Visual treatment system and treatment method for underwater construction of bagged sand

Abstract

The invention relates to the field of underwater construction and computer visualization, and discloses a bagged sand underwater construction visualization processing system and method, and the system comprises a three-dimensional scene model storage which is used for storing a three-dimensional scene model which can be dynamically updated, and the three-dimensional scene model represents the physical boundary of an underwater construction area; the data acquisition module is configured to acquire motion sensor data of to-be-placed bagged sand in real time and construct a dynamic state vector containing the motion state of the bagged sand based on the motion sensor data; and the parameter identification module is connected with the data acquisition module. By arranging a parameter identification module, before the bagged sand is released, key physical parameters such as an equivalent hydrodynamic damping coefficient and a local environment water flow velocity are identified on line by utilizing the moving process of the bagged sand in water, so that a dynamic model for track prediction can dynamically adapt to a real and variable underwater environment; and errors caused by dependence on fixed or inaccurate empirical parameters are avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of underwater construction and computer visualization, in particular to a visualization processing system and method for underwater construction of bagged sand. Background Art

[0002] In water conservancy projects, marine engineering projects, and port and waterway construction, operations such as dam reinforcement, underwater pipeline protection, and caisson foundation laying often require the precise placement and deployment of a large number of components underwater. Bag sand is widely used due to its cost-effectiveness, adaptability, and ease of construction. Traditional underwater installation of bag sand relies primarily on lifting equipment (such as cranes or grab buckets) on surface construction platforms (such as ships or barges). The surface platform is positioned using methods such as the Global Positioning System, and the bag sand is then lowered to the desired location for release.

[0003] However, existing technologies suffer from inherent technical flaws. Underwater environments, especially turbid waters, have extremely low visibility, making the entire installation process a "blind operation." Operators on the surface cannot directly observe the actual movement of the bagged sand underwater, its trajectory, or its interaction with existing structures or terrain.

[0004] More importantly, after the bagged sand is released from the water's surface, its trajectory through the water isn't a simple vertical drop; instead, it's subject to complex and dynamically changing hydrodynamic forces. For one thing, the bagged sand's irregular shape and tumbling behavior during its fall make its hydrodynamic drag coefficient difficult to precisely predetermine. Furthermore, localized currents, whose strength and direction vary with depth and time, are prevalent in the construction area. These invisible undercurrents can significantly shift the bagged sand. The combined effects of these uncertainties often result in significant, unpredictable deviations between the bagged sand's final landing point and the theoretical target point estimated based on the water surface. Consequently, operators are largely forced to rely on personal experience through repeated trial and error, which not only leads to low construction efficiency and material waste, but also makes it difficult for the resulting underwater structure to meet precise design requirements.

[0005] In addition, even if the existing auxiliary installation methods utilize the initial underwater topographic survey data, they generally ignore the dynamic changes in the underwater environment during the construction process. Each successful placement of bagged sand will become a new physical entity in the underwater scene, thereby changing the local topography. The placement and stability of subsequent bagged sand will be directly affected by these previously placed components. Existing technical solutions generally lack an effective feedback and update mechanism, and are unable to dynamically and cumulatively reflect the completed construction results in the environmental model on which they rely. As a result, their understanding of the environment is gradually disconnected from the actual construction status, making subsequent installation guidance information increasingly distorted and unable to meet the needs of continuous, large-scale, and high-precision construction. Summary of the Invention

[0006] In response to the shortcomings of the existing technology, the present invention provides a visualization processing system and method for underwater construction of bagged sand, which solves the problem that in blind underwater operation environments, due to the uncertainty of environmental factors such as hydrodynamics and dynamic changes in construction scenes, it is impossible to accurately predict the landing point and achieve precise placement of components such as bagged sand.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: a bagged sand underwater construction visualization processing system, comprising: A three-dimensional scene model memory, used to store a dynamically updateable three-dimensional scene model, wherein the three-dimensional scene model represents the physical boundary of the underwater construction area; a data acquisition module configured to acquire in real time motion sensor data of the bagged sand to be placed, and construct a dynamic state vector containing the motion state of the bagged sand based on the motion sensor data; a parameter identification module, connected to the data acquisition module, and configured to identify physical parameters characterizing underwater environmental characteristics online based on the dynamic state vector before the bagged sand is released; The trajectory calculation module is connected to the parameter identification module, the data acquisition module and the three-dimensional scene model memory, and is configured as follows: Calculating a predicted trajectory of the bagged sand after self-release based on the physical parameters and the dynamic state vector; and Determining a predicted landing point of the bagged sand by detecting a collision between the predicted trajectory and the three-dimensional scene model stored in the three-dimensional scene model memory; a visualization presentation module, connected to the trajectory calculation module, configured to display the predicted trajectory and the predicted landing point in an augmented reality display in a visualization interface; The scene update module is connected to the three-dimensional scene model memory and is configured to: after the bagged sand is placed, update the three-dimensional scene model stored in the three-dimensional scene model memory according to the detection data of the placed bagged sand obtained by the detection equipment.

[0008] Preferably, the parameter identification module is specifically configured as follows: An extended Kalman filter is adopted, the physical parameters are used as state variables to be estimated, and the real-time motion information in the dynamic state vector is used to perform online optimal estimation of the physical parameters.

[0009] Preferably, the physical parameters include: Equivalent hydrodynamic damping coefficient or local ambient water velocity vector.

[0010] Preferably, the trajectory solving module is specifically configured as follows: A dynamic equation of the bagged sand in water is established, and a numerical integration method is used to solve the equation in combination with the physical parameters to obtain the predicted trajectory.

[0011] Preferably, the predicted landing point is expressed as a probabilistic placement area on the surface of the three-dimensional scene model.

[0012] Preferably, the probabilistic placement area is generated based on a state covariance matrix output by an extended Kalman filter.

[0013] Preferably, the visualization presentation module is specifically configured as follows: The predicted trajectory is displayed in the form of a virtual trajectory line, and the predicted landing point or its corresponding probabilistic placement area is displayed in the form of a highlighted area or a heat map in a superimposed manner in the visualization interface.

[0014] Preferably, the scene update module is specifically configured as follows: After the bagged sand is confirmed to be placed, instructing the detection device to scan the newly placed bagged sand to obtain a three-dimensional solid model thereof; and Through a model fusion operation, the three-dimensional entity model is incorporated into the three-dimensional scene model to form an updated three-dimensional scene model.

[0015] Preferably, the dynamic state vector includes at least one information selected from the following group: The position, speed, acceleration, attitude and angular velocity of the bagged sand.

[0016] The present invention also provides a method for visualizing underwater construction of bagged sand, comprising the following steps: Step 1: Establish a dynamically updateable 3D scene model to represent the physical boundaries of the underwater construction area; Step 2: acquiring motion sensor data of the bagged sand to be placed in real time, and constructing a dynamic state vector containing the motion state of the bagged sand; Step 3: Based on the dynamic state vector, before the bagged sand is released, online identification of physical parameters characterizing underwater environmental characteristics; Step 4: Based on the physical parameters and the dynamic state vector, a predicted trajectory of the bagged sand after self-release is calculated, and a predicted landing point of the bagged sand is determined by detecting a collision between the predicted trajectory and the three-dimensional scene model; Step 5: Display the predicted trajectory and the predicted landing point in an augmented reality display in a visualization interface; Step 6: After the bagged sand is placed, the three-dimensional scene model is updated according to the detection data of the bagged sand obtained by the detection equipment for subsequent installation operations.

[0017] The present invention provides a visual processing system and method for underwater construction of bagged sand, which has the following beneficial effects: 1. By setting up a parameter identification module, the present invention uses the movement process of the bagged sand in the water to online identify key physical parameters such as the equivalent hydrodynamic damping coefficient and the local environmental water velocity before the bagged sand is released. These real-time identified parameters are applied to the subsequent trajectory solution, so that the dynamic model of trajectory prediction can dynamically adapt to the real and changeable underwater environment, avoiding the errors caused by relying on fixed or inaccurate empirical parameters, and significantly improving the accuracy and reliability of the predicted trajectory and predicted landing point.

[0018] 2. The present invention sets up a visualization presentation module to intuitively present the high-precision predicted trajectory and predicted landing point in the form of augmented reality such as virtual trajectory lines and highlighted areas in a visualization interface superimposed with a three-dimensional scene model, thereby transforming the invisible blind operation underwater into a clear closed-loop control task based on visual feedback, enabling the operator to perform precise pre-alignment, effectively improving the positioning accuracy and operating efficiency of a single placement.

[0019] 3. This invention incorporates a scene update module. After each confirmed placement of sand bags, this module uses detection data from the detection equipment to dynamically and cumulatively update the 3D scene model. This creates a digital twin that evolves synchronously with the real underwater construction environment. This mechanism ensures that subsequent trajectory predictions fully account for the physical occlusion and support effects of placed components, enhancing the system's adaptability to complex and dynamic environments and ensuring long-term prediction effectiveness during continuous construction. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is the main framework diagram of the present invention; Figure 2 This is a schematic diagram of the motion sensor process of the present invention; Figure 3This is a schematic diagram of the inertial measurement unit process of the present invention; Figure 4 This is a schematic diagram of the initialization unit flow of the present invention; Figure 5 Schematic diagram of the parameter identification module flow of the present invention. DETAILED DESCRIPTION

[0021] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0022] Please see the attached Figure 1 -Attached Figure 5 An embodiment of the present invention provides a visualization processing system for underwater construction of bagged sand. By setting a parameter identification module, before the bagged sand is released, its movement process in the water is used to online identify key physical parameters such as the equivalent hydrodynamic damping coefficient and the local environmental water flow velocity, and these real-time identified parameters are applied to the subsequent trajectory solution, so that the dynamic model of trajectory prediction can dynamically adapt to the real and changeable underwater environment, avoiding the errors caused by relying on fixed or inaccurate empirical parameters, and significantly improving the accuracy and reliability of the predicted trajectory and predicted landing point.

[0023] Bag sand underwater construction visualization processing system, including: A three-dimensional scene model memory, used to store a dynamically updateable three-dimensional scene model, which represents the physical boundary of the underwater construction area; a data acquisition module configured to acquire in real time motion sensor data of the bagged sand to be placed, and construct a dynamic state vector containing the motion state of the bagged sand based on the motion sensor data; A parameter identification module, connected to the data acquisition module, is configured to identify physical parameters representing underwater environmental characteristics online based on a dynamic state vector before the bagged sand is released; The trajectory calculation module connects the parameter identification module, the data acquisition module and the 3D scene model memory, and is configured as follows: Calculate the predicted trajectory of the bagged sand after self-release based on physical parameters and dynamic state vectors; and Determining the predicted landing point of the bagged sand by detecting the collision between the predicted trajectory and the three-dimensional scene model stored in the three-dimensional scene model memory; A visualization presentation module, connected to the trajectory solving module, configured to display the predicted trajectory and predicted landing point in an augmented reality in a visualization interface; The scene update module is connected to the three-dimensional scene model memory and is configured to: after the bagged sand is placed, update the three-dimensional scene model stored in the three-dimensional scene model memory according to the detection data of the placed bagged sand obtained by the detection equipment.

[0024] First, before the entire underwater installation operation starts, the system needs to initialize the 3D scene model memory to establish an initial 3D scene model that represents the original physical boundary.

[0025] Preferably, the initialization process uses external detection equipment, such as a multi-beam bathymetric system or a three-dimensional scanning sonar, to conduct comprehensive topographic mapping of the planned construction waters, thereby obtaining initial three-dimensional point cloud data of the riverbed or seabed with high spatial resolution.

[0026] The acquired raw point cloud data will go through a series of pre-processing operations, including but not limited to coordinate transformation, denoising filtering and outlier removal, and then be constructed into a geometric model that is easy to process in computer graphics, preferably a triangular mesh model or a voxel grid model. This model is the initial 3D scene model, denoted as , which is loaded and stored in the 3D scene model memory as the benchmark zero state for all subsequent simulation calculations and visualization presentations.

[0027] In addition, to achieve guided operation, the CAD design model containing the engineering design outline or theoretical placement target points is also loaded and accurately spatially aligned with the initial 3D scene model in a unified global coordinate system.

[0028] More importantly, the three-dimensional scene model stored in the three-dimensional scene model memory is not static, but dynamically evolves with the construction progress through interaction with the scene update module. This constitutes the technical basis for the present invention to solve the problem of accumulated physical influences between components.

[0029] This dynamic update process is triggered when the placement of a bag of sand is confirmed. The scene update module instructs the detection equipment, preferably a forward-looking high-frequency imaging sonar mounted on an underwater robot or grab, to perform a close-range, high-resolution scan of the bag of sand that has just been placed on the bottom.

[0030] The purpose of this scanning is to obtain the final, true three-dimensional point cloud data of the newly placed component, including its actual deformation and posture caused by water impact and interaction with the environment.

[0031] Subsequently, the system processes the point cloud data obtained from the scan and generates an independent, closed, and accurate three-dimensional solid model that describes the geometric shape of the single bag of sand, which is recorded as ,in The serial number of the current installation.

[0032] Next, the system performs a model fusion operation to transform the newly generated 3D solid model , and the three-dimensional scene model currently stored in the three-dimensional scene model memory Preferably, the model fusion operation is a Boolean union operation of three-dimensional models, and its mathematical expression can be summarized as: Through this operation, the entity model of the new component is seamlessly and permanently integrated into the scene model, forming an updated version that better reflects the actual physical boundaries of the "after completion" 3D scene model The updated model will overwrite the old version and become the latest data in the storage to serve the next job. This historical state accumulation mechanism makes the 3D scene model storage play a role far more than a simple data warehouse.

[0033] It becomes a self-evolving digital twin world. When performing subsequent installation operations, the trajectory solution module will retrieve the latest 3D scene model from the 3D scene model memory in real time. , serving as the physical boundary for collision detection. This ensures that the predicted trajectory of subsequent bagged sand correctly interacts with all previously placed components, rather than "penetrating" them, significantly improving the accuracy of trajectory prediction during continuous construction. Simultaneously, the visualization module retrieves the latest model from storage as the background scene for augmented reality overlay rendering, allowing the operator to view a 3D environment fully synchronized with the actual underwater construction progress.

[0034] Specifically, the data acquisition module operates throughout the entire process from the time the bagged sand is grabbed by the bucket, moves through the water, and is finally released. This process provides a crucial data foundation for the subsequent online parameter identification of the present invention.

[0035] In order to achieve comprehensive perception of the movement state of the bagged sand, the data acquisition module is configured to obtain data from a set of preferred multi-source heterogeneous sensor systems.

[0036] Preferably, the sensor system includes: a differential global positioning system (RTK / DGPS) installed on a surface work mother ship or platform, which is used to provide a high-precision global geographic location reference; an ultra-short baseline (USBL) acoustic positioning beacon installed on the underwater grab or a nearby location, which is used to obtain the three-dimensional position of the grab underwater in real time; and an inertial measurement unit (IMU) integrated on the grab, which has a built-in accelerometer and gyroscope, which is used to directly measure the linear acceleration and angular velocity of the grab.

[0037] At the data acquisition level, the data collection module continuously performs the following operations: It obtains the position information of bagged sand from the ultra-short baseline system ; It obtains three-axis linear acceleration information from the inertial measurement unit , three-axis angular velocity information , and attitude information usually expressed in the form of quaternions .

[0038] After acquiring raw sensor data, the core task of the data acquisition module is to construct and update the dynamic state vector in real time. This vector is a mathematical representation of the instantaneous physical state of the bagged sand. It not only includes directly measured kinematic quantities but also reserves data structure space for unknown parameters to be identified.

[0039] The complete form of the dynamic state vector can be expressed as: ; In this vector, the origin and function of each component are as follows: Position vector and attitude quaternion , angular velocity vector and acceleration vector , which is directly derived from the real-time measurement values of the above sensors.

[0040] Velocity vector , preferably by the data acquisition module through the continuous position vector It is obtained by performing time difference or derivative operation, which reflects the instantaneous movement rate and direction of the bagged sand.

[0041] Physical parameters, i.e. equivalent hydrodynamic damping coefficient and the local ambient water velocity vector , their values are unknown at the initial stage of this vector construction. They are included in the vector structure as placeholders, and their true values will be estimated and filled online by the subsequent parameter identification module. This structural design ensures the uniformity and integrity of the data flow.

[0042] Effective mass , represents the mass corresponding to the apparent weight of the bagged sand in water. It is usually a constant preset according to the specifications of the bagged sand and is included in the vector for subsequent dynamic calculations.

[0043] Preferably, given the potential differences in data update frequencies between different sensors, the data acquisition module also includes a data synchronization unit. This unit is responsible for time-aligning and synchronizing the data streams from different sensors to ensure that each generated dynamic state vector accurately reflects the physical state at the same moment, thus ensuring the accuracy of subsequent algorithms.

[0044] Ultimately, the dynamic state vector, constructed and output in real time by the data acquisition module, serves as the most direct and core data input and is passed to the parameter identification module and the trajectory solution module. The parameter identification module uses the time series of this vector to identify unknown physical parameters, while the trajectory solution module uses the vector at the moment of bagged sand release as the initial condition for its physical simulation.

[0045] In this embodiment, the parameter identification module is the core technical unit that enables the adaptability and high-precision prediction capabilities of the bagged sand underwater construction visualization processing system. Its fundamental task is to address the technical challenges of unknown and dynamically changing key physical parameters in underwater environments. Through online identification, it provides accurate model input that conforms to current real-world conditions for subsequent trajectory calculations.

[0046] The key perturbation affecting the trajectory of bagged sand underwater is primarily fluid dynamics. The magnitude and direction of this force depend not only on the motion of the bagged sand itself but also, more importantly, on two difficult-to-predict physical parameters: the equivalent hydrodynamic damping coefficient, which characterizes the interaction between the bagged sand and the water; and the local ambient water velocity vector, which characterizes the water's own motion. These parameters are highly nonlinear and time-varying due to the bagged sand's irregular shape and variable falling posture, as well as undercurrents and eddies in the water. Using fixed empirical values for trajectory prediction inevitably introduces significant errors.

[0047] To address this issue, the parameter identification module begins its work before the bagged sand is officially released, while the grab bucket is moving through the water. It connects to the data acquisition module, continuously acquiring the time series of the dynamic state vectors generated by the latter. Based on this information, it uses an advanced filter estimation algorithm for online parameter identification.

[0048] Preferably, in this embodiment, the parameter identification module uses the Extended Kalman Filter (EKF) as the core algorithm to implement this function. This choice is because the EKF is particularly suitable for processing the state estimation problem of nonlinear systems. Its implementation principle is as follows: First, the module takes the physical parameters to be identified, namely the equivalent hydrodynamic damping coefficient and the local ambient water velocity vector, as the state variables to be estimated, and together with the core kinematic state (position and velocity) of the bagged sand, forms an augmented EKF state vector, which can be in the form of: By embedding the physical parameters in the state vector, the filter can use the dynamic response of the system to reversely infer and estimate the values of these parameters during the iteration process.

[0049] Next, the module establishes a state transition equation, or process model, that describes how the state vector evolves over time. The function here is a discrete-time expression based on Newton's laws of motion. For the physical parameters and, a random walk model is preferably used to describe their time-varying characteristics. This assumes they are approximately constant over short periods of time, but allows them to vary slowly over time, consistent with physical reality. Finally, the module establishes a measurement equation that relates the state vector to the actual sensor measurements. In this invention, the measurements are primarily the real-time positions of the bagged sand provided by the ultrashort baseline system, so the function primarily extracts the position component from the state vector. With this model in place, the parameter identification module operates in a continuous prediction-update cycle: In the prediction step, the module uses the state transition equation to predict the a priori estimate of the state vector and its covariance at the current moment based on the optimal estimate at the previous moment. In the update step, when the module receives a new sensor measurement from the data acquisition module, it calculates the residual between that measurement and the value predicted based on the a priori estimate. Based on this residual and combined with the uncertainty of the system (represented by the covariance matrix), the module calculates the Kalman gain and uses it to correct the prior estimate to obtain the optimal estimate of the posterior state at the current moment that incorporates the latest measurement information.

[0050] By repeating this prediction-update cycle as the bagged sand moves before being released, the physical parameter components (and) in the state vector will quickly converge from unknown initial values (which can be set to a reasonable guess) to stable values that can best explain the observed movement behavior of the bagged sand.

[0051] Finally, when trajectory prediction is ready, the parameter identification module provides its converged, optimally estimated physical parameter values to the trajectory solution module. Furthermore, the module outputs a final state covariance matrix, which quantifies the uncertainty in the estimated values of all state variables (including the identified parameters). This covariance matrix is also used by subsequent modules to generate probabilistic placement zones, providing operators with an intuitive basis for assessing the reliability of the prediction.

[0052] The trajectory calculation module is typically triggered by an operator command or automatically activated when the system determines that the bagged sand has entered an area suitable for release. Once activated, the module immediately obtains all the input information necessary for accurate calculation from other associated modules.

[0053] Specifically, it obtains the physical parameters that have just been identified online and characterize the current underwater environment characteristics from the parameter identification module, namely the equivalent hydrodynamic damping coefficient and the local ambient water velocity vector When the sand bag is released, it obtains the dynamic state vector of the sand bag at the moment of release from the data acquisition module, and extracts the position vector as the initial condition of the simulation. and velocity vector In addition, it retrieves the latest version of the 3D scene model from the 3D scene model memory, which contains all previously placed components. , serving as the geometric boundary for subsequent collision detection. After obtaining all the above input information, the trajectory solution module first establishes a dynamic equation describing the free fall of the bagged sand in the water after it leaves the grab bucket. This equation is based on Newton's second law and can be expressed in vector form as: In this equation, is the mass of bagged sand, is its acceleration vector.

[0054] is the net weight of the bagged sand in water, i.e. the vector sum of its own weight and the buoyancy force. This term can usually be regarded as a constant driving force.

[0055] The core, nonlinear hydrodynamic resistance term is the key to the prediction accuracy. In this embodiment, the calculation of this force fully utilizes the online identification results of the present invention, and its expression is preferably: ; in, is the density of water, and A is the equivalent cross-sectional area of the bagged sand. It is crucial that the relative velocity vector , which takes into account the local ambient water velocity vector provided by the parameter identification module In addition, the equivalent hydrodynamic damping coefficient in the formula is The latest value identified online by the parameter identification module is also directly used. This approach enables the dynamic model to dynamically adapt to the real environment, rather than relying on fixed, potentially inaccurate empirical parameters. Since the above dynamic equations are nonlinear ordinary differential equations, the trajectory solution module uses a high-precision numerical integration method to solve them. Preferably, the fourth-order Runge-Kutta method (RK4) is used, with the initial conditions ( ) as the starting point, in a small time step By forward iteration, the position and velocity of the bagged sand at each moment in the future are calculated step by step. All these consecutive position points are connected in three-dimensional space to form the predicted trajectory of the bagged sand.

[0056] At each time step of numerical integration, the trajectory solver module must perform a collision detection. The module compares the latest calculated position of the bagged sand with its geometric envelope and the 3D scene model retrieved from the 3D scene model memory. The intersection operation is performed. This process continues until the first collision is detected. This first collision point is determined by the trajectory calculation module as the predicted landing point of the bagged sand.

[0057] To provide operators with quantitative information on the reliability of the prediction results, the trajectory calculation module in this embodiment can further generate a probabilistic placement zone (PPZ). This is achieved by utilizing the final state covariance matrix output by the extended Kalman filter in the parameter identification module. This matrix represents the uncertainty in the estimates of the physical parameters and initial states. Through error propagation theory or a small number of Monte Carlo simulations (i.e., performing multiple random samplings of the input parameters within their uncertainty range and repeating the trajectory calculation), the input uncertainty can be mapped into a two-dimensional probability distribution area on the surface of the three-dimensional scene model for the predicted landing point. This area intuitively illustrates the most likely landing range of the bagged sand.

[0058] Specifically, the visualization module runs under the support of a 3D graphics rendering engine. The module runs in real time and continuously. As an information fusion and presentation terminal, it receives data streams from multiple upstream modules and unifies them in a visualization interface. First, it retrieves the latest version of the 3D scene model from the 3D scene model memory. The underwater construction area is then rendered as a complete and accurate context for the operator.

[0059] Secondly, it connects to the data acquisition module to obtain the current position and posture of the bagged sand to be placed in real time, and renders a corresponding virtual model in the 3D scene, allowing the operator to constantly understand the actual state of the bagged sand. In addition, pre-set project targets, such as the theoretical placement position or design outline, are also superimposed and rendered in the scene as reference elements.

[0060] On this basis, the core function of the visualization module is to display the predicted information in augmented reality. It receives the core output of the trajectory solution module, namely the predicted trajectory and predicted landing point, and presents them using a specific visual encoding strategy.

[0061] As for the predicted trajectory, in this embodiment, it is preferably rendered as a clear virtual trajectory line starting from the current virtual model position of the bagged sand and extending to the predicted landing point.

[0062] The trajectory can be given unique visual attributes, such as a semi-transparent material to avoid obscuring important background information, or a gradient color to indicate speed or time at different points along the trajectory. This dynamically updated trajectory acts as a "ghost guide," intuitively revealing to the operator the most likely path the bag of sand would follow if released at that moment.

[0063] The visualization method for the predicted landing point depends on its specific form.

[0064] If the predicted landing point is a certain three-dimensional point, the module can render a striking mark at the point, such as a three-dimensional cross cursor or a flashing indicator.

[0065] More preferably, when the predicted landing point is represented as a probabilistic placement zone (PPZ), the visualization module renders it as a highlighted area or heat map on the surface of the 3D scene model. The extent of this area intuitively represents the uncertainty of the prediction, while the color or brightness distribution of the heat map further indicates the high and low probability density—for example, the center area is warmer or brighter, indicating the highest landing probability.

[0066] By superimposing and displaying the above-mentioned real state information, prediction information and target information on the same screen, the present invention transforms the originally complex underwater blind operation that relies on operator experience into a clear closed-loop control task based on visual feedback.

[0067] The operator's task is no longer to guess, but to operate the lifting equipment, observe the changes in the virtual trajectory line and highlighted area in real time, and use it as a guide to continuously adjust the position of the bagged sand before release until the rendered predicted landing point or the center of the probabilistic placement area is precisely aligned with the preset engineering target placement point on the screen.

[0068] The scenario update module operates not continuously but on an event-triggered basis. Specifically, once a bag of sand has been placed and confirmed to be stable at the bottom by the operator or the automated system, the module is activated to perform a cumulative scenario update.

[0069] After activation, the scene update module first executes the detection data acquisition process after placement.

[0070] To this end, the module issues instructions to a specific detection device. Preferably, the detection device is a forward-looking high-frequency imaging sonar or a 3D scanning sonar installed on an underwater robot or a lifting grab, because it can provide high-resolution 3D point cloud data at a relatively close distance.

[0071] The goal of this scan is to accurately capture the final three-dimensional shape of the newly placed sand bags, including their actual posture resulting from impact with the riverbed and other structures, as well as deformation caused by water flow and gravity. This step aims to obtain real-world data "as built," rather than relying on predictive models.

[0072] After obtaining the original point cloud data containing the information of the newly placed bagged sand, the scene update module then executes the individual model generation process.

[0073] The process includes a series of processing of the original point cloud data, such as filtering, denoising and surface reconstruction, and finally generates an independent, closed, and accurate three-dimensional solid model that describes the geometric shape of the single newly placed bag of sand. The model generated by the installation operation is recorded as The significance of this step is to separate the new information introduced by this installation operation from the complex environment and encapsulate it into an independent digital object.

[0074] Next, the scene update module performs its most core model fusion operation. This operation connects to the 3D scene model memory and adds the newly generated 3D solid model to the 3D scene model memory. , and the version currently stored in the memory is 3D scene model Merge. In 3D computer graphics, this model fusion operation is preferably implemented as a Boolean union operation of the 3D models. Its mathematical expression can be summarized as: ; The physical meaning of this formula is that the entity model of the new component is seamlessly and incrementally incorporated into the existing scene geometry, forming a version number of The updated 3D scene model The updated model fully incorporates the All physical components after the installation is completed. Finally, the scene update module will update the 3D scene model Write back to the 3D scene model memory to overwrite the old version.

[0075] Through this cycle of “operation-detection-fusion-update”, the present invention builds a digital twin environment that can evolve dynamically. Its key role is: in the next (i.e. During the installation operation, the trajectory solver module will perform collision detection based on this data which contains the actual historical installation results. This ensures that subsequent predictions correctly account for the physical occlusion and support effects of placed components. Furthermore, the visualization module renders based on this updated model, providing operators with a visual environment that is fully synchronized with the actual underwater construction progress. This ensures that the entire system's prediction and guidance capabilities remain accurate as construction progresses.

[0076] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. Visual processing system for underwater construction of bagged sand, characterized by: include: A three-dimensional scene model memory, used to store a dynamically updateable three-dimensional scene model, wherein the three-dimensional scene model represents the physical boundary of the underwater construction area; a data acquisition module configured to acquire in real time motion sensor data of the bagged sand to be placed, and construct a dynamic state vector containing the motion state of the bagged sand based on the motion sensor data; a parameter identification module, connected to the data acquisition module, and configured to identify physical parameters characterizing underwater environmental characteristics online based on the dynamic state vector before the bagged sand is released; The trajectory calculation module is connected to the parameter identification module, the data acquisition module and the three-dimensional scene model memory, and is configured as follows: Calculating a predicted trajectory of the bagged sand after self-release based on the physical parameters and the dynamic state vector; and Determining a predicted landing point of the bagged sand by detecting a collision between the predicted trajectory and the three-dimensional scene model stored in the three-dimensional scene model memory; a visualization presentation module, connected to the trajectory calculation module, configured to display the predicted trajectory and the predicted landing point in an augmented reality display in a visualization interface; The scene update module is connected to the three-dimensional scene model memory and is configured to: after the bagged sand is placed, update the three-dimensional scene model stored in the three-dimensional scene model memory according to the detection data of the placed bagged sand obtained by the detection equipment.

2. The bagged sand underwater construction visualization processing system according to claim 1 is characterized in that: The parameter identification module is specifically configured as follows: An extended Kalman filter is adopted, the physical parameters are used as state variables to be estimated, and the real-time motion information in the dynamic state vector is used to perform online optimal estimation of the physical parameters.

3. The bagged sand underwater construction visualization processing system according to claim 1 is characterized in that: The physical parameters include: Equivalent hydrodynamic damping coefficient or local ambient water velocity vector.

4. The bagged sand underwater construction visualization processing system according to claim 1 is characterized in that: The trajectory solution module is specifically configured as follows: A dynamic equation of the bagged sand in water is established, and a numerical integration method is used to solve the equation in combination with the physical parameters to obtain the predicted trajectory.

5. The bagged sand underwater construction visualization processing system according to claim 1 is characterized in that: The predicted landing point is expressed as a probabilistic placement area on the surface of the three-dimensional scene model.

6. The bagged sand underwater construction visualization processing system according to claim 5 is characterized in that: The probabilistic placement region is generated based on a state covariance matrix output by an extended Kalman filter.

7. The bagged sand underwater construction visualization processing system according to claim 1 is characterized in that: The visualization presentation module is specifically configured as follows: The predicted trajectory is displayed in the form of a virtual trajectory line, and the predicted landing point or its corresponding probabilistic placement area is displayed in the form of a highlighted area or a heat map in a superimposed manner in the visualization interface.

8. The bagged sand underwater construction visualization processing system according to claim 1 is characterized in that: The scene update module is specifically configured as follows: After the bagged sand is confirmed to be placed, instructing the detection device to scan the newly placed bagged sand to obtain a three-dimensional solid model thereof; and Through a model fusion operation, the three-dimensional entity model is incorporated into the three-dimensional scene model to form an updated three-dimensional scene model.

9. The bagged sand underwater construction visualization processing system according to claim 1, characterized in that: The dynamic state vector includes at least one information selected from the following group: The position, speed, acceleration, attitude and angular velocity of the bagged sand.

10. A method for visualizing the processing of bagged sand underwater construction, according to the visualizing the processing system for bagged sand underwater construction according to any one of claims 1 to 9, characterized in that: The following steps are involved: Step 1: Establish a dynamically updateable 3D scene model to represent the physical boundaries of the underwater construction area; Step 2: acquiring motion sensor data of the bagged sand to be placed in real time, and constructing a dynamic state vector containing the motion state of the bagged sand; Step 3: Based on the dynamic state vector, before the bagged sand is released, online identification of physical parameters characterizing underwater environmental characteristics; Step 4: Based on the physical parameters and the dynamic state vector, a predicted trajectory of the bagged sand after self-release is calculated, and a predicted landing point of the bagged sand is determined by detecting a collision between the predicted trajectory and the three-dimensional scene model; Step 5: Display the predicted trajectory and the predicted landing point in an augmented reality display in a visualization interface; Step 6: After the bagged sand is placed, the three-dimensional scene model is updated according to the detection data of the bagged sand obtained by the detection equipment for subsequent installation operations.

Images