A project status management method and system for intelligent point cloud modeling

Abstract

This invention discloses a project status management method and system for intelligent point cloud modeling, relating to the field of project status management technology. The method includes registering point cloud block files, label files, and 3D model files generated during point cloud segmentation, point cloud partitioning, geometric fitting, and pipeline parameter adjustment as subsequent versions and recording operation impact information; constructing a 3D spatial index structure to associate and store point cloud block files, pipeline objects, and corresponding operation impact information; and when a rollback request is received, determining the set of objects that intersect with the target 3D spatial range and whose semantic category belongs to the target semantic category set, and determining the set of affected objects whose version number is later than the target version identifier. This invention can, upon receiving a rollback request containing the target spatial range and the target semantic category set, only locate the affected local object set and perform a local rollback, which helps to preserve the modeling results of unaffected areas and reduces redundant processing and manual verification work.

Description

Technical Field

[0001] This invention relates to the field of project status management technology, and in particular to a project status management method and system for point cloud intelligent modeling. Background Technology

[0002] With the development of 3D measurement technologies such as laser scanning and structured light measurement, point cloud data is widely used for 3D modeling and operation and maintenance management in scenarios such as industrial plants, pipe corridors, and building spaces. A typical intelligent point cloud modeling process usually includes multiple steps such as: raw point cloud acquisition, point cloud preprocessing and segmentation, point cloud semantics and instance segmentation, geometric fitting, pipeline parameter adjustment, and model verification and output, which are often completed collaboratively by different software modules or algorithm components.

[0003] In existing technologies, to manage the aforementioned multi-step processing, engineering practices typically employ the following methods: One approach involves manually maintaining project files at different stages at the operating system folder level, distinguishing between processing stages through file naming or directory copying; another approach utilizes general version management tools or the built-in project management functions of point cloud processing software to create versions or snapshots of the overall project status, allowing the project to be restored to a historical version when needed. This type of approach generally enables backtracking of the point cloud processing process at the project or dataset level, and is suitable for saving and restoring the overall state.

[0004] However, in practical applications of point cloud intelligent modeling, a common scenario is that only a specific local spatial region or a certain type of semantic object (such as some pipelines or local equipment) needs to be rolled back or rebuilt, while other areas in the project have already met engineering requirements and should not be rolled back as well. If project-level rollback is still used, the modeling results of areas without problems may be rolled back entirely, increasing the workload of repeated calculations and manual verification. Simultaneously, the point cloud modeling process generates a large number of intermediate results; some can be obtained through recalculation, while others should not be discarded. How to reasonably balance storage overhead and recalculation overhead is also a problem that needs to be considered. Furthermore, in scenarios where multiple people collaborate to modify the pipeline network model, when different personnel make adjustments to the same pipeline or adjacent fittings on different branches, how to effectively identify and handle topological conflicts during version merging to ensure the connectivity and engineering rationality of the merged pipeline network model is also a practical need. Summary of the Invention

[0005] In view of the aforementioned existing problems, the present invention is proposed.

[0006] Therefore, this invention provides a project status management method for intelligent point cloud modeling to solve the problems of lack of fine-grained version management, partial rollback and resource optimization in point cloud modeling.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0008] In a first aspect, the present invention provides a project status management method for intelligent point cloud modeling, comprising: establishing a project directory for the intelligent point cloud modeling project in a storage medium; generating a project list file within the project directory; registering the original point cloud data as an initial version; and setting a current version pointer to point to the initial version in the project list file; during the process of point cloud segmentation, point cloud partitioning, geometric fitting, and pipeline parameter adjustment of the original point cloud data, registering the point cloud block files, label files, and 3D model files generated in each step as subsequent versions; and recording operation impact information including 3D spatial bounding boxes and semantic category masks for each version in the project list file; and constructing a system based on the spatial location of the point cloud block files and pipeline objects. A three-dimensional spatial index structure is used to associate and store the point cloud block file, the pipeline object, and the corresponding operation impact information. When a rollback request including a target version identifier, a target three-dimensional spatial range, and a target semantic category set is received, the set of objects that intersect with the target three-dimensional spatial range and whose semantic category belongs to the target semantic category set is determined according to the three-dimensional spatial index structure. The set of versions that affect the set of objects and whose version number is later than the target version identifier is determined in the project list file. According to the target version identifier and the version set, a partial rollback is performed on the set of objects to restore it to the state corresponding to the target version identifier. The restored state of the set of objects is registered as a new rollback version in the project list file.

[0009] As a preferred embodiment of the project status management method for intelligent point cloud modeling described in this invention, the step of performing a local rollback on the object set includes:

[0010] When the rollback mode is soft rollback, only the mapping relationship between the object set and the target version identifier in the project manifest file is updated to the target version identifier, without deleting any files in the version directory;

[0011] When the rollback mode is a hard rollback, while updating the mapping relationship, delete the local data files generated by the version set that do not belong to the version before the target version identifier, and delete the corresponding records in the project manifest file.

[0012] As a preferred embodiment of the project status management method for intelligent point cloud modeling described in this invention, the original point cloud data, the point cloud blocks, the label files and the three-dimensional model files are abstracted into multiple data nodes, and a directed acyclic data lineage graph is constructed with the input-output relationship of each processing step as the edge.

[0013] For each data node in the data lineage graph, record cost parameters including at least storage size and recalculation cost.

[0014] As a preferred embodiment of the project status management method for intelligent point cloud modeling described in this invention, wherein: in the data lineage graph, a target data node set is determined according to the object set, and the total cost value of each data node in the target data node set is calculated using two strategies: reading from existing storage and recalculating from the parent node;

[0015] Based on the total cost value, a dynamic programming algorithm or a greedy algorithm is used to determine the strategy combination for each target data node, so as to minimize the sum of the total cost values ​​of all target data nodes, and to perform reading or recalculation on the target data node according to the strategy combination.

[0016] As a preferred embodiment of the project status management method for point cloud intelligent modeling described in this invention, the pipeline model in the point cloud intelligent modeling result is represented as a graph structure containing nodes and edges, wherein nodes represent pipe connection points and edges represent pipe segments with attached pipe diameter and material properties;

[0017] For different version branches derived from the common ancestor version, the pipeline modification process of each branch relative to the common ancestor version is recorded as a graph editing sequence;

[0018] When performing a version merge between two branches, the graph editing sequence is replayed based on the graph structure corresponding to the common ancestor version to obtain the change sets of the two branches. The change sets are compared to identify mutually exclusive modifications to the same node or edge, and topological conflicts are determined based on the identification results.

[0019] As a preferred embodiment of the project status management method for intelligent point cloud modeling described in this invention, the following steps are performed: connectivity checks and preset process constraint checks are performed on the merged graph structure; if a disconnected pipe segment or illegal loop is detected, the graph editing operation that causes the disconnected pipe segment or illegal loop will be kept as a conflict marker.

[0020] As a preferred embodiment of the project status management method for intelligent point cloud modeling described in this invention, wherein: a processing flow including point cloud segmentation, point cloud division, geometric fitting and pipeline parameter adjustment is identified as a transaction, and the intermediate results generated by each processing step in the processing flow are written into a temporary version directory corresponding to the transaction;

[0021] When all processing steps in the process are successfully completed and pass the consistency check, the data in the temporary version directory will be migrated to the official version directory and registered as the new version in the project manifest file.

[0022] If any processing step in the process fails, the data in the temporary version directory is deleted and the transaction is undone.

[0023] Secondly, the present invention provides a project status management system for intelligent point cloud modeling, comprising: a project directory management unit, used to establish a project directory for the intelligent point cloud modeling project in a storage medium, generate a project list file in the project directory, register the original point cloud data as the initial version, and set the current version pointer to the initial version in the project list file; a version registration unit, used to register the point cloud block files, label files, and 3D model files generated by each processing step as subsequent versions during point cloud segmentation, point cloud partitioning, geometric fitting, and pipeline parameter adjustment of the original point cloud data, and record operation impact information including 3D spatial bounding boxes and semantic category masks for each version in the project list file; and a spatial index construction unit, used to construct a spatial index based on the spatial location of the point cloud block files and pipeline objects. A three-dimensional spatial index structure is provided, in which the point cloud block file, the pipeline object, and the corresponding operation impact information are associated and stored. A rollback determination unit is used to determine, upon receiving a rollback request including a target version identifier, a target three-dimensional spatial range, and a target semantic category set, a set of objects that intersect with the target three-dimensional spatial range and whose semantic category belongs to the target semantic category set according to the three-dimensional spatial index structure, and a set of versions that affect the set of objects and whose version number is later than the target version identifier in the project manifest file. A partial rollback execution unit is used to perform a partial rollback on the set of objects according to the target version identifier and the set of versions to restore it to the state corresponding to the target version identifier, and register the restored state of the set of objects as a new rollback version in the project manifest file.

[0024] Thirdly, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, it implements any step of the project status management method for intelligent point cloud modeling as described in the first aspect of the present invention.

[0025] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the project state management method for intelligent point cloud modeling as described in the first aspect of the present invention.

[0026] The beneficial effects of this invention are as follows: When receiving a rollback request containing the target three-dimensional spatial range and the target semantic category set, this invention can locate only the affected local object set and perform local rollback on the set, which is beneficial to retain the modeling results of the unaffected areas and reduce unnecessary repetitive processing and manual verification work; in addition, it can reduce the overall resource consumption while ensuring the rollback effect, which is beneficial to balancing storage consumption and computing time in large-scale point cloud projects. Attached Figure Description

[0027] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1 This is a flowchart of a project status management method used for intelligent point cloud modeling.

[0029] Figure 2 A flowchart for project initialization, version registration, and index building.

[0030] Figure 3 A flowchart for partial rollback, soft / hard rollback, and cost optimization.

[0031] Figure 4 This is a flowchart of branch merging and transaction processing in the pipeline model. Detailed Implementation

[0032] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0033] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0034] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0035] Reference Figures 1-4As one embodiment of the present invention, this embodiment provides a project status management method for intelligent point cloud modeling, including the following steps:

[0036] S1. Establish a project directory for the point cloud intelligent modeling project in the storage medium, generate a project list file in the project directory, register the original point cloud data as the initial version, and set the current version pointer to point to the initial version in the project list file.

[0037] Specifically, the process begins by first concatenating the root directory path of the storage medium with the name of the intelligent point cloud modeling project to generate the project directory path, and then creating the project directory for the intelligent point cloud modeling project within the storage medium. A project manifest file with a fixed filename is then created within the project directory. Next, the original point cloud data file corresponding to the original point cloud data path is read and copied to the original point cloud data subdirectory within the project directory, with attributes such as file size and checksum hash value recorded. A unique version identifier is generated in memory, and an initial version directory corresponding to the unique version identifier is created in the version subdirectory within the project directory. A version record containing the unique version identifier, the original point cloud data storage path, and the original point cloud data attributes is written into the project manifest file. Finally, a current version pointer field is set in the project manifest file, with the value being the unique version identifier.

[0038] S2. During the process of point cloud segmentation, point cloud division, geometric fitting and pipeline parameter adjustment of the original point cloud data, the point cloud block files, label files and three-dimensional model files generated in each step are registered as subsequent versions, and the operation impact information including three-dimensional spatial bounding boxes and semantic category masks is recorded for each version in the project list file.

[0039] Furthermore, the project manifest file is read and parsed to obtain the version identifier corresponding to the current version pointer and the path of the original point cloud data file associated with the version identifier; based on the original point cloud data file path, the original point cloud data is split into multiple point cloud block files; a new version directory is created in the version subdirectory in the project directory of the point cloud intelligent modeling project, and a new version identifier is generated.

[0040] Store the point cloud block files in a new version directory, and calculate the 3D spatial bounding box and the corresponding semantic category mask fragment for each point cloud block file; add a version record to the project manifest file, and write the new version identifier, the relative path of the point cloud block file, and the operation impact information composed of the 3D spatial bounding box and the semantic category mask fragment in the version record.

[0041] Read the point cloud block file in the version directory pointed to by the current version pointer, perform point cloud segmentation to generate a label file, calculate the 3D spatial bounding box and semantic category mask in the point cloud space covered by the label file, create a new version directory to write the label file, register the operation impact information including version identifier, relative path of label file, 3D spatial bounding box and semantic category mask in the project manifest file, and update the current version pointer.

[0042] Read the point cloud block file and label file in the version directory pointed to by the current version pointer, perform geometric fitting and pipeline parameter adjustment to generate a 3D model file, calculate the 3D spatial bounding box and semantic category mask for the spatial range covered by the 3D model file, create a new version directory and write the 3D model file, register the operation impact information including version identifier, relative path of 3D model file, 3D spatial bounding box and semantic category mask in the project manifest file, and update the current version pointer again.

[0043] S3. Construct a three-dimensional spatial index structure based on the spatial location of the point cloud block file and the pipeline object, and associate and store the point cloud block file, the pipeline object and the corresponding operation impact information.

[0044] Specifically, the process iterates through all point cloud blocks and pipeline objects registered in the project manifest file in memory, reads the 3D bounding box parameters corresponding to each point cloud block and pipeline object, and determines the global 3D bounding box covering all point cloud blocks and pipeline objects. Using the global 3D bounding box as the root node, spatial regions are recursively divided according to preset partitioning rules to construct an octree-like 3D spatial index structure based on 3D spatial coordinates. The 3D bounding boxes of each point cloud block and pipeline object are sequentially inserted into the 3D spatial index structure. The leaf nodes intersecting with the 3D bounding boxes store the point cloud block identifier, pipeline object identifier, and location references to the operation impact information in the project manifest file. The node hierarchy of the 3D spatial index structure and the point cloud block identifier, pipeline object identifier, and operation impact information location references stored within the nodes are serialized to generate a 3D spatial index structure file, which is stored in the project directory of the point cloud intelligent modeling project. The path of the 3D spatial index structure file is registered in the project manifest file, thereby achieving the associated storage of point cloud blocks, pipeline objects, and corresponding operation impact information in the 3D spatial index structure.

[0045] It should be noted that after obtaining the global 3D bounding box, the maximum span L in the three-axis direction is calculated, with the center of the global 3D bounding box as the center point; the root node space is determined as a cube with a side length of L; the number of all point cloud blocks and all pipeline objects is counted, and the average volume ratio and average density occupied by each object are calculated. Based on this, the maximum number of objects that a single spatial node can contain and the maximum number of allowed partitioning layers are determined in memory; for example, the "maximum number of point cloud blocks and pipeline objects that each node can contain" and the "maximum octree depth" are automatically calculated based on the total number of objects and the global volume.

[0046] After the above parameters are determined, the rule that "the root node covers a cube with a side length of L. If the number of point cloud blocks and pipeline objects contained in the node exceeds the number threshold and the current depth has not reached the maximum number of division layers, then the node is divided into eight sub-regions along the midpoint of the three coordinate axes; otherwise, the division is stopped" is fixed as the preset division rule for this point cloud intelligent modeling project.

[0047] S4. When a rollback request including the target version identifier, the target three-dimensional spatial range, and the target semantic category set is received, the set of objects that intersect with the target three-dimensional spatial range and whose semantic category belongs to the target semantic category set is determined according to the three-dimensional spatial index structure, and the set of versions that affect the set of objects and whose version number is later than the target version identifier is determined in the project list file.

[0048] Furthermore, the system receives rollback requests through an interactive interface. These requests sequentially define the target version identifier, the target 3D spatial extent, and the target semantic category set. Specifically, a version list parsed from the project manifest file is displayed next to the 3D view. Each row in the version list corresponds to a version identifier and a version creation time. When a user clicks on a row in the version list, the version identifier corresponding to that row is used as the target version identifier and temporarily stored in the rollback request structure in memory. The system receives the spatial region selected by the user through selection, dragging, or inputting coordinate ranges in the 3D view. Using the 3D coordinates of all visible points within the selected spatial region as input, the system calculates the minimum bounding box aligned to the bounding axis of this point cloud in the 3D coordinate system. The minimum and maximum values ​​of the three axes of this bounding box are used as the numerical representation of the target 3D spatial extent and written into the rollback request structure. The system also lists all semantic category names registered in the project manifest file and tag file in the interface sidebar. It receives one or more semantic category names selected by the user via checkboxes, assembles all selected semantic category names into a target semantic category set, and writes it into the rollback request structure.

[0049] Furthermore, upon receiving a rollback request that includes the target version identifier, the target 3D spatial extent, and the target semantic category set, a spatial query is performed in the 3D spatial index structure to locate the leaf node intersecting with the target 3D spatial extent. Point cloud block identifiers and pipeline object identifiers are read from the leaf nodes. Based on the point cloud block identifiers and pipeline object identifiers, their respective semantic category information is read from the 3D spatial index structure or the project manifest file. Only point cloud blocks and pipeline objects whose semantic categories belong to the target semantic category set are retained to form an object set. Subsequently, based on each point cloud block identifier and pipeline object identifier in the object set, the version records arranged in chronological order are traversed in the project manifest file to find version records whose operation impact information contains any identifier of the object set and whose version number is greater than the target version identifier. Version identifiers that meet the conditions are added to the version set.

[0050] S5. Based on the target version identifier and the version set, perform a partial rollback on the object set to restore it to the state corresponding to the target version identifier, and register the restored object set state as a new rollback version in the project manifest file.

[0051] Furthermore, when the rollback mode is soft rollback, the mapping relationship between the object set and the target version identifier is only updated to the target version identifier in the project manifest file, without deleting any files in the version directory;

[0052] When the rollback mode is a hard rollback, while updating the mapping relationship, delete the local data files generated by the version set that do not belong to the version before the target version identifier, and delete the corresponding records in the project manifest file.

[0053] The original point cloud data, the point cloud blocks, the label file, and the 3D model file are abstracted into multiple data nodes, and a directed acyclic data lineage graph is constructed with the input-output relationship of each processing step as the edge.

[0054] For each data node in the data lineage graph, record cost parameters including at least storage size and recalculation cost.

[0055] In the data lineage graph, a target data node set is determined based on the object set, and the total cost value of each data node in the target data node set is calculated using two strategies: reading from existing storage and recalculating from the parent node.

[0056] Based on the total cost value, a dynamic programming algorithm or a greedy algorithm is used to determine the strategy combination for each target data node, so as to minimize the sum of the total cost values ​​of all target data nodes, and to perform reading or recalculation on the target data node according to the strategy combination.

[0057] Preferably, the pipeline network model in the point cloud intelligent modeling result is represented as a graph structure containing nodes and edges, where nodes represent pipe connection points and edges represent pipe segments with pipe diameter and material properties; wherein, the model object is preferably a pipeline network model, but can also be a structural model such as a steel structure model, and the present invention does not limit the type of model object.

[0058] For different version branches derived from the common ancestor version, the pipeline modification process of each branch relative to the common ancestor version is recorded as a graph editing sequence;

[0059] When performing a version merge between two branches, the graph editing sequence is replayed based on the graph structure corresponding to the common ancestor version to obtain the change sets of the two branches. The change sets are compared to identify mutually exclusive modifications to the same node or edge, and topological conflicts are determined based on the identification results.

[0060] The merged graph structure is subjected to connectivity checks and preset process constraint checks. If a disconnected pipe segment or illegal loop is detected, the graph editing operation of the disconnected pipe segment or illegal loop will be marked as a conflict for manual processing.

[0061] A single processing flow, including point cloud segmentation, point cloud division, geometric fitting, and pipeline parameter adjustment, is identified as a transaction. The intermediate results generated by each processing step in the processing flow are written to the temporary version directory corresponding to the transaction.

[0062] When all processing steps in the process are successfully completed and pass the consistency check, the data in the temporary version directory will be migrated to the official version directory and registered as the new version in the project manifest file.

[0063] If any processing step in the process fails, the data in the temporary version directory is deleted and the transaction is undone.

[0064] The operational impact information recorded in the project list file also includes the identifier of the processing step that generated the operation and the processing algorithm parameters.

[0065] This embodiment also provides a project status management system for intelligent point cloud modeling, including: a project directory management unit, used to establish a project directory for intelligent point cloud modeling projects in a storage medium, generate a project list file in the project directory, register the original point cloud data as the initial version, and set the current version pointer to the initial version in the project list file;

[0066] The version registration unit is used to register the point cloud block files, label files and 3D model files generated by each processing step as subsequent versions during the point cloud segmentation, point cloud division, geometric fitting and pipeline parameter adjustment of the original point cloud data, and to record the operation impact information including the 3D spatial bounding box and semantic category mask for each version in the project list file.

[0067] The spatial index construction unit is used to construct a three-dimensional spatial index structure based on the spatial location of the point cloud block file and the pipeline object, and to associate and store the point cloud block file, the pipeline object and the corresponding operation impact information in the three-dimensional spatial index structure.

[0068] The rollback determination unit is used to determine, upon receiving a rollback request including a target version identifier, a target three-dimensional spatial range, and a target semantic category set, a set of objects that intersect with the target three-dimensional spatial range and whose semantic category belongs to the target semantic category set according to the three-dimensional spatial index structure, and to determine, in the project manifest file, a set of versions that affect the set of objects and whose version number is later than the target version identifier.

[0069] The partial rollback execution unit is used to perform a partial rollback on the object set according to the target version identifier and the version set to restore it to the state corresponding to the target version identifier, and register the restored object set state as a new rollback version in the project manifest file.

[0070] This embodiment also provides a computer device applicable to a project status management method for intelligent point cloud modeling, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the project status management method for intelligent point cloud modeling as proposed in the above embodiment.

[0071] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0072] This embodiment also provides a storage medium storing a computer program. When executed by a processor, the program implements the project state management method for intelligent point cloud modeling as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0073] In summary, by recording the 3D bounding box and semantic category mask for each operation during version registration and constructing a 3D spatial index for point cloud blocks and pipeline objects, this invention can locate only the affected local object set and perform local rollback on that set when a rollback request containing the target 3D spatial extent and target semantic category set is received. Compared to methods that only support project-level or dataset-level rollback, this invention helps to preserve the modeling results of unaffected areas, reducing unnecessary repetitive processing and manual verification work.

[0074] This invention abstracts the original point cloud, segmented blocks, segmentation results, pipeline model, and parameter set into data nodes, constructs a directed acyclic data lineage graph, and records cost parameters such as storage size and recalculation cost for each node. During rollback, this invention selects a combination of strategies—either "directly reading existing results" or "recalculating from the parent node"—on the lineage graph based on the data nodes involved in the target object. This ensures effective rollback while minimizing overall resource overhead, facilitating a balance between storage consumption and computation time in large-scale point cloud projects.

[0075] This invention represents the pipeline network modeling results as a graph structure composed of nodes and edges, and records the modifications of different branches relative to the common ancestor version as a graph editing sequence. During branch merging, mutually exclusive modifications to the same node or edge are identified based on graph difference, and conflicts are marked for modifications that may affect pipeline connectivity or violate preset process constraints. If necessary, connectivity checks and loop checks are combined to provide prompts. Therefore, in multi-person collaborative scenarios, this helps maintain the consistency of the pipeline network model structure and the engineering rationality when merging versions.

[0076] The solution of this invention centers on methodological steps, using project directory structure, project list files, spatial indexes, and data lineage diagrams as universal data organization and management tools. It does not rely on the internal implementation of specific point cloud processing software and can be adapted to different point cloud processing and modeling systems. Through a unified version registration and rollback method, various point cloud processing modules can work collaboratively under the same project status management framework, which is beneficial for improving the maintainability and scalability of intelligent point cloud modeling projects.

[0077] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A project status management method for intelligent point cloud modeling, characterized in that: include, A project directory is established for the point cloud intelligent modeling project in the storage medium, a project list file is generated in the project directory, the original point cloud data is registered as the initial version, and the current version pointer is set to point to the initial version in the project list file. During the process of point cloud segmentation, point cloud division, geometric fitting and pipeline parameter adjustment of the original point cloud data, the point cloud block files, label files and 3D model files generated in each step are registered as subsequent versions, and the operation impact information including 3D spatial bounding boxes and semantic category masks is recorded for each version in the project list file. A three-dimensional spatial index structure is constructed based on the spatial location of the point cloud block file and the pipeline object, and the point cloud block file, the pipeline object and the corresponding operation impact information are associated and stored. When a rollback request including the target version identifier, the target three-dimensional spatial range, and the target semantic category set is received, the set of objects that intersect with the target three-dimensional spatial range and whose semantic category belongs to the target semantic category set is determined according to the three-dimensional spatial index structure, and the set of versions that affect the set of objects and whose version number is later than the target version identifier is determined in the project manifest file; Based on the target version identifier and the version set, a partial rollback is performed on the object set to restore it to the state corresponding to the target version identifier, and the restored object set state is registered as a new rollback version in the project manifest file.

2. The project status management method for intelligent point cloud modeling as described in claim 1, characterized in that: The partial rollback on the set of objects includes, When the rollback mode is soft rollback, only the mapping relationship between the object set and the target version identifier in the project manifest file is updated to the target version identifier, without deleting any files in the version directory; When the rollback mode is a hard rollback, while updating the mapping relationship, delete the local data files generated by the version set that do not belong to the version before the target version identifier, and delete the corresponding records in the project manifest file.

3. The project status management method for intelligent point cloud modeling as described in claim 2, characterized in that: The original point cloud data, the point cloud blocks, the label file, and the 3D model file are abstracted into multiple data nodes, and a directed acyclic data lineage graph is constructed with the input-output relationship of each processing step as the edge. For each data node in the data lineage graph, record cost parameters including at least storage size and recalculation cost.

4. The project status management method for intelligent point cloud modeling as described in claim 3, characterized in that: In the data lineage graph, a target data node set is determined based on the object set, and the total cost value of each data node in the target data node set is calculated using two strategies: reading from existing storage and recalculating from the parent node. Based on the total cost value, a dynamic programming algorithm or a greedy algorithm is used to determine the strategy combination for each target data node, and the target data node is read or recalculated according to the strategy combination.

5. The project status management method for intelligent point cloud modeling as described in claim 4, characterized in that: The pipeline network model in the point cloud intelligent modeling result is represented as a graph structure containing nodes and edges, where nodes represent pipe connection points and edges represent pipe segments with pipe diameter and material properties. For different version branches derived from the common ancestor version, the pipeline modification process of each branch relative to the common ancestor version is recorded as a graph editing sequence; When performing a version merge between two branches, the graph editing sequence is replayed based on the graph structure corresponding to the common ancestor version to obtain the change sets of the two branches. The change sets are compared to identify mutually exclusive modifications to the same node or edge, and topological conflicts are determined based on the identification results.

6. The project status management method for intelligent point cloud modeling as described in claim 5, characterized in that: The merged graph structure is subjected to connectivity checks and preset process constraint checks. If a disconnected pipe segment or illegal loop is detected, the graph editing operation that caused the disconnected pipe segment or illegal loop will be marked as a conflict.

7. The project status management method for intelligent point cloud modeling as described in claim 6, characterized in that: A single processing flow, including point cloud segmentation, point cloud division, geometric fitting, and pipeline parameter adjustment, is identified as a transaction. The intermediate results generated by each processing step in the process flow are written to the temporary version directory corresponding to the transaction. When all processing steps in the process are successfully completed and pass the consistency check, the data in the temporary version directory will be migrated to the official version directory and registered as the new version in the project manifest file. If any processing step in the process fails, the data in the temporary version directory is deleted and the transaction is undone.

8. A project status management system for intelligent point cloud modeling, based on the project status management method for intelligent point cloud modeling according to any one of claims 1 to 7, characterized in that: include, The project directory management unit is used to establish a project directory for the point cloud intelligent modeling project in the storage medium, generate a project list file in the project directory, register the original point cloud data as the initial version, and set the current version pointer to the initial version in the project list file. The version registration unit is used to register the point cloud block files, label files and 3D model files generated by each processing step as subsequent versions during the point cloud segmentation, point cloud division, geometric fitting and pipeline parameter adjustment of the original point cloud data, and to record the operation impact information including the 3D spatial bounding box and semantic category mask for each version in the project list file. The spatial index construction unit is used to construct a three-dimensional spatial index structure based on the spatial location of the point cloud block file and the pipeline object, and to associate and store the point cloud block file, the pipeline object and the corresponding operation impact information in the three-dimensional spatial index structure. The rollback determination unit is used to determine, upon receiving a rollback request including a target version identifier, a target three-dimensional spatial range, and a target semantic category set, a set of objects that intersect with the target three-dimensional spatial range and whose semantic category belongs to the target semantic category set according to the three-dimensional spatial index structure, and to determine, in the project manifest file, a set of versions that affect the set of objects and whose version number is later than the target version identifier. The partial rollback execution unit is used to perform a partial rollback on the object set according to the target version identifier and the version set to restore it to the state corresponding to the target version identifier, and register the restored object set state as a new rollback version in the project manifest file.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the project status management method for intelligent point cloud modeling as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the project status management method for intelligent point cloud modeling as described in any one of claims 1 to 7.

Images