Three-dimensional model rendering system and rendering method

Abstract

The three-dimensional model rendering system comprises a management system, a three-dimensional model rendering server and terminal equipment, and the management system, the three-dimensional model rendering server and the terminal equipment communicate through a network and cooperatively complete three-dimensional model rendering. Through the arrangement of the authority control module, the minimum authority MinV required for completing the request can be calculated in real time according to the user authority V and the interface ID carried in the rendering request. Therefore, the request is confirmed to be valid only when V is not lower than MinV. The causal logic has the remarkable beneficial effects that the authority control is improved from simple'access' to a refined level of'access at what precision / cost ', illegal or unauthorized rendering requests are effectively prevented, the security of core digital assets is greatly protected, and the security of the core digital assets is greatly improved. And invalid occupation and abuse of system rendering resources are avoided from the source.

Description

Technical Field

[0001] This invention relates to the field of 3D model rendering, and more particularly to a 3D model rendering system and rendering method. Background Technology

[0002] 3D model rendering technology, as a crucial bridge connecting the digital world and physical reality, has become an indispensable core engine for modern scientific computing and the cultural and creative industries. Its applications have expanded from early scientific visualization and film special effects to deeply penetrate current cutting-edge fields such as industrial design, Building Information Modeling (BIM), smart cities, next-generation games, and even the metaverse. With the exponential growth of model complexity, scene scale, and user interaction demands, traditional single-machine rendering modes have shown serious bottlenecks in terms of computing power, data storage, and collaborative sharing. To address this challenge, the industry's technological paradigm has generally shifted to distributed rendering and cloud rendering architectures based on networked and clustered architectures.

[0003] However, it has revealed serious inherent deficiencies in the refined management of system resources and service quality assurance, specifically in the following aspects: (1) The access control mechanism is too static and crude. The role-based access control (RBAC) model used in the existing technology is essentially a "one-size-fits-all" static authorization mode, which only determines whether a user has the permission to perform a specific operation (such as "rendering") based on the user's fixed role. This model cannot perceive and evaluate the resource consumption inherent in a single rendering request. Whether it is requesting to render a simple geometry with low resolution or requesting to render a highly complex scene with complex ray tracing and a large number of polygons, as long as the user's role and permissions match, there is no difference between the two in the eyes of the access control module. This lack of dynamic perception mechanism leads to a significant "resource perception blind spot", which fails to achieve on-demand and fine-grained access control of system resources. As a result, legitimate users may unintentionally or maliciously submit rendering tasks that consume extreme resources in a way that is disproportionate to their permission level, resulting in the system's lack of accurate identification and effective interception capabilities for such behavior during the access phase.

[0004] (2) Lack of intelligent dynamic adjustment capability for task priorities. Due to the information gap between the access control module and the task scheduling engine, once a request passes the admission verification, the initiator's identity and permission information are discarded in the subsequent scheduling process. When arranging queues and allocating resources, the task scheduling engine mainly relies on the submission sequence of tasks and the real-time load of the server, and cannot effectively distinguish whether a task entering the queue comes from a high-priority paid user or a low-privilege free user. More seriously, the system lacks a memory and feedback mechanism for users' historical behavior. For example, a user who frequently makes unreasonable requests may be frequently rejected by the system or cause rendering timeouts, but their subsequent requests will not be subject to any punitive adjustment in priority, and they can still queue "fairly" in the same queue as high-priority users with good behavior. This design, which lacks feedback on unreasonable behavior, reduces the service response efficiency and the ability to protect the user experience.

[0005] (3) Blindness and unfairness in resource allocation. The fragmented design leads to a lack of intelligent judgment in resource allocation. When the system load is low, the above defects may not be easily noticed, but during peak periods of high concurrency and heavy load, these problems become more obvious. High-load rendering tasks submitted by low-privilege users may occupy valuable computing nodes for a long time, thereby blocking lightweight tasks submitted by high-priority users that should be processed first, ultimately leading to a degradation of the Service Level Agreement (SLA). The system cannot proactively and forward-lookingly allocate limited computing resources to high-value users or high-priority tasks, nor can it effectively suppress potential resource abuse. This not only weakens the stability and fairness of the rendering service, but also affects the ability to operate commercially, constituting a core bottleneck of existing technical solutions in practical applications. Summary of the Invention

[0006] The purpose of this invention is to provide a three-dimensional model rendering system and rendering method to solve the problems mentioned in the background art, such as the inability to perceive and evaluate the resource consumption inherent in a single rendering request, the lack of feedback design, and the lack of intelligent judgment in resource allocation.

[0007] To achieve the above objectives, the present invention provides the following technical solution: A 3D model rendering system includes a management system, a 3D model rendering server, and terminal devices. The management system, the 3D model rendering server, and the terminal devices communicate with each other via a network and work together to complete the 3D model rendering. The operation process of the 3D model rendering system is as follows: S1. The terminal device is used to generate a 3D model rendering request and to receive and display the rendering result processed by the 3D model rendering server. S2. The management system receives the 3D model rendering request from the terminal device in step S1 and generates a 3D rendering task. The management system includes an access control module, a task scheduling module, and an access behavior priority adjustment module. The access control module includes a structure and input table and an interface resource consumption table. The structure and input table receives 3D model rendering requests from the terminal device. The rendering request includes at least an interface identifier and a user access level. Based on the interface identifier, the module obtains the corresponding resource consumption evaluation value from the interface resource consumption table and calculates the access level. The access control module compares the user access level in the 3D model rendering request with the calculated access level, and then sends the comparison result along with the original request to the task scheduling module for verification to generate a rendering task, which is then sent to the 3D model rendering server. S3. The 3D model rendering server receives the rendering task sent by the management system in step S2, adds it to the rendering task queue, and generates the 3D model rendering result for output according to the priority adjustment module scheduling of the rendering task; the 3D model rendering server is composed of a server cluster consisting of one or more computing nodes.

[0008] Preferably, the interface identifier points to at least one of the API interfaces for preview rendering, high-definition output, and ray tracing rendering, and the user permission level V is a value pre-set for each user account, representing its service level.

[0009] Preferably, the method for verifying the task scheduling module is as follows: Scenario 1: When the user permission level V ≥ the permission level MinV, the permission verification passes, and the task scheduling module will put it into a rendering task queue, which will then be sent to the rendering server. Scenario 2: When the user permission level V < the permission level MinV, the permission verification fails, and a secondary judgment mechanism is initiated.

[0010] Preferably, the secondary judgment mechanism is to determine whether it is a peak period, and the peak period is determined by at least one of the following: a time preset by the system administrator or the average CPU utilization of the cluster exceeds 80%; If it is determined to be a peak period, the current request will be rejected immediately, and an insufficient permission prompt will be returned to the terminal device to protect system resources from being occupied by non-compliant requests during peak periods; If the current period is determined to be "off-peak", a lenient strategy is adopted, and the request is placed in the rendering task queue and marked as "low priority". The "low priority" task will only be scheduled and executed after all normal priority tasks in the queue have been processed.

[0011] Preferably, the interface resource consumption table sets a standardized resource consumption evaluation value R for each interface identifier. The R value is a dimensionless numerical value, which is pre-calibrated by the system administrator or through performance testing, and evaluates the system resources required to complete a typical task by calling the interface, including at least one CPU, GPU, memory, and network bandwidth system resources.

[0012] Preferably, the terminal device includes a user interface module, a request generation and encapsulation module, a network communication module, and a result receiving and decoding module. The user interface module is used to provide a user-operable graphical interface (GUI) and generate rendering instructions through the graphical interface (GUI). The request generation and encapsulation module receives the rendering instruction generated by the user interaction interface module, and the request generation and encapsulation module collects the necessary parameter information to form a structured data packet and transmits it to the network communication module. The network communication module sends the request data packet generated by the encapsulation module to the management system according to a secure network transmission protocol such as HTTPS or WSS. The network communication module is responsible for all data exchange between the terminal device and the cloud system, and is also responsible for listening to and receiving the rendering result data generated by the 3D model rendering server, and managing the status of the network connection. The result receiving and decoding module receives the rendering result data, and the result receiving and decoding module parses, decodes and decompresses the rendering result data. The decoded final result is submitted to the user interaction interface module for real-time display or saving.

[0013] Preferably, the graphical user interface (GUI) includes loading a 3D model, setting the scene, adjusting the camera view, configuring lighting parameters, selecting rendering quality, and finally submitting rendering commands.

[0014] Preferably, the parameter information includes at least the user's identity credentials such as an API token, an interface identifier selected and determined by the user, scene 3D data or its index, and all rendering configurations, and the parameter information is connected to a predefined communication protocol of the management system.

[0015] Preferably, the 3D model rendering server includes a rendering core module, a cache management module, and a transmission optimization module. The rendering core module generates a two-dimensional image that is a transformation of three-dimensional scene data by at least using a graphics rasterization rendering algorithm, a ray tracing rendering algorithm, and a path tracing rendering algorithm. The rendering core module is a unit that performs actual pixel calculations. The cache management module caches at least the 3D model, texture map, and material asset data in the local cache of the server, and the cache management module provides cached data to the rendering core module. The transmission optimization module uses adaptive bitrate streaming technology to compress the image or video stream in real time, such as H.264, H.265, or AV1 encoding, and then sends the rendering result back to the terminal device. It also dynamically adjusts the bitrate and resolution according to the network environment of the terminal device.

[0016] Preferably, the access behavior priority adjustment module is used to continuously record and analyze the historical request records of each user, and monitor the record analysis unit of the permission control module for each user request verification result, namely, whether the user permission level V ≥ the permission level MinV or the user permission level V < the permission level MinV, and monitor the verification result of the user in each task scheduling module. Based on the historical request records, the "historical access success rate" is dynamically calculated for each user by dividing the number of successful requests within a sliding time window by the total number of requests. The "historical access success rate" is converted into a priority weight and applied to the task queue of the task scheduling module.

[0017] The present invention also provides a rendering method for a three-dimensional model rendering system, used to implement the above-mentioned three-dimensional model rendering system, characterized in that: S202: The terminal device generates and submits a rendering request; S203: The management system receives the rendering request; S204: The management system performs permission verification on the rendering request and calculates the minimum permission threshold to determine whether the permission is satisfied. S205: If the permission verification passes, proceed with S206 - S209: S206: Perform dynamic priority scheduling; S207: Submit a task to the rendering server; S208: Perform rendering and return the result; S209: Perform adaptive optimization; If the permission verification fails, execute S210; S210: Returns insufficient permissions information, forming a rendering request processing flow that includes permission verification, dynamic scheduling, and adaptive optimization.

[0018] The beneficial effects of this invention are: This invention, through a permission control module, can calculate in real time the minimum permission (MinV) required to complete a rendering request based on the user permission (V) and interface ID carried in the request. Therefore, a request is only considered valid if V is not lower than MinV. This causal logic brings significant benefits: it elevates permission control from a simple "whether access is possible" to a more refined level of "whether access is possible with what precision / cost," effectively preventing illegal or unauthorized rendering requests. This not only greatly protects the security of core digital assets but also prevents the ineffective use and abuse of system rendering resources at the source.

[0019] This invention establishes a task scheduling module that schedules rendering requests that have passed permission verification according to preset queue / priority rules. Therefore, high-priority tasks can jump the queue or be processed first. This mechanism completely changes the previous indiscriminate queuing model, ensuring that rendering requests for critical business operations receive the fastest response. In high-concurrency scenarios, the system's average response time and effective throughput of task processing are significantly improved, resulting in a qualitative improvement in user experience.

[0020] This invention uniquely incorporates an access behavior priority adjustment module. This module continuously analyzes historical rendering task success rates, time consumption, and other historical behaviors, dynamically adjusting and outputting priority weights, which are then fed back to the task scheduling module. Therefore, the system can learn and evolve: for users or task types with high historical success rates and good performance, the system automatically increases their priority weights, and vice versa. This design forms an intelligent closed-loop feedback system, enabling system resources to continuously and automatically tilt towards the most efficient and stable direction, ultimately achieving a dynamic balance between ensuring rendering quality and improving overall system operating efficiency—a feature completely absent in existing technologies. Attached Figure Description

[0021] Figure 1 This is an overall diagram of a three-dimensional model rendering system and rendering method according to the present invention; Figure 2 This is a flowchart of a three-dimensional model rendering system and rendering method according to the present invention; Figure 3 This is a schematic diagram of an unordered task queue in the prior art; Figure 4 This is a schematic diagram of the ordered task queue of the present invention; Figure 5 This is a schematic diagram of the background monitoring interface of a three-dimensional model rendering system and rendering method according to the present invention; Figure 6 This is a schematic diagram illustrating the resource utilization monitoring of the present invention; Figure 7 This is a schematic diagram illustrating the trend of response time variation in this invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] Example 1 Please see Figure 1 The present invention provides a three-dimensional model rendering system, including a management system 100, a three-dimensional model rendering server 200 and a terminal device 300. The management system 100, the three-dimensional model rendering server 200 and the terminal device 300 communicate with each other through a network and work together to complete the three-dimensional model rendering. Terminal device 300 includes at least one of the following: personal computer, workstation, mobile device (such as mobile phone, tablet computer), or virtual reality (VR) / augmented reality (AR) head-mounted display device, possessing network communication and graphics display capabilities. 3D model rendering server 200, as the core computing unit of the system, is responsible for executing specific, computationally intensive rendering tasks. Management system 100 processes requests through a series of innovative internal verification, arbitration, and scheduling mechanisms, and efficiently assigns compliant tasks to 3D model rendering server 200. Management system 100 plays a crucial bridging role between terminal device 300 and rendering server 200.

[0024] The operation process of the 3D model rendering system is as follows: S1. The terminal device 300 is used to generate 3D model rendering requests and receive and display the rendering results processed by the 3D model rendering server 200. The terminal device 300 is the client through which the user interacts with the entire rendering system, and it is the initiator of the rendering request and the carrier for presenting the final result. It does not perform high-intensity rendering calculations itself, but focuses on providing a smooth user experience and efficient data communication.

[0025] S2. The management system 100 receives the 3D model rendering request from the terminal device 300 in step S1 and generates a 3D rendering task. This includes generating a 3D rendering task queue; The management system 100 includes an access control module 110, a task scheduling module 120, and an access behavior priority adjustment module 130. The access control module 110 includes a structure and input table and an interface resource consumption table. The structure and input table receives the 3D model rendering request from the terminal device 300. The rendering request includes at least an interface identifier and a user access level. Based on the interface identifier, the corresponding resource consumption evaluation value in the interface resource consumption table is obtained, and the access level is calculated. The access control module 110 compares the user access level in the 3D model rendering request with the calculated access level, and sends the comparison result along with the original request to the task scheduling module 120 for verification to generate a rendering task, which is then sent to the 3D model rendering server 200. The formula for calculating the permission level is: (1) Here, is the square root operation, and floor is the floor function. This formula is designed with non-linear characteristics: for tasks with low resource consumption R, the requirement for the MinV privilege level increases relatively gradually; while for tasks with high resource consumption R, the requirement for the MinV privilege level increases sharply. This design effectively limits low-privilege users' access to expensive system resources.

[0026] S3. The 3D model rendering server receives the rendering tasks sent by the management system in step S2, adds them to the rendering task queue, and generates 3D model rendering results for output according to the priority adjustment module's scheduled rendering tasks; the 3D model rendering server 200 consists of a server cluster composed of one or more computing nodes.

[0027] Preferably, the interface identifier points to at least one of the API interfaces for preview rendering, high-resolution output, and ray tracing rendering, and the user permission level is a pre-set value for each user account, representing its service level. For example, free users V=10, monthly fee users V=50, and enterprise users V=200.

[0028] Preferably, 3. The three-dimensional model rendering system and rendering method according to claim 1, characterized in that: the method for verifying the result by the task scheduling module 120 is as follows: Scenario 1: When the user's permission level V ≥ the permission level MinV, the permission verification passes, and a rendering task is generated and sent to the rendering server 200; this indicates that the user's permission level is sufficient to support the resource consumption requested. In this case, the task is considered compliant, and the task scheduling module 120 will place it in a main rendering task queue, waiting to be assigned to the rendering server 200. Its specific priority in the queue will be further affected by the access behavior priority adjustment module 130.

[0029] Scenario 2: When the user permission level V < the permission level MinV, the permission verification fails, and a secondary judgment mechanism is initiated. In this case, the system will not immediately reject the request as in existing technologies, but will instead initiate a secondary judgment mechanism.

[0030] Preferably, the secondary judgment mechanism is to determine whether it is a peak period, and the peak period is determined by at least one of the following: a time preset by the system administrator or the average CPU utilization of the cluster exceeds 80%; If a peak period is identified, the current request is immediately rejected, and an insufficient permission message is returned to the terminal device 300 to protect system resources from being occupied by non-compliant requests during peak hours. If the current period is determined to be a peak period, the request is immediately rejected, and an insufficient permission message is returned to the terminal device 300. This measure aims to resolutely protect system resources from being occupied by non-compliant requests during peak hours.

[0031] If the current period is determined to be "off-peak," a lenient strategy is adopted, placing the request in the rendering task queue and marking it as "low priority." This "low priority" task will only be scheduled for execution after all normal priority tasks in the queue have been processed. If the system determines that the current period is not peak, a more lenient strategy will be adopted. The request will not be rejected but will be placed in the rendering queue, but will be marked as "low priority." This means that the task will only be scheduled for execution after all normal priority tasks in the queue have been processed. This utilizes the system's idle resources without affecting the experience of normal users.

[0032] Preferably, the interface resource consumption table assigns a standardized resource consumption evaluation value R to each interface identifier. The R value is a dimensionless numerical value, pre-calibrated by the system administrator or through performance testing, and assesses the system resources (CPU, GPU, memory, network bandwidth) required to complete a typical task by calling the interface. For example, the R value for the "Preview Rendering" interface might be 5, the R value for the "High-Resolution Output" interface might be 60, and the R value for "4K Cinematic Ray Tracing Rendering" might be as high as 200.

[0033] Preferably, the terminal device 300 includes a user interface module 310, a request generation and encapsulation module 320, a network communication module 330, and a result receiving and decoding module 340. The user interface module 310 is used to provide a user-operable graphical user interface (GUI) and generate rendering instructions through the GUI; the user interface module 310 is a direct window for human-computer interaction, ensuring that the user can view it.

[0034] The request generation and encapsulation module 320 receives the rendering instruction generated by the user interface module 310, and collects necessary parameter information to form a structured data packet which is then transmitted to the network communication module 330; the request generation and encapsulation module 320 prepares for network transmission.

[0035] The network communication module 330 sends the request data packet generated by the encapsulation module 320 to the management system 100 according to a secure network transmission protocol such as HTTPS or WSS. The network communication module 330 is responsible for all data exchanges between the terminal device 300 and the cloud system, and is also responsible for listening to and receiving the rendering result data generated by the 3D model rendering server 200, and managing the network connection status; ensuring the confidentiality and integrity of the request data packet during transmission, and preventing data from being eavesdropped on or tampered with.

[0036] The result receiving and decoding module 340 receives the rendering result data. The module 340 parses, decodes, and decompresses the rendering result data. The final decoded result is then submitted to the user interface module 310 for real-time display or saving. For example, it decodes the received H.264 video stream into continuous image frames, or decodes PNG / JPG format image data into a displayable bitmap.

[0037] Preferably, the graphical user interface (GUI) includes loading a 3D model, setting a scene, adjusting the camera view, configuring lighting parameters, selecting rendering quality, and finally submitting rendering commands.

[0038] Preferably, the parameter information includes at least the user's identity credentials such as an API token, an interface identifier selected and determined by the user, scene 3D data or its index, and all rendering configurations, and the parameter information communicates with the management system 100 via a predefined communication protocol.

[0039] Preferably, the 3D model rendering server 200 includes a rendering core module 210, a cache management module 220, and a transmission optimization module 230. The rendering core module 210 generates a two-dimensional image that is a transformation of three-dimensional scene data by at least using a graphics rasterization rendering algorithm, a ray tracing rendering algorithm, and a path tracing rendering algorithm. The rendering core module 210 is a unit that performs actual pixel calculations. The cache management module 220 caches at least the 3D model, texture map and material asset data in the local cache of the server (such as in memory or SSD), and the cache management module 220 provides cached data to the rendering core module 210; The transmission optimization module 230 uses adaptive bitrate streaming technology to compress the image or video stream in real time (e.g., H.264, H.265, AV1 encoding), and then sends the rendering result back to the terminal device 300. It also dynamically adjusts the bitrate and resolution based on the network environment of the terminal device 300, thereby ensuring a smooth viewing experience for the user.

[0040] Preferably, the access behavior priority adjustment module 130 is used to continuously record and analyze the historical request records of each user, and monitor the verification results of the user by the task scheduling module 120 each time. Based on the historical request records, the "historical access success rate" is dynamically calculated for each user by dividing the number of successful requests within a sliding time window by the total number of requests. The "historical access success rate" is converted into a priority weight and applied to the task queue of the task scheduling module 120. For example, a user with a historical success rate as high as 99% will have their compliant tasks (i.e., tasks with V>= MinV) given a higher weight in the queue, allowing them to "jump the queue" before other compliant users with lower historical success rates. Conversely, a user who frequently attempts unauthorized operations and has a historical success rate of only 20% will have their task's priority weight in the queue lowered, even if one of their requests is compliant. This mechanism forms a positive incentive loop: encouraging users to make reasonable requests within their own permissions while punishing and suppressing probing and abusive system access behaviors.

[0041] The present invention also provides a control method for a three-dimensional model rendering system and a rendering method, for implementing the aforementioned three-dimensional model rendering system. S202: The terminal device generates and submits a rendering request; the user operates through the user interface module 310 of the terminal device 300, triggering the request generation and encapsulation module 320 to generate a structured rendering request containing the user identity identifier (userID), user permission level (V), request interface ID (interfaceID), and other rendering parameters (such as resolution, format, etc.). This request is securely submitted to the management system 100 through the network communication module 330.

[0042] S203: The management system receives the rendering request; the access control module 110 of the management system 100, as the entry point, receives the rendering request from the terminal device 300.

[0043] S204: The management system performs permission verification on the rendering request, calculates the minimum permission threshold to determine whether the permission is satisfied; the permission control module 110 performs fine-grained permission verification on the received request. The core of this step is to dynamically calculate the minimum permission threshold (MinV) required to execute this specific request.

[0044] S205: If the permission verification passes, proceed with S206 - S209: S206: Perform dynamic priority scheduling; A valid request that passes the permission check is forwarded to the task scheduling module 120. This module calculates a comprehensive priority score for the task, which is a weighted combination of the user's static priority and the dynamic historical behavior weights obtained from the access behavior priority adjustment module 130. The task is then inserted into a rendering task queue sorted by its comprehensive priority score.

[0045] S207: Submit a task to the rendering server; the task scheduling module 120 retrieves one or more tasks from the head of the queue according to the priority order of the queue, encapsulates them into an instruction format that the rendering server 200 can recognize, and submits them to the cache management module 220 or the rendering core module 210 of the 3D model rendering server 200 for execution.

[0046] S208: Execute rendering and return results; After rendering is completed, rendering server 200 returns the results, including detailed execution status information such as whether the task was successful, the rendering time, and the amount of resources consumed, to management system 100.

[0047] S209: Perform adaptive optimization; the access behavior priority adjustment module 130 captures and records the execution result returned by step S208, and updates the user's historical behavior weights based on the analysis results to affect the priority calculation of subsequent tasks. After this step is completed, the entire method flow ends, forming a closed loop.

[0048] If the permission verification fails, execute S210; S210: Returns insufficient permissions information, forming a rendering request processing flow that includes permission verification, dynamic scheduling, and adaptive optimization.

[0049] The following are three optimizations of control methods for a 3D model rendering system and rendering method. The first approach, based on the formula MinV = BasePermission * ComplexityFactor in S204 and the formula PriorityScore = (0.7 * UserStaticPriority) + (0.3 * HistoricalBehaviorWeight) in S206, allows administrators to observe through the backend monitoring interface 600 that the average system waiting time (area 602) remains stable at a low level, and the task queue (area 601) always maintains an orderly state with higher priority tasks at the front, thus avoiding chaos.

[0050] The second approach involves introducing the DataSensitivity factor based on the MinV calculation formula in S204. When a low-privilege user attempts to access a highly sensitive interface, their request will be deemed a failure in step S205, and an error will be returned in step S210. Furthermore, the administrator will see a "High-Risk Access Attempt Alert" on the backend monitoring interface 600. Simultaneously, in the user behavior reputation score area 603, the HistoricalBehaviorWeight of the malicious user will be immediately reduced to the minimum value and highlighted, thereby maximizing the protective capabilities of the access control module 110.

[0051] Thirdly, when the weight w1 of ServiceLevelBonus, representing the service level, is set to an extremely high 0.9 according to the S206 PriorityScore calculation formula, the administrator can set a monitoring item for the average waiting time of VIP tasks in the KPI area 602 of the backend monitoring interface 600, and verify that it always meets the requirements of the Service Level Agreement (SLA). The task queue area 601 will clearly show that all tasks marked as VIP are always arranged before non-VIP tasks, thus ensuring the absolute priority of VIP tasks 501a.

[0052] Example 2 Suppose there are two users in the system: User A is a professional designer with a permission level of V=150; User B is a free trial user with a permission level of V=20.

[0053] The system provides two rendering interfaces: "Quick Preview" with a resource consumption rating of R=10 and "Ultra-HD Output" with a resource consumption rating of R=90.

[0054] The system's peak hours are set to 9:00-18:00 on weekdays.

[0055] Scenario 1: User B requests "ultra-high definition output" during off-peak hours (such as 2 a.m.).

[0056] 1. Permission Control: User B (V = 20) requests "Ultra-High-Definition Image Generation" (R = 90). The permission control module 110 calculates the minimum required permission MinV = floor(5 * sqrt(90) + 1) = floor(5 * 9.48 + 1) = floor(47.4 + 1) = 48.

[0057] 2. Permission Comparison: V(20) < MinV(48), permission verification fails.

[0058] 3. Task Scheduling: The task scheduling module 120 receives the "failure" result and checks that the current time is 2 am, which is a non-peak period. Therefore, this request is accepted but marked as "low priority" and placed in the queue.

[0059] 4. Behavior Adjustment: The access behavior priority adjustment module 130 records a failed request from User B. Its historical access success rate decreases, and the priority weight of its future tasks will also be correspondingly reduced.

[0060] Scenario 2: User A requests "Ultra-High-Definition Image Generation" during the peak period (e.g., 10 am).

[0061] 1. Permission Control: User A (V = 150) requests "Ultra-High-Definition Image Generation" (R = 90). The calculated MinV is still 48.

[0062] 2. Permission Comparison: V(150) >= MinV(48), permission verification passes.

[0063] 3. Task Scheduling: The task scheduling module 120 receives the "pass" result and places this task in the main rendering queue.

[0064] 4. Behavior Adjustment: The access behavior priority adjustment module 130 records a successful request from User A. Assuming User A has always behaved well, its high success rate will give this task a very high priority weight, making it be processed preferentially in the queue.

[0065] Through this technical solution, the system no longer blindly processes all seemingly legitimate requests, but constructs a closed-loop and refined resource management and scheduling system that can perceive resource consumption, identify user value, and provide intelligent feedback based on user historical behavior, thereby greatly enhancing the stability, fairness, and overall operation efficiency of the system.

[0066] System Architecture and Main Components Applied to the Application in Embodiment 2 The 3D model rendering system of the present invention consists of the following modules: Management System (100): The core, responsible for permission verification and task scheduling. It internally includes: Access control module (110): Implements dynamic access control and resource consumption monitoring.

[0067] Task scheduling module (120): Responsible for intelligent scheduling of tasks to be processed.

[0068] Access behavior priority adjustment module (130): Dynamically adjusts task priority based on user history behavior.

[0069] 3D model rendering server (200): Executes specific rendering tasks, transforms user requests into visual results, and returns data after rendering is complete.

[0070] Terminal device (300): User interface, responsible for user interaction and task submission, and displaying rendering results.

[0071] As shown in the diagram below, the system architecture is briefly described as follows: User ---> Terminal Device ---> Management System ---> Rendering Server Regarding performance test data To verify the system performance of this invention, comprehensive tests were conducted on rendering tasks under different environments. The results are shown in the table below: Table 1 Statistical Table of Rendering Task Test Information The above data clearly demonstrates the significant improvements of this invention in terms of task response speed, resource utilization, and request processing efficiency, indicating its effectiveness and advantages in practical applications.

[0072] In practical applications, the system can flexibly adjust task scheduling priorities based on the request content and current resource status. For example: User A (a professional designer) submitted a high-value task, passed the permission verification, and received priority processing.

[0073] User B (a free user) submitted a resource-intensive request, but the permission verification failed. During off-peak hours, the system processed the request with low priority.

[0074] The above embodiments demonstrate the ability of the present invention to respond to different user needs and its flexibility in resource management.

[0075] The various performance tests applied in Example 2 are as follows: 1. Testing Standards GB / T 19823-2005: Standard for Testing the Performance of Computer Graphics GB / T 20843-2014: Performance Evaluation of Computer Systems ISO 9193:2011: Visualization performance evaluation and rendering system testing Enterprise Standard No. Q / YZ 2023001: Performance Evaluation Standard for 3D Graphics Rendering Systems 2. Measurement methods Performance benchmark testing The performance of the rendering system is benchmarked using standard testing tools (such as SPECviewperf, GLBenchmark, etc.). The test content includes frame rate, rendering time, and scene complexity handling capabilities. The test scene is set according to the actual application.

[0076] 3 Delay Test Use response time measurement tools (such as JMeter and LoadRunner) to simulate concurrent user requests and measure the system's response time under different user concurrency.

[0077] Resource utilization monitoring Use system monitoring software (such as Prometheus and Grafana) to monitor the utilization of resources such as CPU, GPU, memory, and disk I / O in real time, and analyze their utilization efficiency through the generated charts.

[0078] 4. Stability Test The system's stability under high load was observed through long-term stress testing (72 hours of continuous operation), and any crashes, anomalies, or performance degradation events were recorded.

[0079] 5. Security Testing Use penetration testing tools (such as OWASP ZAP) to conduct security tests on the system and evaluate its resistance to attacks and data protection capabilities.

[0080] 6. Test Results The following is a summary of the performance test results for the 3D model rendering system of the present invention. The specific data is shown in the table below: Table 2 Summary of Performance Test Results 7. Summary Based on the above test results, the 3D model rendering system of the present invention has shown significant advantages in performance, efficiency and security, meets the expected technical indicators, and can effectively meet the needs of high-load and high-complexity business scenarios.

[0081] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions or improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A three-dimensional model rendering system, characterized in that: It includes a management system (100), a 3D model rendering server (200), and a terminal device (300). The management system (100), the 3D model rendering server (200), and the terminal device (300) communicate with each other via a network and work together to complete the 3D model rendering. The operation process of the 3D model rendering system is as follows: S1. The terminal device (300) is used to generate a three-dimensional model rendering request and to receive and display the rendering result processed by the three-dimensional model rendering server (200); S2. The management system (100) receives the three-dimensional model rendering request from the terminal device (300) in step S1 and generates a three-dimensional rendering task. The management system (100) includes an access control module (110), a task scheduling module (120), and an access behavior priority adjustment module (130). The access control module (110) includes a structure and input and an interface resource consumption table. The structure and input receive the three-dimensional model rendering request from the terminal device (300). The rendering request includes at least an interface identifier and a user permission level. Based on the interface identifier, the corresponding resource consumption evaluation value in the interface resource consumption table is obtained, and the permission level is calculated. The access control module (110) compares the user permission level in the three-dimensional model rendering request with the calculated permission level, and sends the comparison result along with the original request to the task scheduling module (120) for verification to generate a rendering task, which is then sent to the three-dimensional model rendering server (200). S3. The three-dimensional model rendering server (200) receives the rendering task sent by the management system (100) in step S2, adds it to the rendering task queue, and generates the three-dimensional model rendering result according to the rendering task scheduled by the priority adjustment module (130) for output. The 3D model rendering server (200) consists of a server cluster consisting of one or more computing nodes.

2. A three-dimensional model rendering system according to claim 1, characterized in that: The interface identifier refers to at least one of the API interfaces for preview rendering, high-definition output, and ray tracing rendering. The user permission level is a value pre-set for each user account, representing its service level.

3. The three-dimensional model rendering system according to claim 1, characterized in that: The method for verifying the results of the task scheduling module (120) is as follows: Case 1: When the user's permission level is greater than or equal to the permission level MinV, the permission verification passes and the rendering task is sent to the rendering server (200). Scenario 2: When the user's permission level is less than the permission level MinV, the permission verification fails, and a secondary judgment mechanism is initiated.

4. A three-dimensional model rendering system according to claim 3, characterized in that: The secondary judgment mechanism is to determine whether it is a peak period. The peak period is determined by at least one of the following methods: the system administrator presets a time or the average CPU utilization of the cluster exceeds 80%. If it is determined to be a peak period, the current request is immediately rejected, and an insufficient permission prompt is returned to the terminal device (300) to protect system resources from being occupied by non-compliant requests during peak periods; If the current period is determined to be "off-peak", a lenient strategy is adopted, and the request is placed in the rendering task queue and marked as "low priority". The "low priority" task will only be scheduled and executed after all normal priority tasks in the queue have been processed.

5. A three-dimensional model rendering system according to claim 1, characterized in that: The interface resource consumption table assigns a standardized resource consumption evaluation value R to each interface identifier. The R value is a dimensionless numerical value, which is pre-calibrated by the system administrator or through performance testing, and evaluates the system resources required to complete a typical task by calling the interface, including at least one CPU, GPU, memory, and network bandwidth system resources.

6. A three-dimensional model rendering system according to claim 1, characterized in that: The terminal device (300) includes a user interface module (310), a request generation and encapsulation module (320), a network communication module (330), and a result receiving and decoding module (340). The user interface module (310) is used to provide a user-operable graphical interface and generate rendering instructions through the graphical interface. The request generation and encapsulation module (320) receives the rendering instruction generated by the user interaction interface module (310), and the request generation and encapsulation module (320) collects the necessary parameter information and forms a structured data packet to be transmitted to the network communication module (330). The network communication module (330) sends the request to generate a structured data packet from the encapsulation module (320) to the management system (100) according to a secure network transmission protocol. The network communication module (330) is responsible for all data exchanges between the terminal device (300) and the cloud system, and is also responsible for listening to and receiving the rendering result data generated by the 3D model rendering server (200) and managing the network connection status. The result receiving and decoding module (340) receives the rendering result data. The result receiving and decoding module (340) parses, decodes and decompresses the rendering result data. The final result after decoding is submitted to the user interaction interface module (310) for real-time display or saving.

7. A three-dimensional model rendering system according to claim 6, characterized in that: The graphical interface includes loading a 3D model, setting a scene, adjusting the camera angle, configuring lighting parameters, selecting rendering quality, and finally submitting rendering instructions. The parameter information includes at least the user's identity credentials, the interface identifier selected by the user, the scene's 3D data or its index, and all rendering configurations. The parameter information forms a predefined communication protocol with the management system (100) and is encapsulated into a structured data packet to prepare for network transmission.

8. A three-dimensional model rendering system according to claim 1, characterized in that: The 3D model rendering server (200) includes a rendering core module (210), a cache management module (220), and a transmission optimization module (230). The rendering core module (210) generates a two-dimensional image that is a transformation of three-dimensional scene data through at least a graphics rasterization rendering algorithm, a ray tracing rendering algorithm and a path tracing rendering algorithm. The rendering core module (210) is a unit that performs actual pixel calculations. The cache management module (220) caches at least the 3D model, texture map and material asset data in the local cache of the server, and the cache management module (220) provides cached data to the rendering core module (210); The transmission optimization module (230) uses adaptive bitrate streaming technology to compress the image or video stream in real time, and then sends the rendering result back to the terminal device (300), and dynamically adjusts the bitrate and resolution according to the network environment of the terminal device (300).

9. A three-dimensional model rendering system according to claim 1, characterized in that: The access behavior priority adjustment module (130) is used to continuously record and analyze the historical request records of each user, and monitor the verification results of each user by the task scheduling module (120). Based on the historical request records, the "historical access success rate" is dynamically calculated for each user by dividing the number of successful requests within a sliding time window by the total number of requests. The "historical access success rate" is converted into a priority weight and applied to the task queue of the task scheduling module (120).

10. A rendering method for a three-dimensional model rendering system, used to implement the three-dimensional model rendering system of any one of claims 1 to 10, characterized in that: S202: The terminal device generates and submits a rendering request; S203: The management system receives the rendering request; S204: The management system performs permission verification on the rendering request and calculates the minimum permission threshold to determine whether the permission is satisfied. S205: If the permission verification passes, proceed with S206 - S209: S206: Perform dynamic priority scheduling; S207: Submit a task to the rendering server; S208: Perform rendering and return the result; S209: Perform adaptive optimization; If the permission verification fails, execute S210; S210: Returns insufficient permissions information, forming a rendering request processing flow that includes permission verification, dynamic scheduling, and adaptive optimization.

Images