A ship cabin environment dynamic rendering method and device, electronic equipment and medium

Abstract

The application discloses a ship cabin environment dynamic rendering method and device, electronic equipment and medium, and relates to the technical field of computers. The application constructs a semantic-geometry dual-drive scene graph, fuses ship monitoring data flow to output dynamic priority control information, comprehensively calculates cabin node loading priority comprehensive scores based on multidimensional information, completes loading determination and detail level precision distribution, and completes model rendering output on demand. The application realizes intelligent on-demand loading rendering of a complex cabin scene of a ship on a resource-limited terminal, takes into account regular inspection and emergency disposal scenes, reduces memory and rendering overhead, and improves long-time operation stability of the system.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a method, apparatus, electronic device and medium for dynamic rendering of ship cabin environment. Background Technology

[0002] With the development of digital and intelligent ship operation and maintenance, there is a need to realize real-time browsing and rendering of large-scale 3D scenes of complex ship compartments on resource-constrained Augmented Reality (AR) terminals. Existing 3D rendering solutions for ship compartments mainly include full loading solutions, Level of Detail (LOD) solutions based on spatial distance, and static spatial partitioning loading solutions.

[0003] The full-load solution loads all cabin, equipment, and piping models of the entire ship into memory at once. Even if the user only needs to browse a part of the cabin, the remaining large number of models continue to reside in memory, which is very likely to cause memory overflow on domestic ARM tablets and low-power embedded augmented reality terminals. The detail level solution based on spatial distance only changes the number of rendering output patches, but multiple sets of detail level model data still reside in memory, which cannot reduce memory consumption. The static mesh / octree partitioning solution only relies on geometric spatial location to make loading decisions, only focusing on where the space is, without utilizing the semantic attributes of the ship's cabins themselves, ship damage control alarms, and the operational situation information such as the location of on-site personnel.

[0004] The existing technology has the following drawbacks: First, it has high memory consumption and large CPU culling computational overhead, which places high demands on terminal hardware configuration. Resource-constrained augmented reality terminals cannot achieve a usable frame rate. Second, in emergency damage control scenarios, it is impossible to prioritize loading of accident compartments and rescue route compartments. The initial loading wait time in emergency scenarios is long, affecting response decisions. Third, it lacks intelligent loading decisions that combine compartment semantics and real-time ship situation. After the model is loaded, there is no active unloading and recycling mechanism. Long-term maintenance operations will result in continuous memory increases and poor system stability. Fourth, when multiple people are working together, each terminal makes loading decisions independently, which can easily generate a large number of duplicate data requests and consume a lot of bandwidth. Summary of the Invention

[0005] This application provides a method, apparatus, electronic device, and medium for dynamic rendering of ship cabin environments, which at least solves the problems of existing ship cabin 3D rendering technology relying solely on spatial geometric information for loading decisions, resulting in high memory consumption, inability to combine ship damage control and personnel position and situation to achieve differentiated priority loading, long-term memory cannot be reclaimed, and difficulty in stable operation on resource-constrained augmented reality terminals.

[0006] This application provides a method for dynamic rendering of ship cabin environment, including: The ship's 3D model is subjected to semantic parsing of compartments and offline structured parsing to obtain compartment nodes carrying semantic attributes and the topological relationships of the compartment nodes. The topological relationships are then associated and mapped with the spatial octree geometric index to obtain a semantic-geometric dual-drive scene map. Acquire monitoring data streams from the ship, the monitoring data streams including at least ship damage control signals and personnel location data, control the loading priority of each compartment node based on the semantic-geometric dual-drive scenario map and the monitoring data streams, and output dynamic priority adjustment information; The loading priority comprehensive score of each compartment node is determined based on the semantic-geometric dual-drive scenario map, the dynamic priority control information, and the terminal perspective information. The loading judgment of the compartment node and the allocation of loading accuracy level are completed based on the loading priority comprehensive score. Based on the loading determination and loading accuracy level, the model is loaded and an augmented reality 2D rendered image is output.

[0007] Furthermore, the process of performing cabin semantic parsing and offline structured parsing on the ship's 3D model to obtain cabin nodes carrying semantic attributes and their topological relationships is then performed. These topological relationships are then mapped and fused with a spatial octree geometric index to obtain a semantic-geometric dual-drive scene atlas, including: Obtain the ship's 3D model, traverse the component tree of the ship's 3D model to perform semantic parsing, identify geometric entities labeled as compartments, bulkheads, and decks, extract the closed space boundary of each compartment consisting of walls, top plate, and bottom plate, extract the geometric bounding boxes of key equipment, and instantiate each compartment as a compartment node. Identify the connecting paths between the compartment nodes, and construct a compartment connection topology graph with compartments and key equipment as nodes and connecting paths as directed edges; the direction of the directed edges represents the passable direction between compartments, and each directed edge records the connection type, pass width and pass height; Each compartment node is bound to a multidimensional semantic attribute in the form of a key-value dictionary to obtain the compartment node with the bound semantic attribute; the multidimensional semantic attribute includes compartment type, hazard level, deck level, compartment area and volume, carrying capacity, and a list of equipment contained in the compartment; Obtain the pre-built spatial octree index, use the closed spatial boundary of each cabin node as the query range to retrieve the leaf nodes that intersect with it in the octree, and bind the set of model components associated with the intersecting leaf nodes to the corresponding cabin node. The cabin connection topology, cabin nodes with bound multi-dimensional semantic attributes, and spatial octree index are associated, mapped, and fused to output a semantic-geometric dual-drive scene map.

[0008] Furthermore, the step of controlling the loading priority of each compartment node based on the semantic-geometric dual-drive scenario map and the monitoring data stream, and outputting dynamic priority adjustment information, includes: Obtain the configured task anchor personnel and obtain their real-time location from the monitoring data stream; Centered on the cabin where the personnel at the mission anchor point are located, adjacent cabins that extend outward along the directed edge are selected to define a common high-priority loading area, and the loading priority of each cabin node within the common high-priority loading area is increased. Each terminal overlays a personalized view area on the common high-priority loading area to form a terminal view area. Based on the terminal view area, the cabin nodes to be loaded are determined, and model distribution deduplication is performed to avoid multiple terminals repeatedly requesting the same model data. The results are then integrated to obtain complete dynamic priority control information.

[0009] Furthermore, the step of controlling the loading priority of each compartment node based on the semantic-geometric dual-drive scenario map and the monitoring data stream, and outputting dynamic priority adjustment information, also includes: When the monitoring data stream identifies a compartment damage control alarm signal, the compartment node with the alarm signal is marked as an alarm compartment node, and the loading priority of the alarm compartment node is set to the highest level. Based on the topological relationship of the cabin nodes, the loading priority of the cabin nodes that are topologically adjacent to the alarm cabin node is marked as the second highest level; Push priority update instructions to all relevant terminals and output the updated dynamic priority adjustment information.

[0010] Furthermore, before performing model loading based on the loading determination and loading accuracy level, the method further includes: Starting from the user's current location in the compartment node and ending at the current task objective, the optimal safe path to the destination is determined based on the monitoring data stream. The preloading priority of each compartment node on the optimal safe path and its topologically adjacent compartment nodes is increased.

[0011] Furthermore, the step of determining the optimal safe path to the destination based on the monitoring data stream, taking the user's current location as the starting point and the current task objective as the ending point, and increasing the preloading priority of each cabin node on the optimal safe path and its topologically adjacent cabin nodes, includes: The system obtains the location of the task target input by the user or the task target assigned by the system as the endpoint, and combines the topological relationship of the compartment nodes and the damage control status data in the monitoring data stream to assign passage weights to the directed edges that connect the topological compartment nodes to obtain a weighted topology graph. Starting from the user's current location in the compartment node and ending at the current task objective, the optimal weighted safe path is calculated using the Dijkstra shortest path algorithm on the weighted topology graph, and an ordered sequence of compartment nodes is output as the optimal safe path. When the damage control status data is updated, causing a change in the weight of the directed edge, the optimal safe path is recalculated. The cabin nodes traversed by the optimal safety path are marked as high-priority preloaded; the cabin nodes adjacent to the cabin nodes traversed by the optimal safety path are marked as medium-priority preloaded; the remaining non-cabin nodes are marked as low-priority; the cabin node model data marked as high-priority preloaded are added to the head of the preload queue, the cabin node model data marked as medium-priority preloaded are added to the middle of the preload queue, and the cabin node model data marked as low-priority preloaded are marked as releaseable after loading.

[0012] Furthermore, the step of determining the comprehensive score of loading priority for each compartment node based on the semantic-geometric dual-drive scene map, the dynamic priority control information, and the terminal view information, and completing the loading determination and loading accuracy level allocation for the compartment nodes based on the comprehensive score of loading priority, includes: Based on the semantic-geometric dual-drive scene map and the cabin node where the user is currently located, determine the spatial relationship between the user's current camera view frustum and the boundary of the cabin's enclosed space, and calculate the visual visibility score of the cabin node based on the spatial relationship between the user's current camera view frustum and the boundary of the cabin's enclosed space. The cabin type and functional level are determined based on the semantic attributes bound to the cabin node, and the cabin semantic importance score of the cabin node is determined based on the cabin type and functional level. Based on the compartment damage control alarm signals identified by the monitoring data stream, determine the damage control urgency score of the compartment node; Based on the high-priority preloading state, medium-priority preloading state, and low-priority state of each compartment node marked by the optimal safety path, the path passing through the compartment node is determined and scored. Based on the monitoring data stream, the task scenario is determined to be either routine inspection or emergency response. Based on the task scenario, the weight values ​​of the visual visibility score, the semantic importance score of the compartment, the urgency score of damage control, and the path traversal mark score are determined. The loading priority of each compartment node is determined by the visual visibility score and its weight, the semantic importance score of the compartment and its weight, the urgency score of damage control and its weight, and the path traversal mark score and its weight. The loading priority of each compartment node is ranked in descending order based on a comprehensive score, and the loading accuracy level of each compartment node is determined according to a preset loading threshold.

[0013] Furthermore, the method also includes: The monitoring data stream is continuously monitored to determine whether there are preset trigger events. The preset trigger events include at least one of the following: a significant change in viewpoint event, a significant movement of personnel position event, a damage control status update event, and a path replanning event. When the camera's orientation angle changes beyond a preset angle threshold or its position changes beyond a preset length threshold, a significant change in viewing angle event is triggered. When the change in the position of personnel at the task anchor point exceeds a preset distance threshold, a significant personnel movement event is triggered. When a new damage control alarm is received or an existing alarm level changes, a damage control status update event is triggered. When the weight of a directed edge changes, triggering the recalculation of the optimal safe path and generating a new path, a path replanning event is determined to be triggered. Once any preset trigger event is detected, the dynamic priority control information is re-executed, the comprehensive score of loading priority, loading judgment and loading accuracy level allocation of each compartment node are recalculated, and a new round of decision-making cycle begins.

[0014] Furthermore, the step of performing model loading based on the loading determination and the loading accuracy level includes: The terminal requests the model data packets for the corresponding compartment nodes from the external interface of the model management platform according to the loading priority order based on the loading determination; the accuracy of each compartment node is determined according to the loading accuracy level; the model data packet contains geometric mesh, material description and texture map; The model data package is loaded using an asynchronous, non-blocking method. After each frame is loaded, the terminal reports the loading completion status to the server. The loading completion status includes the cabin ID, loading accuracy level, and loading time. The terminal will unload loaded models that are marked as low priority and are not within any view frustum, freeing up memory.

[0015] This application also provides a dynamic rendering device for ship cabin environment, the device comprising: The atlas construction module is used to perform cabin semantic parsing and offline structured parsing processing on the ship's 3D model to obtain cabin nodes carrying semantic attributes and the topological relationships of cabin nodes. The topological relationships are associated and mapped with the spatial octree geometric index to obtain a semantic-geometric dual-drive scene atlas. The situation priority processing module is used to acquire the monitoring data stream in the ship. The monitoring data stream includes at least ship damage control signals and personnel location data. Based on the semantic-geometric dual-drive scenario map and the monitoring data stream, the module controls the loading priority of each compartment node and outputs dynamic priority adjustment information. The weighted decision module is used to determine the comprehensive score of the loading priority of each compartment node based on the semantic-geometric dual-drive scenario map, the dynamic priority control information, and the terminal perspective information, and to complete the loading judgment and loading accuracy level allocation of the compartment node based on the comprehensive score of the loading priority. The terminal rendering module is used to load the model and output an augmented reality 2D rendered image based on the loading determination and loading accuracy level.

[0016] This application also provides an electronic device, including: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of any of the above-described methods for dynamic rendering of ship cabin environments.

[0017] This application also provides a computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, it implements the steps of any of the above-described methods for dynamic rendering of ship cabin environments.

[0018] This application constructs a semantic-geometric dual-drive scene map, integrates dynamic priority control information from ship monitoring data stream output, calculates a comprehensive score for the loading priority of compartment nodes based on multi-dimensional information, completes loading determination and detail level precision allocation, and completes model rendering output on demand; it realizes intelligent on-demand loading and rendering of large scenes of complex ship compartments on resource-constrained terminals, taking into account both routine inspection and emergency response scenarios, reducing memory and rendering overhead, and improving the long-term stability of the system. Attached Figure Description

[0019] To more clearly illustrate the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is an application environment diagram of the dynamic rendering method for ship cabin environment in one embodiment of this application; Figure 2 This is a flowchart illustrating a method for dynamically rendering a ship cabin environment in one embodiment of this application. Figure 3 This is a logic diagram of a dynamic rendering method for ship cabin environment in one embodiment of this application; Figure 4This is a flowchart of the loading priority comprehensive scoring decision-making process in one embodiment of this application; Figure 5 This is a structural block diagram of a ship cabin environment dynamic rendering device in one embodiment of this application; Figure 6 This is an internal structural diagram of a computer device in one embodiment of this application. Detailed Implementation

[0021] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the protection scope of this application.

[0022] It should be noted that, in the description of this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. The terms "first," "second," etc., in this application are used to distinguish similar objects and are not used to describe a specific order or sequence.

[0023] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0024] The dynamic rendering method for ship cabin environments provided in this application can be applied to, for example... Figure 1 In the application environment shown, terminal 102 and server 104 communicate via a network. Terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices. Server 104 can be a standalone server or a server cluster consisting of multiple servers. Server 104 contains a 3D model of a ship and provides augmented reality 2D rendered images based on the perspective information of each terminal 102.

[0025] like Figure 2 , Figure 3 As shown, an embodiment of this application provides a method for dynamically rendering a ship's cabin environment, used on server 104. The method includes the following steps: Step S1: Perform semantic parsing and offline structured parsing on the 3D model of the ship to obtain the cabin nodes carrying semantic attributes and the topological relationships of the cabin nodes. Then, perform association mapping and fusion of the topological relationships with the spatial octree geometric index to obtain the semantic-geometric dual-drive scene map. Step S2: Obtain the monitoring data stream in the ship. The monitoring data stream includes at least ship damage control signals and personnel location data. Based on the semantic-geometric dual-drive scenario map and the monitoring data stream, control the loading priority of each compartment node and output dynamic priority adjustment information. Step S3: Determine the comprehensive score of loading priority for each compartment node based on the semantic-geometric dual-drive scene map, the dynamic priority control information, and the terminal perspective information; and complete the loading judgment and loading accuracy level allocation for the compartment node based on the comprehensive score of loading priority. Step S4: Based on the loading determination and loading accuracy level, perform model loading and output augmented reality 2D rendered image.

[0026] Specifically, by constructing a semantic-geometric dual-drive scene map, integrating dynamic priority control information from ship monitoring data stream output, and comprehensively calculating the loading priority score of compartment nodes based on multi-dimensional information, loading determination and detail level precision allocation are completed, and model rendering output is completed on demand. This enables intelligent on-demand loading and rendering of large scenes in complex ship compartments on resource-constrained terminals, taking into account both routine inspection and emergency response scenarios, reducing memory and rendering overhead, and improving the long-term stability of the system.

[0027] Specifically, during the system deployment phase, the server reads the ship's BIM / IFC 3D model, performs semantic parsing and offline structured parsing of compartments, extracts compartment and key equipment entities, constructs compartment nodes and topological relationships, and integrates these topological relationships with a spatial octree geometric index to generate a semantic-geometric dual-drive scene map. This map is persistently stored on the server-side model management platform. This step is executed only once during deployment and not repeatedly during terminal operation, reducing terminal computing power overhead. The server receives real-time ship monitoring data streams, including fire, water ingress, and breach damage signals, as well as personnel location data output by UWB / Bluetooth AOA. Based on the constructed semantic-geometric dual-drive scene map, loading priority is adjusted in conjunction with the monitoring data stream, and dynamic priority adjustment information is output. By comprehensively considering the scene map, dynamic priority adjustment information, and terminal camera perspective information, a comprehensive score for the loading priority of compartment nodes is calculated, determining which compartments need loading and at what level of detail. The terminal asynchronously requests model data packets according to the decision results and renders and outputs augmented reality 2D images.

[0028] This solution constructs a semantic-geometric dual-drive scene map, simultaneously carrying the semantic topology and spatial geometric index of the cabins; it accesses ship-specific damage control and personnel location monitoring data streams as rendering decision inputs, no longer relying solely on viewpoint distance; it integrates multi-dimensional information to make loading judgments and allocate detail level precision, enabling on-demand loading of ship cabin models, reducing memory usage and CPU / GPU rendering overhead, adapting to resource-constrained augmented reality terminals, and prioritizing the rendering of key cabins based on damage control and personnel location.

[0029] Furthermore, the process of performing cabin semantic parsing and offline structured parsing on the ship's 3D model to obtain cabin nodes carrying semantic attributes and their topological relationships is then performed. These topological relationships are then mapped and fused with a spatial octree geometric index to obtain a semantic-geometric dual-drive scene atlas, including: Obtain the ship's 3D model, traverse the component tree of the ship's 3D model to perform semantic parsing, identify geometric entities labeled as compartments, bulkheads, and decks, extract the closed space boundary of each compartment consisting of walls, top plate, and bottom plate, extract the geometric bounding boxes of key equipment, and instantiate each compartment as a compartment node. Identify the connecting paths between the compartment nodes, and construct a compartment connection topology graph with compartments and key equipment as nodes and connecting paths as directed edges; the direction of the directed edges represents the passable direction between compartments, and each directed edge records the connection type, pass width and pass height; Each compartment node is bound to a multidimensional semantic attribute in the form of a key-value dictionary to obtain the compartment node with the bound semantic attribute; the multidimensional semantic attribute includes compartment type, hazard level, deck level, compartment area and volume, carrying capacity, and a list of equipment contained in the compartment; Obtain the pre-built spatial octree index, use the closed spatial boundary of each cabin node as the query range to retrieve the leaf nodes that intersect with it in the octree, and bind the set of model components associated with the intersecting leaf nodes to the corresponding cabin node. The cabin connection topology, cabin nodes with bound multi-dimensional semantic attributes, and spatial octree index are associated, mapped, and fused to output a semantic-geometric dual-drive scene map.

[0030] The system identifies corridors, doors (watertight doors, fire doors), stairways (straight stairs, inclined stairs), and pipe tunnels (cable tunnels, pipeline tunnels) as connecting paths, using these paths as directed edges to construct a topological graph of the cabin connections. The octree has a maximum depth of 8 layers and a leaf node side length of 0.5m. The system uses the cabin's enclosed spatial boundary as a bounding box to query intersecting leaf nodes in the octree and binds them to corresponding model component sets. Finally, the system fuses the cabin nodes, the cabin connection topology graph, and the octree index to output a semantic-geometric dual-drive scene graph, which is persistently stored on the model management platform.

[0031] The system binds multi-dimensional semantic attribute information to each compartment node, including: compartment type (habitation compartment, engine room, cargo hold, bridge, fuel tank, etc.); hazard level (general area, restricted area, prohibited area, high-risk area); deck level (main deck, upper deck, lower decks 1-3, etc.); compartment area and volume; function (power, navigation, living quarters, storage, etc.); and a list of equipment contained in the compartment (equipment type, number, and a list of 3D component IDs). These semantic attributes are stored in the node data as a key-value pair dictionary for subsequent lookup by the weighted scoring model.

[0032] Furthermore, the step of controlling the loading priority of each compartment node based on the semantic-geometric dual-drive scenario map and the monitoring data stream, and outputting dynamic priority adjustment information, includes: Obtain the configured task anchor personnel and obtain their real-time location from the monitoring data stream; Centered on the cabin where the personnel at the mission anchor point are located, adjacent cabins that extend outward along the directed edge are selected to define a common high-priority loading area, and the loading priority of each cabin node within the common high-priority loading area is increased. Each terminal overlays a personalized view area on the common high-priority loading area to form a terminal view area. Based on the terminal view area, the cabin nodes to be loaded are determined, and model distribution deduplication is performed to avoid multiple terminals repeatedly requesting the same model data. The results are then integrated to obtain complete dynamic priority control information.

[0033] Specifically, the system uses the real-time location of designated personnel (such as on-site operation supervisors or key personnel, configured by the system administrator at the start of the task) as a dynamic "task anchor point." The system continuously tracks the location data of personnel at this anchor point (updating at a frequency of 1Hz) and automatically adjusts the model distribution strategy for all personnel: the compartment where the anchor point is located and its adjacent compartments (within a radius of 2 hops) are marked as a common high-priority loading area for all relevant terminals, and each terminal overlays its own user-specific view area on top of this. This mechanism avoids the waste of server bandwidth caused by multiple terminals repeatedly requesting the same data in collaborative operation scenarios, while ensuring that all collaborating personnel have a consistent scene perception foundation for the core task area.

[0034] The system dynamically adjusts the model distribution strategy for all personnel based on the location of the task anchor point. This avoids bandwidth waste caused by multiple terminals repeatedly loading the same data, improving overall loading efficiency in collaborative work scenarios. The terminal viewpoint area is determined by continuously capturing the user's perspective and predicting the future view frustum range, enabling dynamic on-demand loading and unloading of scene content. The system possesses real-time response capabilities to user behavior and situational changes, ensuring long-term stability and reliability.

[0035] Furthermore, the step of controlling the loading priority of each compartment node based on the semantic-geometric dual-drive scenario map and the monitoring data stream, and outputting dynamic priority adjustment information, also includes: When the monitoring data stream identifies a compartment damage control alarm signal, the compartment node with the alarm signal is marked as an alarm compartment node, and the loading priority of the alarm compartment node is set to the highest level. Based on the topological relationship of the cabin nodes, the loading priority of the cabin nodes that are topologically adjacent to the alarm cabin node is marked as the second highest level; Push priority update instructions to all relevant terminals and output the updated dynamic priority adjustment information.

[0036] When the ship's damage control monitoring system pushes a damage control alarm signal (fire, water ingress, breach, etc.) for a certain compartment, the system's damage control event processor automatically performs the following operations: Marks the loading priority of the compartment node as "emergency upgrade" (priority set to the highest level P0); based on the topology graph, expands outward 1-2 hops along the connected edges, synchronously upgrading the loading priority of topologically adjacent compartments (i.e., compartments directly connected to the accident compartment or connected via an intermediate compartment, such as adjacent compartments, corresponding compartments on upper and lower decks, corridors leading to the accident compartment, etc.) to the second-highest level P1; pushes priority update instructions to all relevant clients. This mechanism ensures that, from the perspective of emergency response personnel, the model of the accident area and evacuation / rescue route area is loaded and presented preferentially and completely, providing timely scene information support for emergency management decisions. Ship damage control alarm signals are directly connected to the rendering and loading decision chain; the accident compartment and topologically adjacent compartments related to rescue and evacuation automatically have their loading priority upgraded; emergency scenarios do not need to wait for the full model to load, and accident-related compartments are loaded and rendered preferentially, providing timely three-dimensional situational support for ship emergency response.

[0037] Damage control signals are directly integrated into the loading decision chain, automatically elevating the loading priority of the triggered compartment and its adjacent compartments to the highest level. The rendering system possesses emergency situation awareness and differentiated support capabilities, ensuring that critical areas receive priority loading support in emergency scenarios, providing timely and complete scene information support for emergency response.

[0038] Furthermore, before performing model loading based on the loading determination and loading accuracy level, the method further includes: Starting from the user's current location in the compartment node and ending at the current task objective, the optimal safe path to the destination is determined based on the monitoring data stream. The preloading priority of each compartment node on the optimal safe path and its topologically adjacent compartment nodes is increased.

[0039] Furthermore, the step of determining the optimal safe path to the destination based on the monitoring data stream, taking the user's current location as the starting point and the current task objective as the ending point, and increasing the preloading priority of each cabin node on the optimal safe path and its topologically adjacent cabin nodes, includes: The system obtains the location of the task target input by the user or the task target assigned by the system as the endpoint, and combines the topological relationship of the compartment nodes and the damage control status data in the monitoring data stream to assign passage weights to the directed edges that connect the topological compartment nodes to obtain a weighted topology graph. Starting from the user's current location in the compartment node and ending at the current task objective, the optimal weighted safe path is calculated using the Dijkstra shortest path algorithm on the weighted topology graph, and an ordered sequence of compartment nodes is output as the optimal safe path. When the damage control status data is updated, causing a change in the weight of the directed edge, the optimal safe path is recalculated. The cabin nodes traversed by the optimal safety path are marked as high-priority preloaded; the cabin nodes adjacent to the cabin nodes traversed by the optimal safety path are marked as medium-priority preloaded; the remaining non-cabin nodes are marked as low-priority; the cabin node model data marked as high-priority preloaded are added to the head of the preload queue, the cabin node model data marked as medium-priority preloaded are added to the middle of the preload queue, and the cabin node model data marked as low-priority preloaded are marked as releaseable after loading.

[0040] The weight calculation formula for the directed edge execution passage weight assignment is as follows: Edge weight = basic passage cost (path length × cabin type coefficient) + damage control penalty (if the edge connects to or passes through an alarm cabin, the weight is multiplied by the damage control penalty factor of 10.0) + danger level coefficient (if the edge passes through a restricted area / high-risk area, the weight is subject to additional penalty).

[0041] The system defines dynamic event interfaces in the server-side model management platform to connect with the ship damage control monitoring system (which receives damage control signals such as fire, water ingress, breaches, and smoke in real time via Modbus TCP or OPC UA protocols) and the personnel positioning system (which receives real-time location coordinate data of each person via UWB or Bluetooth AOA, with an update frequency of no less than 1Hz). It also defines two dynamic event types: "damage control trigger event" and "personnel location change event," as event sources to trigger recalculation of loading decisions.

[0042] Specifically, nodes in the cabins traversed by the optimal safe path (all nodes in the path) are marked as "high-priority preloading (hot)", topologically adjacent nodes (within one hop) of each node on the path are marked as "medium-priority preloading (warm)", and other non-path area nodes are marked as "low-priority (cold)". Model data for hot areas is added to the head of the preloading queue, warm areas are added to the middle of the queue, and loaded models in cold areas are marked as releaseable first. This mechanism ensures that the model of the area the user is about to reach is ready in advance when moving along the path, reducing loading wait during the movement.

[0043] like Figure 4 As shown, the step of determining the comprehensive score of loading priority for each compartment node based on the semantic-geometric dual-drive scene map, the dynamic priority control information, and the terminal view information, and then completing the loading determination and loading accuracy level allocation for the compartment nodes based on the comprehensive score of loading priority, includes: Based on the semantic-geometric dual-drive scene map and the cabin node where the user is currently located, determine the spatial relationship between the user's current camera view frustum and the boundary of the cabin's enclosed space, and calculate the visual visibility score of the cabin node based on the spatial relationship between the user's current camera view frustum and the boundary of the cabin's enclosed space. The cabin type and functional level are determined based on the semantic attributes bound to the cabin node, and the cabin semantic importance score of the cabin node is determined based on the cabin type and functional level. Based on the compartment damage control alarm signals identified by the monitoring data stream, determine the damage control urgency score of the compartment node; Based on the high-priority preloading state, medium-priority preloading state, and low-priority state of each compartment node marked by the optimal safety path, the path passing through the compartment node is determined and scored. Based on the monitoring data stream, the task scenario is determined to be either routine inspection or emergency response. Based on the task scenario, the weight values ​​of the visual visibility score, the semantic importance score of the compartment, the urgency score of damage control, and the path traversal mark score are determined. The loading priority of each compartment node is determined by the visual visibility score and its weight, the semantic importance score of the compartment and its weight, the urgency score of damage control and its weight, and the path traversal mark score and its weight. The loading priority of each compartment node is ranked in descending order based on a comprehensive score, and the loading accuracy level of each compartment node is determined according to a preset loading threshold.

[0044] Specifically, the system defines a multi-dimensional comprehensive weighted scoring model to calculate a comprehensive score for the loading priority of each compartment node in the scenario.

[0045] The formula for the overall score of loading priority is: S_total=α×S_visibility+β×S_semantic+γ×S_damage+δ×S_path; in: S_visibility is the visual visibility score: calculated based on the spatial relationship between the user's current camera frustum and the cabin bounding box. Cabins that are currently visible or about to enter the frustum receive the highest score, while cabins behind or far from the frustum receive the lowest score. The value range is [0,1].

[0046] S_semantic represents the semantic importance score of a cabin: it is pre-configured based on cabin type and functional level, with key functional cabins such as the cockpit, engine room, and damage control station receiving a higher base score than ordinary living quarters. The value range is [0,1].

[0047] S_damage is the damage control urgency score: triggered by a step damage control event, alarm compartment = 1.0, adjacent compartment = 0.7, non-adjacent but on the same deck = 0.3, and others = 0. The value range is [0,1].

[0048] S_path is the path tracing score: determined by the path calculation result, hot=1.0, warm=0.5, cold=0. The value range is [0,1].

[0049] α, β, γ, and δ are the weighting coefficients for each dimension, with initial default values ​​of 0.30, 0.20, 0.30, and 0.20 respectively. These values ​​can be dynamically adjusted by the system administrator based on the task scenario (routine inspection / emergency response). When a damage control alarm is detected, W_dam automatically increases to 0.50, W_vis decreases to 0.15, and the others are scaled proportionally.

[0050] The system sorts the calculated loading priority scores of each compartment from high to low and makes decisions based on preset loading thresholds: (a) S_total≥0.6: High-precision loading (LOD0, full-detail geometry + 2K texture); (b) 0.3≤S_total<0.6: Medium-precision loading (LOD1, simplified geometry + 1K texture); (c) 0.1≤S_total<0.3: Low-precision loading (LOD2, proxy body + 256px texture); (d) S_total<0.1: No loading or marked as unloadable. A maximum data size limit is set for loading per frame (e.g., 50MB). High-priority content exceeding this limit is queued for loading in subsequent frames.

[0051] Furthermore, the method also includes: The monitoring data stream is continuously monitored to see if there are any preset trigger events. The preset trigger events include at least one of the following: a significant change in viewpoint event, a significant movement of personnel position event, a damage control status update event, and a path replanning event. When the camera's orientation angle changes beyond a preset angle threshold or its position changes beyond a preset length threshold, a significant change in viewing angle event is triggered. When the change in the position of personnel at the task anchor point exceeds a preset distance threshold, a significant personnel movement event is triggered. When a new damage control alarm is received or an existing alarm level changes, a damage control status update event is triggered. When the weight of a directed edge changes, triggering the recalculation of the optimal safe path and generating a new path, a path replanning event is determined to be triggered. Once any preset trigger event is detected, the dynamic priority control information is re-executed, the comprehensive score of loading priority, loading judgment and loading accuracy level allocation of each compartment node are recalculated, and a new round of decision-making cycle begins.

[0052] Specifically, the client continuously monitors three types of triggering events: significant changes in user perspective—when the camera's orientation changes by more than 30° or its position changes by more than 2 meters; significant personnel movement—when personnel at the mission anchor point or other key personnel move by more than 5 meters; damage control status updates—when a new damage control alarm is received or an existing alarm level changes; and path replanning—when a path recalculation is triggered due to changes in topology map weights, resulting in a new path. After any event is triggered, the server recalculates the overall score and loading priority of all compartments, entering a new round of decision-making.

[0053] Furthermore, the step of performing model loading based on the loading determination and the loading accuracy level includes: The terminal requests the model data packets for the corresponding compartment nodes from the external interface of the model management platform according to the loading priority order based on the loading determination; the accuracy of each compartment node is determined according to the loading accuracy level; the model data packet contains geometric mesh, material description and texture map; The model data package is loaded in an asynchronous, non-blocking manner, which does not affect the rendering loop of the current frame; After each frame is loaded, the terminal reports the loading completion status to the server. The loading completion status includes the cabin ID, loading accuracy level, and loading time. The terminal will unload loaded models that are marked as low priority and are not within any view frustum, freeing up memory.

[0054] The client-side rendering engine requests the corresponding cabin model data packages (including geometric meshes, material descriptions, and texture maps) from the model management platform's HTTP / HTTPS interface according to the overall score ranking result and LOD precision allocation instructions calculated by the server, in priority order. The loading process adopts an asynchronous non-blocking method and does not affect the rendering loop of the current frame. After each frame is loaded, the client reports the loading completion status (cabin ID, loading precision level, and time taken) to the server for the server's global loading status tracking.

[0055] The above-mentioned dynamic rendering method for ship cabin environments constructs a semantic-geometric dual-drive scene map, integrates dynamic priority control information from ship monitoring data stream output, calculates a comprehensive score for cabin node loading priority based on multi-dimensional information, completes loading judgment and detail level precision allocation, and completes model rendering output on demand. This enables intelligent on-demand loading and rendering of complex ship cabin scenes on resource-constrained terminals, taking into account both routine inspection and emergency response scenarios, reducing memory and rendering overhead, and improving the long-term stability of the system.

[0056] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method.

[0057] In one embodiment, such as Figure 5 As shown, a dynamic rendering device 10 for ship cabin environment is provided, including: a map construction module 1, a situation priority processing module 2, a weighted decision module 3, and a terminal rendering module 4.

[0058] The atlas construction module 1 is used to perform cabin semantic parsing and offline structured parsing processing on the ship's three-dimensional model to obtain cabin nodes carrying semantic attributes and the topological relationships of cabin nodes. The topological relationships are then associated and mapped with the spatial octree geometric index to obtain a semantic-geometric dual-drive scene atlas.

[0059] The situation priority processing module 2 is used to acquire the monitoring data stream in the ship. The monitoring data stream includes at least ship damage control signals and personnel location data. Based on the semantic-geometric dual-drive scenario map and the monitoring data stream, the loading priority of each compartment node is controlled, and dynamic priority adjustment information is output.

[0060] The weighted decision module 3 is used to determine the comprehensive score of the loading priority of each compartment node based on the semantic-geometric dual-drive scenario map, the dynamic priority control information, and the terminal perspective information, and to complete the loading judgment and loading accuracy level allocation of the compartment node based on the comprehensive score of the loading priority.

[0061] The terminal rendering module 4 is used to load the model and output an augmented reality 2D rendered image based on the loading determination and loading accuracy level.

[0062] Furthermore, the process of performing cabin semantic parsing and offline structured parsing on the ship's 3D model to obtain cabin nodes carrying semantic attributes and their topological relationships is then performed. These topological relationships are then mapped and fused with a spatial octree geometric index to obtain a semantic-geometric dual-drive scene atlas, including: Obtain the ship's 3D model, traverse the component tree of the ship's 3D model to perform semantic parsing, identify geometric entities labeled as compartments, bulkheads, and decks, extract the closed space boundary of each compartment consisting of walls, top plate, and bottom plate, extract the geometric bounding boxes of key equipment, and instantiate each compartment as a compartment node. Identify the connecting paths between the compartment nodes, and construct a compartment connection topology graph with compartments and key equipment as nodes and connecting paths as directed edges; the direction of the directed edges represents the passable direction between compartments, and each directed edge records the connection type, pass width and pass height; Each compartment node is bound to a multidimensional semantic attribute in the form of a key-value dictionary to obtain the compartment node with the bound semantic attribute; the multidimensional semantic attribute includes compartment type, hazard level, deck level, compartment area and volume, carrying capacity, and a list of equipment contained in the compartment; Obtain the pre-built spatial octree index, use the closed spatial boundary of each cabin node as the query range to retrieve the leaf nodes that intersect with it in the octree, and bind the set of model components associated with the intersecting leaf nodes to the corresponding cabin node. The cabin connection topology, cabin nodes with bound multi-dimensional semantic attributes, and spatial octree index are associated, mapped, and fused to output a semantic-geometric dual-drive scene map.

[0063] Furthermore, the step of controlling the loading priority of each compartment node based on the semantic-geometric dual-drive scenario map and the monitoring data stream, and outputting dynamic priority adjustment information, includes: Obtain the configured task anchor personnel and obtain their real-time location from the monitoring data stream; Centered on the cabin where the personnel at the mission anchor point are located, adjacent cabins that extend outward along the directed edge are selected to define a common high-priority loading area, and the loading priority of each cabin node within the common high-priority loading area is increased. Each terminal overlays a personalized view area on the common high-priority loading area to form a terminal view area. Based on the terminal view area, the cabin nodes to be loaded are determined, and model distribution deduplication is performed to avoid multiple terminals repeatedly requesting the same model data. The results are then integrated to obtain complete dynamic priority control information.

[0064] Furthermore, the step of controlling the loading priority of each compartment node based on the semantic-geometric dual-drive scenario map and the monitoring data stream, and outputting dynamic priority adjustment information, also includes: When the monitoring data stream identifies a compartment damage control alarm signal, the compartment node with the alarm signal is marked as an alarm compartment node, and the loading priority of the alarm compartment node is set to the highest level. Based on the topological relationship of the cabin nodes, the loading priority of the cabin nodes that are topologically adjacent to the alarm cabin node is marked as the second highest level; Push priority update instructions to all relevant terminals and output the updated dynamic priority adjustment information.

[0065] Furthermore, such as Figure 5 As shown, the ship cabin environment dynamic rendering device 10 further includes: a preloading module 5; the preloading module 5 is used for: Starting from the user's current location in the compartment node and ending at the current task objective, the optimal safe path to the destination is determined based on the monitoring data stream. The preloading priority of each compartment node on the optimal safe path and its topologically adjacent compartment nodes is increased.

[0066] Furthermore, the step of determining the optimal safe path to the destination based on the monitoring data stream, taking the user's current location as the starting point and the current task objective as the ending point, and increasing the preloading priority of each cabin node on the optimal safe path and its topologically adjacent cabin nodes, includes: The system obtains the location of the task target input by the user or the task target assigned by the system as the endpoint, and combines the topological relationship of the compartment nodes and the damage control status data in the monitoring data stream to assign passage weights to the directed edges that connect the topological compartment nodes to obtain a weighted topology graph. Starting from the user's current location in the compartment node and ending at the current task objective, the optimal weighted safe path is calculated using the Dijkstra shortest path algorithm on the weighted topology graph, and an ordered sequence of compartment nodes is output as the optimal safe path. When the damage control status data is updated, causing a change in the weight of the directed edge, the optimal safe path is recalculated. The cabin nodes traversed by the optimal safety path are marked as high-priority preloaded; the cabin nodes adjacent to the cabin nodes traversed by the optimal safety path are marked as medium-priority preloaded; the remaining non-cabin nodes are marked as low-priority; the cabin node model data marked as high-priority preloaded are added to the head of the preload queue, the cabin node model data marked as medium-priority preloaded are added to the middle of the preload queue, and the cabin node model data marked as low-priority preloaded are marked as releaseable after loading.

[0067] Furthermore, the step of determining the comprehensive score of loading priority for each compartment node based on the semantic-geometric dual-drive scene map, the dynamic priority control information, and the terminal view information, and completing the loading determination and loading accuracy level allocation for the compartment nodes based on the comprehensive score of loading priority, includes: Based on the semantic-geometric dual-drive scene map and the cabin node where the user is currently located, determine the spatial relationship between the user's current camera view frustum and the boundary of the cabin's enclosed space, and calculate the visual visibility score of the cabin node based on the spatial relationship between the user's current camera view frustum and the boundary of the cabin's enclosed space. The cabin type and functional level are determined based on the semantic attributes bound to the cabin node, and the cabin semantic importance score of the cabin node is determined based on the cabin type and functional level. Based on the compartment damage control alarm signals identified by the monitoring data stream, determine the damage control urgency score of the compartment node; Based on the high-priority preloading state, medium-priority preloading state, and low-priority state of each compartment node marked by the optimal safety path, the path passing through the compartment node is determined and scored. Based on the monitoring data stream, the task scenario is determined to be either routine inspection or emergency response. Based on the task scenario, the weight values ​​of the visual visibility score, the semantic importance score of the compartment, the urgency score of damage control, and the path traversal mark score are determined. The loading priority of each compartment node is determined by the visual visibility score and its weight, the semantic importance score of the compartment and its weight, the urgency score of damage control and its weight, and the path traversal mark score and its weight. The loading priority of each compartment node is ranked in descending order based on a comprehensive score, and the loading accuracy level of each compartment node is determined according to a preset loading threshold.

[0068] Furthermore, such as Figure 5 As shown, the ship cabin environment dynamic rendering device 10 further includes: an update control module 6; the update control module 6 is used for: The monitoring data stream is continuously monitored to see if there are any preset trigger events. The preset trigger events include at least one of the following: a significant change in viewpoint event, a significant movement of personnel position event, a damage control status update event, and a path replanning event. When the camera's orientation angle changes beyond a preset angle threshold or its position changes beyond a preset length threshold, a significant change in viewing angle event is triggered. When the change in the position of personnel at the task anchor point exceeds a preset distance threshold, a significant personnel movement event is triggered. When a new damage control alarm is received or an existing alarm level changes, a damage control status update event is triggered. When the weight of a directed edge changes, triggering the recalculation of the optimal safe path and generating a new path, a path replanning event is determined to be triggered. Once any preset trigger event is detected, the dynamic priority control information is re-executed, the comprehensive score of loading priority, loading judgment and loading accuracy level allocation of each compartment node are recalculated, and a new round of decision-making cycle begins.

[0069] Furthermore, the step of performing model loading based on the loading determination and the loading accuracy level includes: The terminal requests the model data packets for the corresponding compartment nodes from the external interface of the model management platform according to the loading priority order based on the loading determination; the accuracy of each compartment node is determined according to the loading accuracy level; the model data packet contains geometric mesh, material description and texture map; The model data package is loaded using an asynchronous, non-blocking method. After each frame is loaded, the terminal reports the loading completion status to the server. The loading completion status includes the cabin ID, loading accuracy level, and loading time. The terminal will unload loaded models that are marked as low priority and are not within any view frustum, freeing up memory.

[0070] The aforementioned ship cabin environment dynamic rendering device constructs a semantic-geometric dual-drive scene map, integrates dynamic priority control information from ship monitoring data stream output, calculates a comprehensive score for cabin node loading priority based on multi-dimensional information, completes loading judgment and detail level precision allocation, and completes model rendering output on demand. This enables intelligent on-demand loading and rendering of complex ship cabin scenes on resource-constrained terminals, taking into account both routine inspection and emergency response scenarios, reducing memory and rendering overhead, and improving the system's long-term operational stability.

[0071] For a description of the features in the embodiment of the ship cabin environment dynamic rendering device, please refer to the relevant description of the embodiment of the ship cabin environment dynamic rendering method, which will not be repeated here.

[0072] Embodiments of this application also provide an electronic device, including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the steps in any of the above embodiments of the dynamic rendering method for ship cabin environment.

[0073] In one embodiment, the electronic device may be a server, and its internal structure diagram may be as follows: Figure 6 As shown, the electronic device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores dynamic rendering data of the ship's cabin environment. The network interface communicates with external terminals via a network connection. When the computer program is executed by the processor, it implements a method for dynamic rendering of the ship's cabin environment.

[0074] Embodiments of this application also provide a computer-readable storage medium storing a computer program, wherein the computer program is configured to execute the steps in any of the above embodiments of the dynamic rendering method for ship cabin environment.

[0075] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard disk, magnetic disk, or optical disk.

[0076] The embodiments of this application also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the above embodiments of the dynamic rendering method for ship cabin environment.

[0077] Embodiments of this application also provide another computer program product, including a non-volatile computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps in any of the above embodiments of the dynamic rendering method for ship cabin environment.

[0078] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0079] The foregoing has provided a detailed description of a method, apparatus, electronic device, and medium for dynamic rendering of a ship's cabin environment. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are only intended to aid in understanding the method and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of the claims of this application.

Claims

1. A method for dynamically rendering a ship's cabin environment, characterized in that, include: The ship's 3D model is subjected to semantic parsing of compartments and offline structured parsing to obtain compartment nodes carrying semantic attributes and the topological relationships of the compartment nodes. The topological relationships are then associated and mapped with the spatial octree geometric index to obtain a semantic-geometric dual-drive scene map. Acquire monitoring data streams from the ship, the monitoring data streams including at least ship damage control signals and personnel location data, control the loading priority of each compartment node based on the semantic-geometric dual-drive scenario map and the monitoring data streams, and output dynamic priority adjustment information; The loading priority comprehensive score of each compartment node is determined based on the semantic-geometric dual-drive scenario map, the dynamic priority control information, and the terminal perspective information. The loading judgment of the compartment node and the allocation of loading accuracy level are completed based on the loading priority comprehensive score. Based on the loading determination and loading accuracy level, the model is loaded and an augmented reality 2D rendered image is output.

2. The method for dynamic rendering of ship cabin environment according to claim 1, characterized in that, The process involves performing semantic parsing and offline structured parsing on the ship's 3D model to obtain cabin nodes carrying semantic attributes and their topological relationships. These topological relationships are then correlated and mapped with a spatial octree geometric index to obtain a semantic-geometric dual-drive scene atlas, including: Obtain the ship's 3D model, traverse the component tree of the ship's 3D model to perform semantic parsing, identify geometric entities labeled as compartments, bulkheads, and decks, extract the closed space boundary of each compartment consisting of walls, top plate, and bottom plate, extract the geometric bounding boxes of key equipment, and instantiate each compartment as a compartment node. Identify the connecting paths between the compartment nodes, and construct a compartment connection topology graph with compartments and key equipment as nodes and connecting paths as directed edges; the direction of the directed edges represents the passable direction between compartments, and each directed edge records the connection type, pass width and pass height; Each compartment node is bound to a multidimensional semantic attribute in the form of a key-value dictionary to obtain the compartment node with the bound semantic attribute; the multidimensional semantic attribute includes compartment type, hazard level, deck level, compartment area and volume, carrying capacity, and a list of equipment contained in the compartment; Obtain the pre-built spatial octree index, use the closed spatial boundary of each cabin node as the query range to retrieve the leaf nodes that intersect with it in the octree, and bind the set of model components associated with the intersecting leaf nodes to the corresponding cabin node. The cabin connection topology, cabin nodes with bound multi-dimensional semantic attributes, and spatial octree index are associated, mapped, and fused to output a semantic-geometric dual-drive scene map.

3. The method for dynamic rendering of ship cabin environment according to claim 1, characterized in that, The loading priority of each compartment node is controlled based on the semantic-geometric dual-drive scenario map and the monitoring data stream, and dynamic priority adjustment information is output, including: Obtain the configured task anchor personnel and obtain their real-time location from the monitoring data stream; Centered on the cabin where the personnel at the mission anchor point are located, adjacent cabins that extend outward along the directed edge are selected to define a common high-priority loading area, and the loading priority of each cabin node within the common high-priority loading area is increased. Each terminal overlays a personalized view area on the common high-priority loading area to form a terminal view area. Based on the terminal view area, the cabin nodes to be loaded are determined, and model distribution deduplication is performed to avoid multiple terminals repeatedly requesting the same model data. The results are then integrated to obtain complete dynamic priority control information.

4. The method for dynamic rendering of ship cabin environment according to claim 3, characterized in that, The method of controlling the loading priority of each compartment node based on the semantic-geometric dual-drive scenario map and the monitoring data stream, and outputting dynamic priority adjustment information, also includes: When the monitoring data stream identifies a compartment damage control alarm signal, the compartment node with the alarm signal is marked as an alarm compartment node, and the loading priority of the alarm compartment node is set to the highest level. Based on the topological relationship of the cabin nodes, the loading priority of the cabin nodes that are topologically adjacent to the alarm cabin node is marked as the second highest level; Push priority update instructions to all relevant terminals and output the updated dynamic priority adjustment information.

5. The method for dynamic rendering of ship cabin environment according to claim 1, characterized in that, Before performing model loading based on the loading determination and loading accuracy level, the method further includes: Starting from the user's current location in the compartment node and ending at the current task objective, the optimal safe path to the destination is determined based on the monitoring data stream. The preloading priority of each compartment node on the optimal safe path and its topologically adjacent compartment nodes is increased.

6. The method for dynamic rendering of ship cabin environment according to claim 5, characterized in that, The process of determining the optimal safe path to the target destination based on the monitoring data stream, starting from the user's current location in the cabin node and ending at the current task objective, and increasing the preloading priority of each cabin node on the optimal safe path and its topologically adjacent cabin nodes includes: The system obtains the location of the task target input by the user or the task target assigned by the system as the endpoint, and combines the topological relationship of the compartment nodes and the damage control status data in the monitoring data stream to assign passage weights to the directed edges that connect the topological compartment nodes to obtain a weighted topology graph. Starting from the user's current location in the compartment node and ending at the current task objective, the optimal weighted safe path is calculated using the Dijkstra shortest path algorithm on the weighted topology graph, and an ordered sequence of compartment nodes is output as the optimal safe path. When the damage control status data is updated, causing a change in the weight of the directed edge, the optimal safe path is recalculated. The cabin nodes traversed by the optimal safety path are marked as high-priority preloaded; the cabin nodes adjacent to the cabin nodes traversed by the optimal safety path are marked as medium-priority preloaded; the remaining non-cabin nodes are marked as low-priority; the cabin node model data marked as high-priority preloaded are added to the head of the preload queue, the cabin node model data marked as medium-priority preloaded are added to the middle of the preload queue, and the cabin node model data marked as low-priority preloaded are marked as releaseable after loading.

7. The method for dynamic rendering of ship cabin environment according to claim 5, characterized in that, The process of determining the comprehensive score of loading priority for each compartment node based on the semantic-geometric dual-drive scenario map, the dynamic priority control information, and the terminal view information, and then completing the loading determination and loading accuracy level allocation for the compartment nodes based on the comprehensive score of loading priority, includes: Based on the semantic-geometric dual-drive scene map and the cabin node where the user is currently located, determine the spatial relationship between the user's current camera view frustum and the boundary of the cabin's enclosed space, and calculate the visual visibility score of the cabin node based on the spatial relationship between the user's current camera view frustum and the boundary of the cabin's enclosed space. The cabin type and functional level are determined based on the semantic attributes bound to the cabin node, and the cabin semantic importance score of the cabin node is determined based on the cabin type and functional level. Based on the compartment damage control alarm signals identified by the monitoring data stream, determine the damage control urgency score of the compartment node; Based on the high-priority preloading state, medium-priority preloading state, and low-priority state of each compartment node marked by the optimal safety path, the path passing through the compartment node is determined and scored. Based on the monitoring data stream, the task scenario is determined to be either routine inspection or emergency response. Based on the task scenario, the weight values ​​of the visual visibility score, the semantic importance score of the compartment, the urgency score of damage control, and the path traversal mark score are determined. The loading priority of each compartment node is determined by the visual visibility score and its weight, the semantic importance score of the compartment and its weight, the urgency score of damage control and its weight, and the path traversal mark score and its weight. The loading priority of each compartment node is ranked in descending order based on a comprehensive score, and the loading accuracy level of each compartment node is determined according to a preset loading threshold.

8. The method for dynamic rendering of ship cabin environment according to claim 1, characterized in that, The method further includes: The monitoring data stream is continuously monitored to see if there are any preset trigger events. The preset trigger events include at least one of the following: a significant change in viewpoint event, a significant movement of personnel position event, a damage control status update event, and a path replanning event. When the camera's orientation angle changes beyond a preset angle threshold or its position changes beyond a preset length threshold, a significant change in viewing angle event is triggered. When the change in the position of personnel at the task anchor point exceeds a preset distance threshold, a significant personnel movement event is triggered. When a new damage control alarm is received or an existing alarm level changes, a damage control status update event is triggered. When the weight of a directed edge changes, triggering the recalculation of the optimal safe path and generating a new path, a path replanning event is determined to be triggered. Once any preset trigger event is detected, the dynamic priority control information is re-executed, the comprehensive score of loading priority, loading judgment and loading accuracy level allocation of each compartment node are recalculated, and a new round of decision-making cycle begins.

9. The method for dynamic rendering of ship cabin environment according to claim 1, characterized in that, The step of performing model loading based on the loading determination and loading accuracy level includes: The terminal requests the model data packets for the corresponding compartment nodes from the external interface of the model management platform according to the loading priority order based on the loading determination; the accuracy of each compartment node is determined according to the loading accuracy level; the model data packet contains geometric mesh, material description and texture map; The model data package is loaded using an asynchronous, non-blocking method. After each frame is loaded, the terminal reports the loading completion status to the server. The loading completion status includes the cabin ID, loading accuracy level, and loading time. The terminal will unload loaded models that are marked as low priority and are not within any view frustum, freeing up memory.

10. A dynamic rendering device for ship cabin environment, characterized in that, The device includes: The atlas construction module is used to perform cabin semantic parsing and offline structured parsing processing on the ship's 3D model to obtain cabin nodes carrying semantic attributes and the topological relationships of cabin nodes. The topological relationships are associated and mapped with the spatial octree geometric index to obtain a semantic-geometric dual-drive scene atlas. The situation priority processing module is used to acquire the monitoring data stream in the ship. The monitoring data stream includes at least ship damage control signals and personnel location data. Based on the semantic-geometric dual-drive scenario map and the monitoring data stream, the module controls the loading priority of each compartment node and outputs dynamic priority adjustment information. The weighted decision module is used to determine the comprehensive score of the loading priority of each compartment node based on the semantic-geometric dual-drive scenario map, the dynamic priority control information, and the terminal perspective information, and to complete the loading judgment and loading accuracy level allocation of the compartment node based on the comprehensive score of the loading priority. The terminal rendering module is used to load the model and output an augmented reality 2D rendered image based on the loading determination and loading accuracy level.

11. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of the dynamic rendering method for ship cabin environment as described in any one of claims 1 to 9 when executing the computer program.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, it implements the steps of the dynamic rendering method for the ship cabin environment as described in any one of claims 1 to 9.

Images