Real-time checking method and system for three-dimensional model based on engineering drawings

Abstract

This application provides a method and system for real-time verification of 3D models based on engineering drawings, relating to the field of engineering model verification technology. The method first acquires operational behavior data from the modeling end and generates task awareness information based on this data, including the activity level of modeling operations, verification task attributes, and a task progress log. Then, based on the task awareness information, resource scheduling is performed, allocating core resources to high-frequency active tasks, idle resources to low-frequency active tasks, and minimum guaranteed resources to triggered critical point verification tasks. Verification is then performed according to the resource scheduling results and the engineering drawings, with closed-loop control implemented based on the task progress log and resource scheduling results during the verification process. This method optimizes the utilization of modeling and verification resources by distinguishing between critical point and non-critical point verifications and performing differentiated resource scheduling based on the high-frequency and low-frequency activity states of modeling.

Description

Technical Field

[0001] This application relates to the field of engineering model verification technology, and in particular to a method and system for real-time verification of three-dimensional models based on engineering drawings. Background Technology

[0002] Real-time verification of 3D models refers to a computer-aided design (CAD) process that verifies a 3D engineering model across multiple aspects based on engineering drawings, the physical laws of the model structure, and practical application needs after the model has been constructed. Verification items in real-time 3D model verification can include deviation verification, assembly correctness verification, geometric relationship verification, and manufacturing process verification. Different verification items can be based on different verification rules. After the 3D model passes multiple verifications, it can be applied to backend tasks.

[0003] Real-time verification of 3D models is a crucial step in engineering design to ensure consistency between the model and engineering drawings. Since real-time verification and modeling are performed simultaneously, it's essential to balance the smoothness of the modeling process with the accuracy of the core verification. Therefore, a fixed resource allocation can be used for both 3D modeling and real-time verification, assigning a fixed proportion of resources to each process to create a dual-threaded task for simultaneous real-time verification and modeling.

[0004] However, using a fixed resource allocation method can lead to competition for core resources such as CPU and memory between the modeling and verification processes during high-frequency, high-consumption modeling operations. This can cause modeling to lag or core verification to be interrupted. Conversely, during low-frequency, high-consumption modeling operations, a large amount of resources remain idle, resulting in low verification efficiency. Furthermore, core verification is prone to missing critical deviations due to resource contention failures, affecting the assembly accuracy of engineering structures. In addition, when the verification process is paused and resumed, it must be executed from the beginning, repeatedly consuming resources. Summary of the Invention

[0005] In view of this, embodiments of this application provide a method and system for real-time verification of 3D models based on engineering drawings, in order to solve the problem of unreasonable utilization of modeling and core verification resources.

[0006] According to a first aspect of this application, a method for real-time verification of a 3D model based on engineering drawings is provided, the method comprising:

[0007] Acquire modeling terminal operation behavior data, the operation behavior data including modeling action types;

[0008] Task perception information is generated based on the operational behavior data. The task perception information includes the modeling operation activity level, verification task attributes, and task progress log. The verification task attributes are used to define the task to be verified as a critical point verification task or a non-critical point verification task. The task progress log is used to record the task progress of non-critical point verification tasks.

[0009] Resource scheduling is performed based on the task awareness information to generate resource scheduling results; wherein, when the modeling operation activity level is high frequency, core resources are allocated to the task to be verified; when the modeling operation activity level is low frequency, idle resources are allocated to the task to be verified; when the verification task attribute is a key point verification task and the task to be verified is detected to be triggered, minimum guaranteed resources are allocated to the task to be verified.

[0010] According to the resource scheduling results, the task to be verified is verified using engineering drawings; and, according to the task progress log and the resource scheduling results, closed-loop control is performed. The closed-loop control is used to perform resource reallocation scheduling when the verification task at the critical point is completed; and to perform dynamic resource backfilling when the verification task at the non-critical point is completed or when the modeling operation restarts high-frequency activity.

[0011] In some embodiments, generating task-aware information based on the operational behavior data includes:

[0012] Extract the modeling action type from the operational behavior data;

[0013] Based on the modeling action type, query the resource consumption of the modeling operation;

[0014] The modeling operation is categorized based on the resource consumption, and the action categories include high-consumption actions and low-consumption actions.

[0015] According to the action category, extract the modeling action statistics within the preset time window. The modeling action statistics include the frequency of occurrence of the high-consumption action, the duration of a single action, and the execution interval of the low-consumption action.

[0016] The modeling operation intensity value is calculated based on the modeling action statistics, and the activity level of the modeling operation is determined based on the modeling operation intensity value.

[0017] In some embodiments, calculating the modeling operation intensity value based on the modeling action statistics includes:

[0018] The total duration of high-consumption actions within the preset time window is obtained by summing the duration of each single action.

[0019] Count the number of times high-consumption actions are executed consecutively within the preset time window;

[0020] The continuous execution fine-tuning factor is calculated based on the number of consecutive executions. When high-consumption actions are executed consecutively, the continuous execution fine-tuning factor is the ratio of the number of consecutive executions to the frequency of occurrence. When high-consumption actions are not executed consecutively, the continuous execution fine-tuning factor is a preset fixed value.

[0021] Obtain the average value of the execution interval within the preset time window;

[0022] The modeling operation intensity value is calculated by combining the average value. The modeling operation intensity value is the product of the total duration of the high-consumption action and the fine-tuning factor, and then the ratio of the average value.

[0023] In some embodiments, determining the activity level of modeling operations based on the modeling operation intensity value includes:

[0024] Calculate the average duration of a high-consumption action based on the duration of a single action within a preset time window;

[0025] If the intensity value of the modeling operation is greater than the average duration of a single operation, the activity level of the modeling operation is determined to be high-frequency activity.

[0026] If the intensity value of the modeling operation is less than or equal to the average duration of a single operation, the activity level of the modeling operation is determined to be low-frequency active.

[0027] In some embodiments, generating task-aware information based on the operational behavior data includes:

[0028] Extract the component hierarchy information marked in the engineering drawings, wherein the component hierarchy information includes main components and non-main components;

[0029] Obtain the model component corresponding to the task to be verified;

[0030] A decision set is defined based on the component hierarchy information; wherein, when the model component is a main component marked in the component hierarchy information, the task to be verified is assigned to the priority decision set; when the model component is a non-main component marked in the component hierarchy information, the task to be verified is assigned to the non-priority decision set.

[0031] The correlation of the task to be verified is detected by analyzing the topological structure of the three-dimensional model corresponding to the model components.

[0032] The verification task attributes are determined according to the judgment set and the correlation degree, wherein the verification tasks marked as high correlation tasks in the priority judgment set are defined as key point verification tasks.

[0033] In some embodiments, the correlation of the task to be verified is detected by analyzing the topological structure of the 3D model corresponding to the model components, including:

[0034] Iterate through the number of connections between the model component and other components;

[0035] Based on a preset spatial dependency judgment criterion, the number of dependencies is counted; the spatial dependency judgment criterion includes at least two of the following: assembly constraints between the model component and other components, shared geometric datum, and layout influence.

[0036] When the number of connections is greater than or equal to the number of connections threshold, and the number of dependencies is greater than or equal to the number of dependencies threshold, the task to be verified is set as a high-relevance task.

[0037] When the number of connections is less than the connection number threshold, or when the number of dependencies is less than the dependency number threshold, the task to be verified is set as a low-relevance task.

[0038] In some embodiments, resource scheduling is performed based on the task-aware information to generate resource scheduling results, including:

[0039] Based on the task awareness information, priority verification tasks and non-priority verification tasks are determined; the priority verification tasks include the verification tasks whose modeling operation activity level is high frequency and active, or the verification task attributes are key point verification tasks and the verification tasks are triggered.

[0040] By calling the system resource isolation interface, the execution resources allocated to the priority verification task are physically isolated. The execution resources include core resources and minimum guaranteed resources. The minimum guaranteed resources are the minimum dedicated resources required to maintain the basic operation of the critical point verification task.

[0041] The priority verification task is bound to the first verification process so that the first verification process can perform verification on the priority verification task.

[0042] The execution status of the second verification process corresponding to the non-priority verification task is frozen, and the resources corresponding to the second verification process are released to the modeling process.

[0043] In some embodiments, according to the resource scheduling result, the task to be verified is verified using engineering drawings, including:

[0044] Resource usage data of the modeling process can be obtained through the system resource monitoring interface;

[0045] Based on the resource usage data, the available idle resources are calculated.

[0046] The idle time resources are allocated to the second verification process;

[0047] Obtain historical frozen data of the non-priority verification task, the historical frozen data including process context and real-time progress;

[0048] Using the second verification process, the non-priority verification task is verified according to the historical frozen data.

[0049] In some embodiments, closed-loop control is performed based on the task progress log and the resource scheduling results, including:

[0050] Monitor the verification progress of key points in the task progress log;

[0051] When the key point verification task is completed, the resource isolation status of the key point verification task is released.

[0052] Reclaim the resources occupied by the key point verification task;

[0053] When the modeling process is in a low-frequency active state and there are non-critical point verification tasks that have not been verified, resources are allocated to the non-critical point verification tasks according to the number of uncompleted tasks recorded in the task progress log and the occupied resources.

[0054] According to a second aspect of this application, a real-time verification system for 3D models based on engineering drawings is provided, the system comprising:

[0055] The dual-state perception module is used to acquire modeling terminal operation behavior data and generate task perception information based on the operation behavior data; the operation behavior data includes modeling action type; the task perception information includes modeling operation activity level, verification task attributes, and task progress log; the verification task attributes are used to define the task to be verified as a critical point verification task or a non-critical point verification task; the task progress log is used to record the task progress of non-critical point verification tasks.

[0056] The resource scheduling module is used to perform resource scheduling based on the task awareness information to generate resource scheduling results; wherein, when the modeling operation activity level is high frequency, core resources are allocated to the task to be verified; when the modeling operation activity level is low frequency, idle resources are allocated to the task to be verified; when the verification task attribute is a key point verification task and the task to be verified is detected to be triggered, minimum guaranteed resources are allocated to the task to be verified.

[0057] The closed-loop collaboration module is used to perform verification on the task to be verified using engineering drawings according to the resource scheduling results; and to perform closed-loop control according to the task progress log and the resource scheduling results. The closed-loop control is used to perform resource reallocation scheduling when the verification task at the critical point is completed; and to perform dynamic resource backfilling when the verification task at the non-critical point is completed or when the modeling operation restarts high-frequency activity.

[0058] By employing the above technical solution, this application provides a method and system for real-time verification of 3D models based on engineering drawings. The method first acquires operational behavior data from the modeling end and generates task awareness information, including the activity level of modeling operations, verification task attributes, and a task progress log, based on the operational behavior data. Then, resource scheduling is performed based on the task awareness information, allocating core resources to high-frequency active tasks, idle resources to low-frequency active tasks, and minimum guaranteed resources to triggered critical point verification tasks. Verification is then performed according to the resource scheduling results and the engineering drawings, and closed-loop control is executed during the verification process based on the task progress log and resource scheduling results. This method optimizes the utilization of modeling and verification resources by distinguishing between critical point and non-critical point verification and performing differentiated resource scheduling based on the high-frequency and low-frequency activity states of modeling.

[0059] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below. Attached Figure Description

[0060] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0061] Figure 1 A schematic diagram of the real-time verification method for 3D models based on engineering drawings provided in this application embodiment;

[0062] Figure 2 This is a schematic diagram of the entire real-time verification process provided in the embodiments of this application;

[0063] Figure 3 This is a schematic diagram of the resource scheduling process provided in the embodiments of this application;

[0064] Figure 4 This is a schematic diagram of the closed-loop control process provided in the embodiments of this application;

[0065] Figure 5 This is a schematic diagram of the structure of a real-time verification system for 3D models based on engineering drawings, provided in an embodiment of this application. Detailed Implementation

[0066] The present application will be described in detail below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in the embodiments of the present application can be combined with each other.

[0067] In this embodiment, the three-dimensional model is a digital representation of an object, constructed in a three-dimensional coordinate system using basic geometric primitives such as points, lines, surfaces, and volumes, capable of expressing the spatial position and shape of an object. The three-dimensional model can be created using CAD applications based on Computer-Aided Design (CAD) principles and following a parametric modeling approach.

[0068] Real-time verification of 3D models refers to the CAD process of verifying a 3D model through multiple aspects after its construction, based on engineering drawings, the physical laws of the model structure, and the needs of actual applications. Verification items in real-time 3D model verification can include deviation verification, assembly correctness verification, geometric relationship verification, and machining process verification. Different verification items can be based on different verification rules. After the 3D model passes multi-item verification, it can be applied to backend tasks.

[0069] Real-time verification of 3D models is a crucial step in engineering design to ensure consistency between the model and engineering drawings. Since real-time verification and modeling are performed simultaneously, it is necessary to balance the smoothness of modeling operations with the accuracy of core verification. Therefore, in some embodiments, a fixed resource allocation can be used for both 3D modeling and real-time verification, assigning fixed resource ratios to the modeling and verification processes respectively, forming a dual-threaded task for simultaneous real-time verification and modeling of the 3D model.

[0070] However, using a fixed resource allocation method can lead to competition for core resources such as CPU and memory between the modeling and verification processes during high-frequency, high-consumption modeling operations. This can cause modeling to lag or core verification to be interrupted. Conversely, during low-frequency, high-consumption modeling operations, a large amount of resources remain idle, resulting in low verification efficiency. Furthermore, core verification is prone to missing critical deviations due to resource contention failures, affecting the assembly accuracy of engineering structures. In addition, when the verification process is paused and resumed, it must be executed from the beginning, repeatedly consuming resources.

[0071] To address the issue of inefficient utilization of modeling and core verification resources, this application provides a real-time verification method for 3D models based on engineering drawings in some embodiments. This method acquires modeling operation data in real time and assesses activity levels to distinguish between critical and non-critical points, and establishes a task progress log. Based on activity levels and verification attributes, differentiated scheduling is implemented, including core resource anchoring, idle-time resource redirection, and minimum guaranteed resource allocation, to achieve rational resource allocation. Finally, by comparing the task progress log with the scheduling results, resource reallocation scheduling and dynamic resource repatriation are performed. This method achieves a balance between modeling fluency and verification accuracy, improving resource utilization and verification efficiency.

[0072] The method can be applied to electronic devices with data processing capabilities. These electronic devices include, but are not limited to, computers, servers, mobile terminals, smart wearable devices, and industrial control machines. For ease of description, this application embodiment uses an electronic device as the execution subject of the method. It should be understood that the method can also be applied to other types of execution subjects, which are not illustrated in this application embodiment. Figure 1 As shown, the method includes:

[0073] S101. Obtain operational behavior data from the modeling end.

[0074] During real-time verification of a 3D model, operational behavior data from the modeling end can be acquired first. This data includes modeling action types. Modeling action types represent the specific modeling actions performed by the user during the 3D model creation process. Different modeling actions belong to different modeling action types and correspond to different modeling functions.

[0075] In some embodiments, modeling action types may include 3D component creation, assembly, surface modeling, dimension modification, component position fine-tuning, and attribute information entry. Different modeling action types can be completed using different interactive methods. Taking 3D component creation as an example, the 3D component creation action may involve modeling actions such as sketch creation, solid generation, shape scanning, and viewpoint adjustment. These modeling actions can generate a series of time-related operation behavior data at the modeling end.

[0076] like Figure 2 As shown, to acquire operational behavior data, a dual-state sensing module can be built into the electronic device. This module can capture operational behavior from the modeling end in real time by listening to the user interface. Therefore, during real-time verification of the 3D model, a data acquisition request can be generated first, and in response, the dual-state sensing module can be invoked to capture operational behavior data.

[0077] S102. Generate task awareness information based on operational behavior data.

[0078] After acquiring operational behavior data, task awareness information can be generated based on this data. This task awareness information includes modeling operation activity level, verification task attributes, and a task progress log. Modeling operation activity level reflects the number of times each modeling operation is executed per unit of time; higher activity levels indicate more frequent modeling operations. Verification task attributes define whether the task to be verified is a critical point verification task or a non-critical point verification task. The task progress log records the progress of non-critical point verification tasks.

[0079] Therefore, the dual-state sensing module built into electronic devices can, after acquiring operational behavior data, determine the activity level of the modeling operation based on the current modeling action type, and simultaneously distinguish the verification task attributes in real time to define whether the current task to be verified is a critical point verification or a non-critical point verification. It also constructs a task progress ledger by recording the verification progress of non-critical points.

[0080] To determine the activity level of modeling operations, in some embodiments, when generating task-aware information based on operation behavior data, the modeling action type of the modeling operation can be extracted from the operation behavior data first, and then the resource consumption of the modeling operation can be queried based on the modeling action type. Then, the action category of the modeling operation is categorized based on the resource consumption, wherein the action category includes high-consumption actions and low-consumption actions.

[0081] The high-consumption actions refer to 3D modeling operations that heavily utilize computer hardware resources, have long single-execution times, and require complex processes. These high-consumption actions continuously consume significant computing power during execution, and restarting after an interruption is costly. Examples of high-consumption actions include 3D component creation, assembly, and surface modeling. 3D component creation requires generating a complete 3D geometric model and involves numerous vertex and face calculations and rendering; assembly requires establishing spatial constraints between multiple components and calculating their relative positions and assembly accuracy; and surface modeling requires fitting complex free-form surfaces, placing extremely high demands on the GPU's graphics processing capabilities.

[0082] Low-consumption actions refer to 3D modeling operations that require minimal hardware resources, have short execution times, and involve only parameter modifications. These actions exhibit minimal fluctuations in resource usage during execution and incur no additional cost upon interruption. Examples of low-consumption actions include dimension annotation modification, component position fine-tuning, and attribute information entry. Dimension annotation modification requires adjusting numerical parameters without altering the model's geometry; component position fine-tuning involves translating or rotating the component's spatial position without reconstructing its topology; and attribute information entry only requires filling in textual information such as the component's material and specifications, consuming no computing resources.

[0083] After obtaining the action categories of the modeling operations, statistical information on the modeling actions within a preset time window can be extracted according to the action categories. This statistical information includes the frequency of high-consumption actions, the duration of a single action, and the execution interval of low-consumption actions. Then, the intensity value of the modeling operation is calculated based on the statistical information, and the activity level of the modeling operation is determined based on this intensity value.

[0084] The modeling operation intensity value is used to determine the activity level of modeling operations, that is, to determine whether a certain modeling operation or a certain type of modeling operation is high-frequency or low-frequency. Therefore, the modeling operation intensity value can be calculated based on statistical information about modeling operations, such as the frequency of high-consumption actions, the duration of a single action, and the execution interval of low-consumption actions. Specifically, in some embodiments, when calculating the modeling operation intensity value based on the statistical information of modeling operations, the total duration of high-consumption actions within a preset time window can be obtained by first accumulating the duration of a single action. Then, the number of consecutive executions of high-consumption actions within the preset time window is counted, and a continuous execution fine-tuning factor is calculated based on the number of consecutive executions. Finally, the average execution interval within the preset time window is obtained, and the modeling operation intensity value is calculated in combination with the average value. That is, the modeling operation intensity value is the product of the total duration of high-consumption actions and the fine-tuning factor, divided by the average value.

[0085] For example, after acquiring operational behavior data, statistical analysis can be performed on this data to determine the activity level of modeled operational intensity values ​​within a preset time window. To this end, the frequency N of high-consumption actions and the duration T of each action within the preset time window can be extracted from the operational behavior data. i Low-consumption action execution interval △t.

[0086] Based on the statistical parameters, the continuous execution fine-tuning factor α is determined. When high-consumption actions are executed continuously, the continuous execution fine-tuning factor is the ratio of the number of consecutive executions to the frequency of occurrence. When high-consumption actions are not executed continuously, the continuous execution fine-tuning factor is a preset fixed value.

[0087] For example, the rule for determining the value of the continuous execution fine-tuning factor α is as follows: First, a judgment is made on the continuous execution of high-consumption actions. If, within a preset time window, after the previous high-consumption action is completed, no low-consumption action is inserted or there is no obvious operation pause, and the next high-consumption action is started directly, then the current operation state can be determined as high-consumption actions being executed continuously. At this time, the continuous execution fine-tuning factor α can be obtained by counting the number of consecutive executions M of high-consumption actions within the window and combining it with the frequency N of high-consumption actions within the preset time window. The continuous execution fine-tuning factor α is the ratio of the number of consecutive executions M to the frequency N, i.e., α = M / N, M ≤ N.

[0088] When determining obvious operation pauses, the action execution log of the modeling system is used as the basis. Specifically, the time interval between two consecutive high-consumption actions is compared. If this time interval is less than the minimum time interval between any two low-consumption actions within the same window, it is determined that there is no pause; otherwise, it is determined to be an obvious operation pause. Similarly, if a low-consumption action is inserted after a high-consumption action has been completed, or if there is an obvious operation pause, it is determined that the high-consumption action is not executed continuously. In this case, the continuous execution fine-tuning factor can be set to a preset fixed value. For example, if the high-consumption action is not executed continuously, then α = 0.5.

[0089] Based on the parameters calculated above, the modeling operation intensity value is calculated according to the following formula:

[0090]

[0091] Where S represents the modeling operation intensity value; T i This represents the duration of a single action of the i-th action; N represents the frequency of high-consumption actions within the preset time window. α represents the average value of the low-cost action execution interval Δt within the preset time window; α represents the continuous execution fine-tuning factor.

[0092] As can be seen from the formula above, multiplying the total duration by the continuous execution fine-tuning factor in the numerator better reflects the actual modeling scenario. Since continuously executed high-consumption actions exert significantly more pressure on system resources than intermittently executed high-consumption actions of the same duration, α is introduced to quantify this difference, thus reducing the overestimation of the impact of intermittent operations. The denominator is used to mitigate the interference of scattered low-consumption actions. The shorter the interval between low-consumption actions, the smaller the denominator value, and the corresponding operation intensity value S will not be unnecessarily inflated by high-frequency fine-tuning, ensuring accurate judgment.

[0093] After calculating the modeling operation intensity value, the activity level of the modeling operation can be determined based on this value. To do this, the average duration of a high-consumption action can be calculated based on the duration of a single action within a preset time window, serving as a threshold for judging the modeling operation intensity. If the modeling operation intensity value is greater than the average duration, the modeling operation activity level is determined to be high-frequency active; if the modeling operation intensity value is less than or equal to the average duration, the modeling operation activity level is determined to be low-frequency active.

[0094] For example, by calculating the average single duration T of high-consumption actions within a preset fixed time window. avg The modeling operation intensity value S and the average single duration T are then compared. avg A comparison was performed. The modeling operation intensity value was greater than the average single-attempt duration, i.e., S > T. avgWhen the modeling operation activity level of the task to be verified is high-frequency, it can be determined that the modeling operation activity level is less than or equal to the average duration of a single operation, i.e., S≤T. avg If this is the case, the modeling operation activity level of the task to be verified can be determined to be low-frequency active.

[0095] As can be seen, since the frequency and duration of high-consumption actions directly reflect the operational intensity of modeling, the activity level can be objectively reflected by statistically analyzing the frequency of high-consumption actions, the duration of a single action, and the execution interval of low-consumption actions within a preset time window. At the same time, recording the interval of low-consumption actions can eliminate the interference of meaningless frequent clicks on the judgment results, enabling the dual-state perception module to capture the operational behavior of the modeling end in real time and judge the activity level of the modeling operation based on the current modeling action type, thereby generating task perception information.

[0096] To determine the attributes of the verification task, in some embodiments, when generating task-aware information based on operational behavior data, component hierarchy information marked in engineering drawings can also be extracted. This component hierarchy information includes main components and non-main components.

[0097] Key components can include main load-bearing components, core assembly components, and other important components marked on the engineering drawings. Main load-bearing components refer to the core components in the engineering structure that directly bear and transmit the overall load, determining the structure's stability and safety. Their mechanical properties are core indicators of the entire structural design; any deviation or defect will directly lead to the failure of the entire structure. Core assembly components refer to the critical assembly units that determine the realization of the equipment's core functions; their assembly accuracy directly affects the equipment's operational performance.

[0098] In a 3D model, components other than the main load-bearing components can be referred to as non-load-bearing components. These can include auxiliary connecting components and auxiliary positioning components. Auxiliary connecting components are auxiliary components used to connect main load-bearing components or core assembly components. They do not bear the overall structural load, and their absence will not directly cause the overall structure or equipment to fail; it only affects the stability of the local connection. Auxiliary positioning components are special components used to ensure the accurate positioning of main load-bearing components and core assembly components during construction or assembly. They do not participate in structural load transfer or equipment functionality; they only play a role during installation.

[0099] After extracting the component hierarchy information marked in the engineering drawings, the model components corresponding to the task to be verified can be obtained, and a decision set can be defined based on the component hierarchy information. Specifically, when the model component is a main component marked in the component hierarchy information, the task to be verified is included in the priority decision set; when the model component is a non-main component marked in the component hierarchy information, the task to be verified is included in the non-priority decision set.

[0100] For example, to synchronously and in real-time differentiate the attributes of verification tasks and define whether the current task to be verified is a critical point verification or a non-critical point verification, the component hierarchy information marked in the engineering drawings can be extracted, and the model components involved in the task to be verified can be compared one by one with the components in the drawings. If the model component of the task to be verified corresponds to the verification task of the main load-bearing component or core assembly component clearly marked in the engineering drawings, it is included in the priority decision set. If the model component of the task to be verified corresponds to the verification task of the auxiliary connection component or auxiliary positioning component attached to the engineering drawings, it is included in the non-priority decision set, i.e., the secondary decision set.

[0101] Then, by analyzing the topology of the 3D model corresponding to the model components, the correlation of the task to be verified is detected. In some embodiments, in order to detect the correlation of the task to be verified, the number of other components connected to the model component can be traversed first, i.e., the number of connections, and the number of components with strong spatial dependencies can be counted based on a preset spatial dependency judgment standard, i.e., the number of dependencies. The spatial dependency judgment standard includes at least two of the following: assembly constraints between the model component and other components, shared geometric datum, and layout influence.

[0102] For example, by analyzing the topology of a 3D model, the number of connections and spatial dependencies between the model components and other components corresponding to the task to be verified can be calculated. Based on these two features, the task can be labeled as a high-relevance task or a low-relevance task. The criteria for judging the spatial dependency may include: whether there are assembly constraints between the model components and other components, such as rigid assembly relationships such as fitting, coaxiality, and parallelism; whether the model components and other components share geometric references, such as sharing the same reference plane or reference axis for dimensioning; and whether the spatial position of the model components determines the layout range of other components, such as load-bearing wall components in a building model, whose position directly limits the erection area of ​​beams and floor slabs.

[0103] If any one of the above criteria is met, the model component is determined to have a spatial dependency with its corresponding component; if two or more criteria are met, it is determined to have a strong spatial dependency. Similarly, if zero criteria are met, it is determined to have no spatial dependency.

[0104] Then, based on the number of connections between model components and other components, and the number of dependencies that determine strong spatial dependencies, the task to be verified is determined as a high-relevance task or a low-relevance task. When the number of connections is greater than or equal to a connection threshold, and the number of dependencies is greater than or equal to a dependency threshold, the task to be verified is set as a high-relevance task; when the number of connections is less than a connection threshold, or the number of dependencies is less than a dependency threshold, the task to be verified is set as a low-relevance task.

[0105] For example, count the number of other components directly connected to the model component, i.e., the number of connections C, and distinguish the spatial dependencies between the model component and each connected component, counting the number of dependencies Q of strong spatial dependencies. Then compare each of these with a preset threshold. If the task to be verified satisfies C≥C th And Q≥Q th If the task is highly relevant, it is considered a high-relevance task; otherwise, it is considered a low-relevance task. Where C... th Q is the threshold for the number of connections. th Since both depend on the quantity threshold, they need to be preset according to the actual situation.

[0106] After defining the decision sets and obtaining the relevance as described above, the verification task attributes can be determined based on the decision sets and relevance. When determining the verification task attributes, tasks marked as high-relevance tasks in the priority decision set can be defined as key point verification tasks; otherwise, they are defined as non-key point verification tasks. That is, tasks marked as high-relevance tasks in the priority decision set are defined as key point verification tasks; while other tasks not defined as key point verification tasks are defined as non-key point verification tasks.

[0107] For non-critical point verification tasks, the subsequent idle resource allocation will be affected. Idle resource allocation can only be initiated from the initial non-critical point verification task, requiring previously completed verification work to be repeated, wasting idle resources. Therefore, after determining the critical and non-critical point verification tasks, the verification progress of both can be recorded, and a task progress ledger can be constructed based on this record. The task progress ledger serves as a storage medium for historical pause data. During idle resource allocation, verification can be initiated from the tasks recorded in the task progress ledger, ensuring that previously completed verification work does not need to be repeated.

[0108] Furthermore, for dynamic resource reclaiming and resource reallocation scheduling, the status of non-critical point verifications can be effectively distinguished based on the task progress ledger, reducing issues such as tasks being completed but resources not being reclaimed, or tasks being paused but mistakenly judged as terminated. Therefore, for critical point verification tasks, when resource reallocation scheduling is required after verification, the determination of its completion status can rely on the progress records of the critical points. Without a task progress ledger, dedicated resources may be prematurely reclaimed before critical point verification is completed, leading to interruption of core verification, or critical point verification may be completed, but resources may remain unclaimed for a long time, resulting in resource waste.

[0109] S103. Perform resource scheduling based on task-aware information to generate resource scheduling results.

[0110] After generating task awareness information, resource scheduling can be performed based on this information to generate resource scheduling results. Specifically, by performing resource scheduling, core resources can be allocated to the task to be verified when the modeling operation activity level is high-frequency; idle resources can be allocated to the task to be verified when the modeling operation activity level is low-frequency; and minimum guaranteed resources can be allocated to the task to be verified when the verification task attribute is a key point verification task and the task to be verified is detected to be triggered.

[0111] To this end, a resource scheduling module can be configured in the electronic device. The resource scheduling module can execute the following three resource scheduling methods based on the activity level of the modeling operation and the attributes of the verification task: if the modeling operation is detected to be high-frequency active, core resource anchoring is executed; if the modeling operation is detected to be low-frequency active, idle time resource diversion is executed; if the key point verification task is detected to be triggered, the minimum guaranteed resources need to be allocated for the key point verification.

[0112] like Figure 3 As shown, in order to perform resource scheduling, in some embodiments, priority verification tasks and non-priority verification tasks can be determined based on task-aware information. The priority verification tasks include verification tasks with high-frequency activity levels in modeling operations, or verification tasks whose attributes are key point verification tasks and which have been triggered.

[0113] Then, by calling the system resource isolation interface, the execution resources allocated to the priority verification task are physically isolated. The execution resources include core resources and minimum guaranteed resources; the core resources are the main computing resources of the electronic device when performing modeling and verification; the minimum guaranteed resources are the minimum dedicated resources required to maintain the basic operation of the key point verification task.

[0114] The priority verification task is then bound to the first verification process, so that the first verification process can perform verification on the priority verification task. Meanwhile, the status of the second verification process corresponding to the non-priority verification task is frozen, and the resources corresponding to the second verification process are released to the modeling process.

[0115] For example, based on a resource scheduling module, electronic devices can perform appropriate resource scheduling according to the activity level of modeling operations and the attributes of verification tasks. The activity level of modeling operations determines the intensity of resource consumption; high-frequency activity requires more resources to maintain smooth operation, while low-frequency activity results in idle resources. Verification task attributes determine verification priority; key point verification is related to the core accuracy of the model and cannot be interrupted, while non-key point verification can be executed off-peak. Therefore, combining the activity level of modeling operations with the attributes of verification tasks enables intelligent scheduling, ensuring the real-time performance of key point verification while avoiding resource shortages that could cause high-frequency modeling to lag.

[0116] To resolve the resource contention between modeling and verification, the activity level of modeling operations and the attributes of verification tasks can be introduced. If high-frequency modeling activity is detected, core resource anchoring is executed. By calling the system resource isolation interface, a preset minimum guaranteed resource is physically isolated and bound to the keypoint verification process, preventing the modeling process from occupying this resource. Physical isolation prevents the modeling process from preempting keypoint verification resources, ensuring uninterrupted basic keypoint verification operations, while locking the minimum guaranteed resource prevents the occupation of redundant resources and guarantees resource supply for high-frequency modeling.

[0117] Similarly, if the verification task attribute of the task to be verified is detected as a key point verification task and it is triggered, minimum guaranteed resources can be allocated to the key point verification task. The minimum guaranteed resources refer to the minimal, dedicated resources required to maintain the basic operation of key point verification; they only need to meet the core computing and storage requirements of key point verification and do not occupy redundant resources.

[0118] For example, based on historical running data of keypoint verification, the minimum resource consumption for a single keypoint verification can be calculated. That is, the resource selection is based on just maintaining the keypoint verification without interruption. One or two independent computing cores can be selected on the CPU to achieve physical isolation and not participate in modeling resource scheduling. Fixed physical memory blocks are allocated to store only the real-time data of keypoint verification. The GPU occupies the minimum video memory and only supports geometric feature extraction.

[0119] It should be noted that if a keypoint verification task is detected, minimum guaranteed resources must be allocated for keypoint verification regardless of the current activity level of the modeling operation. Since keypoint verification is directly related to the core accuracy and engineering availability of the 3D model, interruption due to high-frequency modeling activity may lead to the omission of core deviations, requiring a subsequent full verification and increasing time costs. Meanwhile, the basic operational resource requirements for keypoint verification are low, and allocating minimum guaranteed resources will not significantly impact high-frequency modeling.

[0120] For currently executing non-critical point verification processes, a state freeze can be performed, immediately reclaiming the CPU, memory, and GPU resources they occupy and releasing them to the modeling process. Simultaneously, the process context and real-time progress are saved to a separate storage area, retaining only the frozen process data. Freezing the process avoids subsequent re-verification of completed portions, saving time, while targeted resource release maximizes the resource requirements of high-frequency modeling, maintaining modeling smoothness.

[0121] S104. Based on the resource scheduling results, perform verification on the task to be verified using engineering drawings; and perform closed-loop control based on the task progress log and resource scheduling results.

[0122] After resource scheduling is performed, the tasks to be verified can be executed according to the resource scheduling results and using the engineering drawings. Specifically, for frequently active tasks to be verified, a verification process, i.e., the first verification process, can run based on the allocated core resources. The first verification process can execute verification algorithms based on the allocated core resources, using the engineering drawings and corresponding geometric, physical, and assembly verification rules to verify the frequently active tasks to be verified and generate verification result information.

[0123] For low-frequency, active verification tasks, the verification task can be executed based on whether it is a critical point verification task and whether it has been triggered. That is, for verification tasks whose verification task attribute is a critical point verification task and which has been triggered, the verification process can be run using the minimum guaranteed resources to verify the triggered critical point verification task using engineering drawings and corresponding geometric, physical, and assembly verification rules.

[0124] For low-frequency, non-critical verification points, and critical verification point tasks that have not been triggered, idle resources can be utilized for verification. Specifically, in some embodiments, when executing verification of tasks to be verified using engineering drawings according to resource scheduling results, resource usage data of the modeling process can be obtained first through the system resource monitoring interface, and based on this data, available idle resources can be calculated. These idle resources are then allocated to the second verification process, and historical frozen data for non-priority verification tasks is obtained, including process context and real-time progress. Finally, the second verification process is used to perform verification on non-priority verification tasks according to the historical frozen data.

[0125] For example, if the modeling operation is detected to be in a low-frequency active state, then idle resource diversion will be performed. During the idle resource diversion process, the resource usage data of the modeling process can be collected in real time through the system resource monitoring interface, the current idle CPU, memory and GPU resources can be counted, and the available idle resources can be filtered out, but the resources occupied by the system's basic operation and key point verification need to be excluded.

[0126] Not all idle resources during the low-frequency active phase of modeling can be used for non-critical point verification. It is necessary to collect real-time usage data of the modeling process to distinguish between truly idle resources and those reserved for the modeling process, so as to prevent delays in modeling operations due to over-allocation of resources.

[0127] Then, all detected idle resources are reassigned to the verification process at non-critical points. If historical frozen data exists, the frozen process context and real-time progress are invoked, and verification is started from the frozen real-time progress; otherwise, execution is started directly from the initial task of the non-critical point verification.

[0128] Idle resources need to be accurately allocated to non-critical verification processes to avoid them being preempted by other non-core processes, ensuring maximum resource utilization. At the same time, targeted allocation can establish a unique association between resources and processes, preparing for rapid resource reclamation during subsequent modeling restarts at high frequencies.

[0129] While performing verification of the tasks to be verified using engineering drawings according to the resource scheduling results, closed-loop control can also be performed based on the task progress log and resource scheduling results. Specifically, closed-loop control is used to perform resource reallocation scheduling when the verification task at the critical point is completed; and to perform dynamic resource backflow when the verification task at the non-critical point is completed or when the modeling operation restarts high-frequency activity.

[0130] Therefore, the built-in closed-loop collaboration module in electronic devices can also perform closed-loop control through task progress ledgers and resource scheduling results. When the critical point verification is completed, resource reallocation scheduling continues; when non-critical points are verified or high-frequency activity is restarted after modeling operations, dynamic resource backfilling continues.

[0131] like Figure 4 As shown, in order to perform closed-loop control, in some embodiments, when performing closed-loop control based on the task progress log and resource scheduling results, the verification progress of key point verification tasks in the task progress log can be monitored. When the key point verification task is completed, the resource isolation state of the key point verification task is released, and the resources occupied by the key point verification task are reclaimed. Then, when the modeling process is in a low-frequency active state and there are non-key point verification tasks that have not been verified, resources are allocated to the non-key point verification tasks according to the amount of unfinished tasks and occupied resources recorded in the task progress log.

[0132] For example, the closed-loop collaboration module can perform closed-loop control through task progress logs and resource scheduling results. During critical point verification, the system physically locks dedicated resources, completely isolating these resources from modeling and non-critical point verification processes to prevent preemption. However, after critical point verification is completed, this isolation state is no longer necessary. If not released, these resources will remain idle and locked for a long time, unable to be accessed by other processes. Therefore, resource reallocation scheduling can continue upon completion of critical point verification. Specifically, when the task progress log marks the completion of critical point verification, the system immediately releases the resource isolation state of the task, reclaims all its occupied resources, including physically isolated CPU, memory, and GPU resources, and clears the task's process context data.

[0133] The recovered resources are then prioritized for allocation to the modeling process. If the modeling process is infrequently active and has unfinished non-critical point verification tasks, resources are allocated to the non-critical point verification process based on the number of unfinished tasks recorded in the task progress log.

[0134] When allocating resources for non-critical point verification processes, the total number of uncompleted verification items for the target non-critical point verification task in the task progress log can be read, along with the average resource consumption value per verification item generated by the system based on historical running data of similar tasks. Multiplying the two together yields the theoretical total resource requirement for the task to complete the remaining work.

[0135] Next, the total theoretical resource requirement is compared with the total resource recovered from keypoint verification. Simultaneously, 30% of the total keypoint verification resource recovery is reserved in the modeling process resource pool to ensure unforeseen operational needs during infrequent but active modeling tasks. It is evident that resource allocation must adhere to two constraints: it must not exceed the total theoretical resource requirement, and it must not exceed 70% of the total resource recovered from keypoint verification.

[0136] The task progress log records the percentage of non-critical point verifications that are not completed. Allocating resources based on this log avoids over-allocation leading to insufficient modeling resources or under-allocation leading to low verification efficiency. It eliminates the need for preset resource allocation ratios, allowing for dynamic adjustments based on the actual workload.

[0137] When non-critical points complete verification or modeling operations and restart high-frequency activity, dynamic resource backflow continues. That is, the system immediately freezes the non-critical point verification process, saves the progress to the task progress ledger, and reclaims all CPU, memory, and GPU resources it occupies.

[0138] Dynamic resource recharge is designed for scenarios where non-critical point verification is incomplete but modeling restarts with high frequency of activity. If the process is terminated directly, subsequent modeling reverts to low-frequency activity and verification needs to be restarted from scratch, resulting in repetitive work and resource waste. Freezing the process, however, can fully preserve the current verification context data and execution status, providing data support for subsequent resource redirection during idle periods. By directly injecting all reclaimed resources into the resource pool of the modeling process, updating the resource scheduling results, and synchronizing them to the system monitoring module, it can be ensured that the modeling process can call upon newly added resources in real time.

[0139] If the reclaimed resources are redistributed in the public resource pool, they may be preempted by other non-core processes, making it impossible to quickly respond to the high-frequency and active resource demands of modeling. However, updating and synchronizing the resource scheduling results allows the system monitoring module to keep track of the flow of resources from non-critical point verification and reclamation to injection into the modeling process in real time, avoiding problems such as unclear resource flow and out-of-control scheduling status.

[0140] As can be seen, closed-loop control can promptly reclaim and reuse the resources occupied by critical points after verification. Furthermore, resources will not remain idle for extended periods after non-critical point verification. Therefore, by linking the ledger with scheduling results, closed-loop control can be achieved, accurately triggering resource reclamation and reallocation, improving resource utilization, and reducing system resource waste.

[0141] By applying the technical solutions of the above embodiments, the real-time verification method for 3D models based on engineering drawings described in the above embodiments can distinguish between key point and non-key point verification, and perform differentiated resource scheduling based on high-frequency and low-frequency active states of modeling. This not only reserves minimum guaranteed resources for key point verification, but also avoids operational lag caused by resource contention during high-frequency modeling. Furthermore, during low-frequency active modeling, idle resources can be redirected to non-key point verification. When verification is paused, progress is frozen and recorded, and execution resumes from the historical progress point upon resumption, avoiding resource waste caused by repeated verification. The method can also synchronize verification progress and resource scheduling results through a task progress ledger, accurately triggering operations such as resource relocation and dynamic backfilling to ensure timely resource recovery and redistribution.

[0142] In some embodiments, as a specific implementation of the real-time verification method for 3D models based on engineering drawings described in the above embodiments, some embodiments of this application also provide a real-time verification system for 3D models based on engineering drawings, such as... Figure 5 As shown, the system includes:

[0143] The dual-state perception module is used to acquire modeling terminal operation behavior data and generate task perception information based on the operation behavior data; the operation behavior data includes modeling action type; the task perception information includes modeling operation activity level, verification task attributes, and task progress log; the verification task attributes are used to define the task to be verified as a critical point verification task or a non-critical point verification task; the task progress log is used to record the task progress of non-critical point verification tasks.

[0144] The resource scheduling module is used to perform resource scheduling based on the task awareness information to generate resource scheduling results; wherein, when the modeling operation activity level is high frequency, core resources are allocated to the task to be verified; when the modeling operation activity level is low frequency, idle resources are allocated to the task to be verified; when the verification task attribute is a key point verification task and the task to be verified is detected to be triggered, minimum guaranteed resources are allocated to the task to be verified.

[0145] The closed-loop collaboration module is used to perform verification on the task to be verified using engineering drawings according to the resource scheduling results; and to perform closed-loop control according to the task progress log and the resource scheduling results. The closed-loop control is used to perform resource reallocation scheduling when the verification task at the critical point is completed; and to perform dynamic resource backfilling when the verification task at the non-critical point is completed or when the modeling operation restarts high-frequency activity.

[0146] By applying the technical solutions of the above embodiments, the real-time verification system for 3D models based on engineering drawings described in the above embodiments includes a dual-state perception module, a resource scheduling module, and a closed-loop collaboration module. After acquiring operational behavior data from the modeling end, the dual-state perception module can generate task perception information, including the activity level of modeling operations, verification task attributes, and a task progress log, based on the operational behavior data. The resource scheduling module can then perform resource scheduling based on the task perception information, allocating core resources to high-frequency active tasks, idle resources to low-frequency active tasks, and minimum guaranteed resources to triggered critical point verification tasks. The closed-loop collaboration module then performs verification according to the resource scheduling results and the engineering drawings, and performs closed-loop control during the verification process based on the task progress log and resource scheduling results. The system optimizes the utilization of modeling and verification resources by distinguishing between critical point and non-critical point verifications and performing differentiated resource scheduling based on the high-frequency and low-frequency active states of modeling.

[0147] It should be noted that other corresponding descriptions of the functional units involved in the real-time verification system for 3D models based on engineering drawings provided in the embodiments of this application can be found in the corresponding descriptions in the real-time verification method for 3D models based on engineering drawings provided in the above embodiments, and will not be repeated here.

[0148] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0149] The embodiments described above are merely examples of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these modifications and improvements all fall within the protection scope of this application.

Claims

1. A method for real-time checking of a three-dimensional model based on engineering drawings, characterized in that, The method includes: Acquire modeling terminal operation behavior data, the operation behavior data including modeling action types; Based on the operational behavior data, task awareness information is generated, including the activity level of modeling operations, verification task attributes, and task progress logs. The verification task attributes are used to define the task to be verified as a critical point verification task or a non-critical point verification task. The task progress logs are used to record the task progress of non-critical point verification tasks. Generating task awareness information based on the operational behavior data includes: extracting the modeling action type of the modeling operation from the operational behavior data; querying the resource consumption of the modeling operation based on the modeling action type; marking the action category of the modeling operation based on the resource consumption, the action category including high-consumption actions and low-consumption actions; extracting modeling action statistics within a preset time window according to the action category, the modeling action statistics including the frequency of occurrence of high-consumption actions, the duration of a single action, and the execution interval of low-consumption actions; calculating the modeling operation intensity value based on the modeling action statistics; and determining the activity level of the modeling operation based on the modeling operation intensity value. Resource scheduling is performed based on the task awareness information to generate resource scheduling results; wherein, when the modeling operation activity level is high frequency, core resources are allocated to the task to be verified; when the modeling operation activity level is low frequency, idle resources are allocated to the task to be verified; when the verification task attribute is a key point verification task and the task to be verified is detected to be triggered, minimum guaranteed resources are allocated to the task to be verified. According to the resource scheduling results, the task to be verified is verified using engineering drawings; and, according to the task progress log and the resource scheduling results, closed-loop control is performed. The closed-loop control is used to perform resource reallocation scheduling when the verification task at the critical point is completed; and to perform dynamic resource backfilling when the verification task at the non-critical point is completed or when the modeling operation restarts high-frequency activity.

2. The method according to claim 1, characterized in that, The modeling operation intensity value is calculated based on the modeling action statistics, including: The total duration of high-consumption actions within the preset time window is obtained by summing the duration of each single action. Count the number of times high-consumption actions are executed consecutively within the preset time window; The continuous execution fine-tuning factor is calculated based on the number of consecutive executions. When high-consumption actions are executed consecutively, the continuous execution fine-tuning factor is the ratio of the number of consecutive executions to the frequency of occurrence. When high-consumption actions are not executed consecutively, the continuous execution fine-tuning factor is a preset fixed value. Obtain the average value of the execution interval within the preset time window; The modeling operation intensity value is calculated by combining the average value. The modeling operation intensity value is the product of the total duration of the high-consumption action and the fine-tuning factor, and then the ratio of the average value.

3. The method according to claim 2, characterized in that, Determining the activity level of modeling operations based on the modeling operation intensity value includes: Calculate the average duration of a high-consumption action based on the duration of a single action within a preset time window; If the intensity value of the modeling operation is greater than the average duration of a single operation, the activity level of the modeling operation is determined to be high-frequency activity. If the intensity value of the modeling operation is less than or equal to the average duration of a single operation, the activity level of the modeling operation is determined to be low-frequency active.

4. The method according to claim 1, characterized in that, Based on the operational behavior data, task-aware information is generated, including: Extract the component hierarchy information marked in the engineering drawings, wherein the component hierarchy information includes main components and non-main components; Obtain the model component corresponding to the task to be verified; A decision set is defined based on the component hierarchy information; wherein, when the model component is a main component marked in the component hierarchy information, the task to be verified is assigned to the priority decision set; when the model component is a non-main component marked in the component hierarchy information, the task to be verified is assigned to the non-priority decision set. The correlation of the task to be verified is detected by analyzing the topological structure of the three-dimensional model corresponding to the model components. The verification task attributes are determined according to the judgment set and the correlation degree, wherein the verification tasks marked as high correlation tasks in the priority judgment set are defined as key point verification tasks.

5. The method according to claim 4, characterized in that, By analyzing the topological structure of the 3D model corresponding to the model components, the correlation of the task to be verified is detected, including: Iterate through the number of connections between the model component and other components; Based on a preset spatial dependency judgment criterion, the number of dependencies is counted; the spatial dependency judgment criterion includes at least two of the following: assembly constraints between the model component and other components, shared geometric datum, and layout influence. When the number of connections is greater than or equal to the number of connections threshold, and the number of dependencies is greater than or equal to the number of dependencies threshold, the task to be verified is set as a high-relevance task. When the number of connections is less than the connection number threshold, or when the number of dependencies is less than the dependency number threshold, the task to be verified is set as a low-relevance task.

6. The method according to claim 1, characterized in that, Resource scheduling is performed based on the task-aware information to generate resource scheduling results, including: Based on the task awareness information, priority verification tasks and non-priority verification tasks are determined; the priority verification tasks include the verification tasks whose modeling operation activity level is high frequency and active, or the verification task attributes are key point verification tasks and the verification tasks are triggered. By calling the system resource isolation interface, the execution resources allocated to the priority verification task are physically isolated. The execution resources include core resources and minimum guaranteed resources. The minimum guaranteed resources are the minimum dedicated resources required to maintain the basic operation of the critical point verification task. The priority verification task is bound to the first verification process so that the first verification process can perform verification on the priority verification task. The execution status of the second verification process corresponding to the non-priority verification task is frozen, and the resources corresponding to the second verification process are released to the modeling process.

7. The method according to claim 6, characterized in that, Based on the resource scheduling results, the task to be verified is validated using engineering drawings, including: Resource usage data of the modeling process can be obtained through the system resource monitoring interface; Based on the resource usage data, the available idle resources are calculated. The idle time resources are allocated to the second verification process; Obtain historical frozen data of the non-priority verification task, the historical frozen data including process context and real-time progress; Using the second verification process, the non-priority verification task is verified according to the historical frozen data.

8. The method according to claim 7, characterized in that, Based on the task progress log and the resource scheduling results, closed-loop control is performed, including: Monitor the verification progress of key points in the task progress log; When the key point verification task is completed, the resource isolation status of the key point verification task is released. Reclaim the resources occupied by the key point verification task; When the modeling process is in a low-frequency active state and there are non-critical point verification tasks that have not been verified, resources are allocated to the non-critical point verification tasks according to the number of uncompleted tasks recorded in the task progress log and the occupied resources.

9. A real-time verification system for 3D models based on engineering drawings, characterized in that, The system includes: A dual-state perception module is used to acquire modeling terminal operation behavior data and generate task perception information based on the operation behavior data. The operation behavior data includes modeling action types. The task perception information includes modeling operation activity level, verification task attributes, and task progress log. The verification task attributes are used to define the task to be verified as a critical point verification task or a non-critical point verification task. The task progress log is used to record the task progress of non-critical point verification tasks. Generating task perception information based on the operation behavior data includes: extracting the modeling action type of the modeling operation from the operation behavior data; querying the resource consumption of the modeling operation based on the modeling action type; marking the action category of the modeling operation based on the resource consumption, the action category including high-consumption actions and low-consumption actions; extracting modeling action statistics within a preset time window according to the action category, the modeling action statistics including the occurrence frequency of high-consumption actions, the duration of a single action, and the execution interval of low-consumption actions; calculating the modeling operation intensity value based on the modeling action statistics, and determining the modeling operation activity level based on the modeling operation intensity value. The resource scheduling module is used to perform resource scheduling based on the task awareness information to generate resource scheduling results; wherein, when the modeling operation activity level is high frequency, core resources are allocated to the task to be verified; when the modeling operation activity level is low frequency, idle resources are allocated to the task to be verified; when the verification task attribute is a key point verification task and the task to be verified is detected to be triggered, minimum guaranteed resources are allocated to the task to be verified. The closed-loop collaboration module is used to perform verification on the task to be verified using engineering drawings according to the resource scheduling results; and to perform closed-loop control according to the task progress log and the resource scheduling results. The closed-loop control is used to perform resource reallocation scheduling when the verification task at the critical point is completed; and to perform dynamic resource backfilling when the verification task at the non-critical point is completed or when the modeling operation restarts high-frequency activity.

Images