameter matching in the material library, the fitness is analyzed through partial matching, and compensatory matching modeling is performed. In this way, the construction management end can quickly obtain lightweight data and carry out construction operation and maintenance control accordingly, improving the efficiency and accuracy of operation and maintenance control.

[0100] In summary, the embodiments of this application have at least the following technical effects:

[0101] Compared to existing technologies, this application first obtains multi-angle projection images of the initial BIM model of the target component, compares the image shapes with the images entered in the two-dimensional drawings from the same viewpoint to obtain the shape deviation area, and retrieves the two-dimensional design parameters of the missing component accordingly. Thus, through deviation area union analysis, multi-view union analysis is performed to accurately locate the shape deviation area, providing a reliable data foundation for subsequent operation and maintenance management.

[0102] Secondly, this application quantifies the modeling complexity of missing component 2D design parameters by using historical modeling time data, and sorts them from simple to complex to obtain a sequence of design parameters. In this way, a processing sequence from simple to complex is generated, which optimizes the BIM modeling process and improves efficiency.

[0103] Furthermore, this application constructs an index binary tree according to the sequence of design parameters, from easy to difficult, so as to quickly match the parameters in the modeling material library through hierarchical retrieval, thereby improving the efficiency of model retrieval.

[0104] Furthermore, this application avoids blindly traversing the modeling resource library through hierarchical retrieval. The ordered structure of a binary tree allows for rapid location of required parameters, reducing retrieval time and improving efficiency. This is particularly beneficial in large and complex projects where the number of parameters is enormous; this retrieval method significantly reduces retrieval complexity. Moreover, strictly adhering to a parameter retrieval order from basic to complex ensures the completeness and logical consistency of parameters during model construction, preventing model errors caused by parameter omissions or mismatches, and improving the quality of the BIM model.

[0105] Finally, this application sends the first-level matching BIM modeling parameters, the second-level matching BIM modeling parameters, and so on up to the Nth-level matching BIM modeling parameters to the construction management terminal for construction operation and maintenance control. If there is no model combination with complete parameter matching in the material library, the fitness is analyzed through partial matching, and compensatory matching modeling is performed. In this way, the construction management terminal can quickly obtain lightweight data and carry out construction operation and maintenance control accordingly, thereby improving the efficiency and accuracy of operation and maintenance control.

[0106] Through the above technical solution, this application determines the deviation area by comparing images from five views. Then, the design parameters for the deviation area are sorted based on modeling complexity. Based on this, existing modeling parameters are retrieved from the material library and sent to the construction management terminal. In this way, accurate BIM data support is provided for the construction site, realizing intelligent operation and maintenance management and avoiding the inefficiency caused by global comparison and the errors caused by human experience. Thus, the efficiency and accuracy of construction operation and maintenance are improved.

[0107] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0108] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0109] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0110] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0111] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0112] Although preferred embodiments of the invention have been described, those skilled in the art, once they have learned the basic inventive concept, can make other changes and modifications to these embodiments.

[0113] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this invention and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. A smart construction operation and maintenance management method based on BIM technology, characterized in that, include: Obtain multi-angle projection images of the initial BIM model of the target component, compare them with the images entered in the two-dimensional drawings, and obtain the two-dimensional design parameters of the missing component. The modeling complexity of the two-dimensional design parameters of the missing component is sorted from simple to complex to obtain a sequence of design parameters. Construct an index binary tree according to the design parameter sequence; Based on the design parameters of the first-level nodes of the indexed binary tree, first-level matching BIM modeling parameters are retrieved from the modeling material library. When the first-level matching BIM modeling parameters are not empty, second-level matching BIM modeling parameters are retrieved from the modeling material library based on the design parameters of the second-level nodes of the indexed binary tree, until N-level matching BIM modeling parameters are obtained, where N represents the number of design parameters. The first-level matching BIM modeling parameters, the second-level matching BIM modeling parameters, and so on up to the Nth-level matching BIM modeling parameters are sent to the construction management terminal for construction operation and maintenance control. This involves obtaining multi-angle projection images of the initial BIM model of the target component, comparing their shapes with the images entered in the 2D drawings, and obtaining the 2D design parameters of the missing component, including: Images are entered from the two-dimensional drawings, and the images of the front view, rear view, left view, right view, and top view are extracted. Based on the construction positioning parameters, after locating the initial BIM model of the target component, frontal projection image, rear projection image, left projection image, right projection image and top projection image are acquired. The front view paper image, the rear view paper image, the left view paper image, the right view paper image, and the top view paper image are compared with the front view projection image, the rear view projection image, the left view projection image, the right view projection image, and the top view projection image from the same viewing angle to obtain the shape deviation area. Based on the shape deviation area and the image entered from the two-dimensional drawing, the two-dimensional design parameters of the missing component are retrieved.

2. The method as described in claim 1, characterized in that, The front view image, rear view image, left view image, right view image, and top view image are compared with the front projection image, rear projection image, left projection image, right projection image, and top projection image from the same viewing angle to obtain the shape deviation area, including: The front view paper image, the rear view paper image, the left view paper image, the right view paper image, and the top view paper image are compared with the front view projection image, the rear view projection image, the left view projection image, the right view projection image, and the top view projection image from the same viewing angle to obtain the front view paper deviation area, the rear view paper deviation area, the left view paper deviation area, the right view paper deviation area, and the top view paper deviation area. By using a pre-trained deviation region union analysis model, the deviation regions of the front view paper, the rear view paper, the left view paper, the right view paper, and the top view paper are processed to obtain the shape deviation regions. Among them, deviation regions that are marked with the same color in different drawings are the same deviation region.

3. The method as described in claim 2, characterized in that, The construction process of the deviation region union analysis model is as follows: Extract the front view recording image, rear view recording image, left view recording image, right view recording image and top view recording image of BIM modeling, use the first color to randomly select local areas, and obtain the first front view marker image, the first rear view marker image, the first left view marker image, the first right view marker image and the first top view marker image; Based on the BIM modeling, the first color-selected area is clustered at the same location to obtain the first cluster of selected areas up to the Nth cluster of selected areas; Based on the first cluster selection area up to the Nth cluster selection area, N colors different from the first color are used to select the first front view sign image, the first rear view sign image, the first left view sign image, the first right view sign image and the first top view sign image respectively, to obtain the second front view sign image, the second rear view sign image, the second left view sign image, the second right view sign image and the second top view sign image; Using the second front-view sign image, the second rear-view sign image, the second left-view sign image, the second right-view sign image, and the second top-view sign image as supervision, and the first front-view sign image, the first rear-view sign image, the first left-view sign image, the first right-view sign image, and the first top-view sign image as input, a deviation region union analysis model is trained.

4. The method as described in claim 1, characterized in that, A complexity analysis of the modeling of the two-dimensional design parameters of the missing component is performed to obtain a sequence of design parameters, including: Extract the first two-dimensional design parameters from the two-dimensional design parameters of the missing component; Based on the first two-dimensional design parameters, retrieve a set of BIM modeling time records for similar two-dimensional design parameters; Box plot analysis was performed on the set of BIM modeling time records to obtain non-discrete BIM modeling time records. The mode of the recorded non-discrete BIM modeling time is statistically analyzed to obtain the modeling complexity of the first two-dimensional design parameters, and then added to the modeling complexity set. The two-dimensional design parameters of the missing components are sorted in ascending order of the modeling complexity set to obtain the design parameter sequence.

5. The method as described in claim 1, characterized in that, Construct an indexed binary tree according to the design parameter sequence, including: Extract the first sequence design parameter from the design parameter sequence and construct the first-level node of the binary tree; Until the Nth design parameter is extracted from the design parameter sequence, an N-level binary tree node is constructed; The indexed binary tree is constructed based on the first-level nodes of the binary tree up to the Nth-level nodes.

6. The method as described in claim 1, characterized in that, The first-level matching BIM modeling parameters, the second-level matching BIM modeling parameters, and so on up to the Nth-level matching BIM modeling parameters are sent to the construction management terminal for construction operation and maintenance control, including: Obtain a first BIM material model combination up to the Qth BIM material model combination that has all the first-level matching BIM modeling parameters, the second-level matching BIM modeling parameters, up to the Nth-level matching BIM modeling parameters; Extract the first BIM material model combination until the minimum number of combinations of the Qth BIM material model combination is reached, set as the target BIM material model combination, and send it to the construction management terminal for construction operation and maintenance control.

7. The method as described in claim 6, characterized in that, Also includes: When the number of BIM material model combinations having all the first-level matching BIM modeling parameters, the second-level matching BIM modeling parameters up to the Nth-level matching BIM modeling parameters is equal to 0, a number of BIM material model combinations having some of the first-level matching BIM modeling parameters, the second-level matching BIM modeling parameters up to the Nth-level matching BIM modeling parameters are obtained. The modeling parameters of the several BIM material model combinations are summed in series to obtain the fitness of the several BIM material model combinations. Extract the maximum value of the fitness of the several BIM material model combinations to obtain the target BIM material model combination to be compensated, and send it to the construction management terminal for construction operation and maintenance control in combination with the BIM modeling parameters to be compensated.

8. An electronic device, characterized in that, include: Memory, used to store computer software programs; A processor is used to read and execute the computer software program, thereby realizing the intelligent construction operation and maintenance management method based on BIM technology as described in any one of claims 1-7.

9. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements a smart construction operation and maintenance management method based on BIM technology as described in any one of claims 1-7.

Images