A system and method for three-dimensional data acquisition management

Abstract

The application relates to the technical field of data collection and discloses a three-dimensional data collection management system and method, which comprises the following steps: S1, a three-dimensional space model of a region to be collected is acquired, and the space is discretized into a plurality of sampling units to form a voxel grid; S2, an occluded region in the three-dimensional space model is identified, and an effective three-dimensional collection space is constructed; S3, initial layout points are generated according to device field-of-view parameters; S4, collection parameters of the device and positions of the sampling units are used as bases; S5, a joint optimization objective function is constructed; and S6, under the constraint condition, an optimal device layout point and a path control function are solved by an optimization method. In the application, a three-dimensional space collection model is constructed, a joint optimization objective function is used, and the dynamic optimization of device layout points and paths, energy consumption and occluded region punishment and other factors are considered, so that an optimal device layout scheme and a path control strategy are obtained, and the collection efficiency and the rationality of path planning are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data acquisition, in particular to a three-dimensional data acquisition management system and method. BACKGROUND

[0002] With the rapid development of three-dimensional data acquisition technology, in many application fields such as environmental monitoring, city modeling, robot navigation, etc., how to efficiently and accurately acquire and process three-dimensional spatial data has become a research hotspot. However, the traditional three-dimensional data acquisition method usually relies on fixed layout points and simple path planning strategy, lacking dynamic optimization of acquisition area and device configuration. This method has several problems in practical application, especially as the complexity of the acquisition area increases, the problems of device energy consumption, the influence of obstacles and acquisition coverage become more prominent, resulting in low efficiency and high energy consumption of the data acquisition process, and the available resources cannot be fully utilized.

[0003] In the prior art, the common three-dimensional data acquisition management system usually adopts a static path planning method, lacking dynamic selection of device layout points and real-time optimization of paths. Even some methods consider the optimization of the acquisition path, but usually only focus on a single factor, such as minimizing energy consumption or maximizing acquisition coverage, ignoring the need for multi-objective optimization. In addition, due to the existence of obstacles in the environment, the traditional system fails to effectively handle the occlusion problem, resulting in frequent data omission or incomplete acquisition in practical application. SUMMARY

[0004] In order to make up for the above shortcomings, the present application provides a three-dimensional data acquisition management system and method, aiming to improve the problem that the existing method usually focuses on a single target, ignoring how to balance multiple optimization targets, resulting in insufficient acquisition coverage or excessive energy consumption, and failing to provide an optimal solution after considering various factors.

[0005] In a first aspect, the present application provides the following technical solution, a three-dimensional data acquisition management method, comprising the following steps:

[0006] S1, obtaining a three-dimensional space model of an area to be acquired, and discretizing the space into a plurality of sampling units to form a voxel grid;

[0007] S2, identifying an occluded area in the three-dimensional space model, and constructing an effective reachable three-dimensional acquisition space;

[0008] S3, generating initial layout points according to device field of view parameters, and establishing an acquisition path control function for each device;

[0009] S4, establishing a field of view coverage relationship based on the acquisition parameters of the device and the position of the sampling unit, for judging whether each sampling unit is effectively covered by the device;

[0010] S5, constructing a joint optimization objective function considering the acquisition coverage, path energy cost, and spatial conflict penalty of the path and the occluded region;

[0011] S6, solving the optimal device layout point and path control function by an optimization method under the constraint condition, and outputting the optimal result for guiding the layout and scheduling of the actual three-dimensional data acquisition device.

[0012] Preferably, in S2, the effective reachable space is obtained by voxel modeling of the occluded region and excluding points with occlusion in any acquisition direction based on ray analysis, thereby obtaining the device layout region.

[0013] Preferably, in S3, the device path control function is established by a kinematics model, and the kinematics model is a time continuous function, the velocity input of which is constrained by the maximum velocity and acceleration of the device, and the initial layout point is taken as the path starting point.

[0014] Preferably, in S4, the established view coverage relationship is based on the following conditions:

[0015] The distance between the device and the sampling unit does not exceed the effective acquisition radius of the device, and the device is directed to the unit, and there is no occluded voxel between the line of sight and the unit.

[0016] Further comprising constructing a support graph structure between the device views, wherein each node in the support graph corresponds to a device view, and if the views of two devices have spatial overlap, a connection edge is established in the graph, thereby forming a continuous nested field of view chain structure.

[0017] Preferably, the constructed joint optimization objective function includes the following three sub-objectives:

[0018] S501, the proportion of the sampling unit not covered by any device in the total sampling unit;

[0019] S502, the total energy cost of all device acquisition paths, wherein the energy cost is represented by the integral of the path velocity function on the path interval;

[0020] S503, the penalty value generated by the traversal of the occluded region by all device paths during execution, which is determined by the spatial intersection measure of the path trajectory and the occluded voxel set.

[0021] Preferably, in S6, the optimization method includes an alternating direction multiplier method or a mixed integer programming method, and the layout point variable, the path variable, and the coverage relationship variable are jointly optimized to minimize the joint objective function value.

[0022] The path optimization is modeled based on optimal control theory, a Hamilton function is constructed, and a Lagrange multiplier method is used to determine optimal control variables, so that the energy consumption of the device path is minimized.

[0023] Preferably, the initial deployment point generation in the method adopts a heuristic sampling method, the method generates a plurality of candidate points according to the spatial boundary condition and the obstacle distribution, and selects a point with the highest visible coverage as the initial deployment position.

[0024] In a second aspect, the present application provides the following technical solution, a three-dimensional data acquisition management system, the system comprises:

[0025] A space modeling module is configured to discretely model a grid of an acquisition area and identify an obstacle area to form an acquisition space.

[0026] A deployment generation module is configured to generate an initial deployment point according to device parameters and establish a device path control function.

[0027] A view modeling module is configured to construct a view coverage relationship between a device and a space sampling unit and construct a support graph structure.

[0028] An optimization solving module is configured to construct a joint optimization target of coverage, path energy consumption and occlusion penalty, and solve an optimal device deployment scheme and path.

[0029] A scheduling control module is configured to issue the optimal result to a three-dimensional data acquisition device, and schedule and control device deployment and trajectory execution tasks.

[0030] In a third aspect, the present application provides the following technical solution, a computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor executes the computer program to implement the above-mentioned three-dimensional data acquisition management method.

[0031] In a fourth aspect, the present application provides the following technical solution, a readable storage medium, the readable storage medium stores a computer program, and the computer program is executed by a processor to implement the above-mentioned three-dimensional data acquisition management method.

[0032] The present application has the following beneficial effects:

[0033] 1. In the present application, a three-dimensional space acquisition model is constructed, a joint optimization objective function is adopted, and dynamic optimization of device deployment points and paths is realized, considering multiple factors such as acquisition coverage, energy consumption and occlusion area penalty, so as to obtain an optimal device deployment scheme and path control strategy, and effectively improve the acquisition efficiency and the rationality of path planning.

[0034] 2、The present application establishes the relationship between the field of view of the device and the spatial voxel coverage, so that the collection coverage of each voxel can be comprehensively evaluated, combined with the optimal control and path optimization algorithm, to achieve higher spatial collection coverage and accuracy, reduce the uncovered area, and improve the integrity and accuracy of the collection task.

[0035] 3、In the present application, by introducing the Hamilton optimal control theory and the path energy consumption model, the energy consumption of the device during path execution can be effectively reduced while optimizing the device collection path, the energy consumption and path conflict of the device operation are minimized to the greatest extent, and the operation efficiency and sustainability of the entire collection system are improved. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 A method flowchart for a three-dimensional data collection management method is proposed for the present application;

[0037] Figure 2 A system module relationship architecture diagram for a three-dimensional data collection management system is proposed for the present application. DETAILED DESCRIPTION

[0038] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0039] Embodiment one

[0040] Reference Figure 1 In the first embodiment of the present application, the present application provides a three-dimensional data collection management method, comprising the following steps:

[0041] S1, obtaining a three-dimensional space model of the area to be collected, and discretizing the space into a plurality of sampling units to form a voxel grid;

[0042] S2, identifying the occluded area in the three-dimensional space model, and constructing an effective three-dimensional collection space;

[0043] S3, generating an initial layout point according to the field of view parameters of the device, and establishing a collection path control function for each device;

[0044] S4, based on the collection parameters of the device and the position of the sampling unit, establishing a field of view coverage relationship for judging whether each sampling unit is effectively covered by the device;

[0045] S5, construct a joint optimization objective function, which considers the acquisition coverage, path energy cost and path and occlusion region space conflict penalty;

[0046] S6, under the constraint condition, the optimal device layout point and path control function are solved by optimization method, and the optimal result is output to guide the layout and scheduling of actual three-dimensional data acquisition device.

[0047] In S2, the effective reachable space is obtained by voxel modeling of the occlusion region and excluding the points that are blocked in any acquisition direction based on ray analysis, so as to obtain the device judgment layout region.

[0048] Specifically, the construction of the effective reachable space relies on the voxel modeling result of the occlusion region, and combines the multi-directional ray projection analysis method to realize the accurate identification of the device layout position. The system first discretely processes all known obstacle bodies in the three-dimensional space, maps them into voxel grid form, and forms an obstacle voxel set Each voxel unit has a volume boundary definition, and its voxel center is denoted as The obstacle region is no longer continuously processed in the form of surface or body, but is converted into a point set expression with uniform spatial scale. This conversion is direction-independent and facilitates subsequent algorithm processing. The next key processing link is to determine whether there is an occlusion voxel blocking in each direction based on the ray-tracing based visibility check. The processing takes a to-be-laid point As the starting point, and performs spatial path simulation along multiple preset directions

[0049] ;

[0050] Wherein, is the maximum projection distance (usually set as the upper limit of the maximum perception range of the device), and the vector can be uniformly distributed on the spherical surface to ensure the uniformity of direction coverage. For each path , it is discretized into a continuous point column with a step size of , and each point is checked whether it passes through the obstacle voxel set .

[0051] If there is any point in the path falling within the range of the obstacle voxel in a certain direction, it is considered that the field of view in this direction is limited. The judgment not only depends on whether the center points overlap, but also considers the voxel boundary buffer value (a -neighborhood judgment mechanism can be introduced), to prevent the occlusion judgment from being too harsh. Finally, only the points that meet the following conditions are reserved as the layout positions:

[0052] ​ ;

[0053] This means that the device from this point, there is at least one unobstructed direction can be executed perception task. If all directions exist obstacles block, the point is rejected from the layout candidate set. It is worth noting that this way based on voxel occlusion decision compared to traditional geometric model or triangle net occlusion judgment more robust, and can effectively avoid the boundary geometry calculation error. By voxel scale uniformity, can guarantee the consistency of different space region occlusion analysis, facilitate uniform processing. Occlusion judgment support parallel execution, and the path judgment way is naturally applicable to GPU or spatial index optimization, after the construction is completed, the system outputs the voxel set As the device can actually be deployed area. The set of stable spatial distribution, not with modeling error fluctuation, with repeatability and engineering feasibility.

[0054] S3, the device path control function is established by kinematic model, the kinematic model is a time continuous function, the speed input is constrained by the maximum speed and acceleration of the device, and the initial layout point is taken as the path starting point.

[0055] Specifically, in the device scheduling system, the generation of the path is not an isolated process, which depends on the dynamic behavior modeling based on the layout area. After the determination of the device layout candidate point set is completed, the next stage of constructing the device path control function is entered. The path is not only a connecting line in the geometric sense, but also a time-continuous action planning constrained by the motion law and execution constraints. Therefore, the introduction of kinematic modeling becomes the key, and the device path control function adopts a continuous time domain expression form to describe the position change process of the device in the three-dimensional space. Let the path function of the device in time be:

[0056] ;

[0057] Where , represents the spatial position at time . The initial condition is that the device starts from the layout point , and the terminal time can be regarded as the planning period or the completion time of the task, and the change of the path control function is limited by the actual motion ability of the device, including the maximum linear speed and the maximum acceleration. For this purpose, the velocity function is defined as:

[0058] ;

[0059] And the acceleration function is:

[0060] ;

[0061] The following constraints are applied in the present application:

[0062] ;

[0063] - maximum acceleration constraint:

[0064] ;

[0065] wherein, and are upper limits set in advance according to the physical structure or platform specifications of the device in the system. Such constraints avoid non-physical movements in the path, such as instantaneous acceleration, U-turns, etc.

[0066] The specific construction method of the path function can vary depending on the device category. For example, for flight devices (such as drones), it is preferred to use a spline curve (such as a cubic B-spline) to interpolate the path modeling, as it naturally satisfies the continuity of position and velocity. The number of spline parameter nodes can be dynamically adjusted according to the complexity of the sampling area, both preserving the expression ability of the path and avoiding excessive redundancy, and the path initial value constraint ensures that the path starts from the device deployment point without drift error. That is:

[0067] ;

[0068] In the process of establishing the path control model, to avoid problems such as excessive curvature of the path and frequent changes in direction, a curvature constraint or soft constraint penalty term is introduced to impose a regulating factor on the second derivative of the path function. This term can be used as an energy consumption proxy in the subsequent optimization model, and it also helps to generate a natural, smooth, and executable path. The path function not only has temporal continuity in data representation, but also fully reflects the dynamic constraint conditions faced by the device in actual operation. Establishing a path control function on the deployable area is an expression of spatial freedom, and the rationality of the path function determines the feasibility of the final scheduling scheme and the ability to ensure high-quality collection.

[0069] The field of view coverage relationship established in S4 is based on the following conditions:

[0070] The distance between the device and the sampling unit does not exceed the effective sampling radius of the device, and the device is oriented towards the unit, and there is no occlusion voxel between the line of sight and the unit;

[0071] It also includes constructing a support graph structure between the device fields of view, where each node in the support graph corresponds to a device field of view. If the fields of view of two devices overlap in space, a connection edge is established in the graph, forming a continuous nested field of view chain structure.

[0072] Specifically, first is the distance condition. The straight-line distance between the device and the target point must not exceed its preset maximum collection radius. This constraint reflects the physical limit of the device's perception ability. In practical applications, the perception radius can be flexibly set according to the device type, resolution requirements, or optical conditions;

[0073] Second is the directionality judgment. Not all voxels within the radius are perceptible. The device must be oriented towards the point, and the target voxel must fall within its field of view. In other words, from the current orientation of the device, the angle formed by the voxel position and the device must be within the allowed fan-shaped area for collection. If the back of the device is directly opposite the voxel, even if the distance is close, it is considered invalid. Third, the line of sight is unobstructed. Even if the distance is appropriate and the direction is correct, if the line of sight passes through an obstacle (which has been labeled as an obstructing voxel) in the middle, then this point is still considered invisible. The system will determine whether the line connecting the device and the voxel is penetrated by any obstacle. If there is an obstruction, it is considered "obstructed";

[0074] Only voxels that meet all three conditions will be marked as effectively covered by the device. Such mapping results have a direct impact on subsequent calculations of collection coverage, path benefit evaluation, etc. The system will define the set of all voxels that can be covered by the device as its field of view support set, as the effective perception range under the current path control state;

[0075] A support graph structure between the fields of view of devices is also introduced. This is an abstract data structure that describes the spatial coupling relationship between the fields of view of multiple devices;

[0076] The field of view of each device at any moment in the path is considered a node in the graph. If the fields of view of two devices have an intersection, i.e., there is at least one voxel that can be perceived by both devices, a non-directional connection edge is established between the two nodes. This connection represents the information exchangeability or coverage overlap relationship between the markets. The construction of this support graph not only helps to judge coverage redundancy, but also provides a structured input for subsequent joint optimization models. It can reveal which devices are locally collaborative, which areas have visual islands, and it can also find paths with high overlap, thereby assisting in dynamic adjustment of path planning and point allocation. In addition, the support graph naturally has scalability, and as the number of devices increases or the path changes, the graph structure can be updated in real time, ensuring the timeliness and accuracy of the entire system field of view topology.

[0077] The constructed joint optimization objective function includes the following three sub-goals:

[0078] S501, the proportion of sampling units not covered by any device in the total sampling units;

[0079] S502, all devices collect the total energy cost of the path, wherein the energy cost is represented by the integral of the path speed function on the path interval;

[0080] S503, all devices path in the execution process through the occlusion region generated by the penalty value, the penalty value is determined by the spatial intersection measure of path trajectory and occlusion voxel set.

[0081] Specifically, first in order to achieve efficient, complete, stable data acquisition in three-dimensional space, the system needs to integrate multiple objectives, under the given conditions of the layout area and path control function, by solving a joint optimization model, to obtain the optimal coverage, reasonable energy consumption and safe obstacle avoidance scheduling scheme. To this end, the following joint objective function is constructed:

[0082] ;

[0083] Wherein: : the proportion of uncovered voxels : path energy consumption : occlusion conflict penalty : weight coefficient, used to adjust the importance of the three

[0084] Second, this item measures how many target voxels are not covered by any device. Define the target sampling voxel set as If a voxel is not covered by any device, it is marked as "uncovered". Define the binary function , when is covered, it is 1, otherwise it is 0:

[0085] ;

[0086] This item reflects the overall sampling accuracy, the lower the value, the more comprehensive the coverage. The system considers the motion energy consumption generated by each device along its path, not simply the path length measurement, but based on the speed function to model the energy consumption by integral. Let the path of the th device be a time continuous function , the speed is , then its energy consumption can be approximated as the integral of the square norm of the speed:

[0087] ;

[0088] Where is the path execution time of the th device. This model implicitly prefers smooth paths to avoid frequent acceleration / deceleration. If the device path traverses the occlusion region during execution, i.e. the trajectory has a non-empty intersection with , a penalty is generated. Herein,

[0089] ;

[0090] wherein is an indicator function of the obstacle region, taking value 1 when the path point falls into the obstacle region, otherwise 0. The integral result represents the "length of time" that the device interacts with the obstacle region during the running process, and the larger the value is, the more path exists unpassable crossing behavior.

[0091] In S6, the optimization method includes an alternating direction multiplier method or a mixed integer programming method, and the joint objective function value is minimized by jointly optimizing the layout point variable, the path variable and the covering relationship variable;

[0092] The path optimization is modeled based on optimal control theory, the optimal control variable is determined by constructing a Hamilton function and combining a Lagrange multiplier method, so that the energy consumption of the device path is minimized.

[0093] Specifically, in order to minimize the energy consumption of the device path, a static planning or heuristic path approximation method is not used, but the path optimization process is converted into a typical optimal control problem for solving. This modeling method can fine-tune the system performance index while meeting the motion constraints, and the device path function is defined as a state variable that changes over time, and the control variable of the device is the instantaneous speed , and the goal is to control how the device "walks" to walk out an optimal trajectory with minimum energy consumption and constraint satisfaction between the starting point and the ending point, wherein the state control model is defined as: state variable: ;

[0094] Control variable: , and the dynamic equation is written as:

[0095] ;

[0096] Boundary conditions: the initial position is the layout point: , if there is a terminal point constraint, then specify , otherwise free termination in space, and the objective function is the minimum energy consumption path, that is, the quadratic cost function of the control input:

[0097] ;

[0098] In order to introduce the optimal control theory framework, the Hamilton function is constructed by the present application:

[0099] ;

[0100] wherein, For the co-state variable (i.e. the Lagrange multiplier), representing the sensitivity of the system state to the objective function, the optimal control is derived from the Pontryagin's minimum principle The Hamiltonian must be minimized at each time instant and must satisfy the following set of necessary conditions:

[0101] ;

[0102] ;

[0103] ;

[0104] From the above conditions, it can be seen that there is a linear proportional relationship between the optimal velocity and the co-state variable, which is a constant vector in this example. This means that the optimal path corresponds to a smooth trajectory with uniform energy distribution, and the control input changes linearly with time, which is consistent with the characteristics of the actual device dynamic model.

[0105] The initial deployment point generation in the method uses a heuristic sampling method, and the method generates a plurality of candidate points according to the spatial boundary conditions and the obstacle distribution, and selects the point with the highest visible coverage as the initial deployment position.

[0106] Specifically, the initial deployment point generation in the method uses a heuristic sampling method, and the method generates a plurality of candidate points according to the spatial boundary conditions and the obstacle distribution, and selects the point with the highest visible coverage as the initial deployment position, and the generation area of the candidate points is limited within the space region determined as the deployable space region in the previous step, that is, the voxel set On this basis, the system constructs the candidate point set according to the following heuristic rules: boundary constraint: all candidate points need to satisfy the spatial closed boundary condition, to avoid high-risk areas such as close to the wall and the edge of the site. By setting the edge buffer distance, the critical points can be automatically excluded, obstacle avoidance: the minimum Euclidean distance between the candidate point and any obstacle voxel should be greater than or equal to the device volume safety radius and the perception field boundary, to prevent the initial penetration of the subsequent path from being blocked, direction distribution uniformity: to ensure comprehensive field coverage, the sampling direction should be as dispersed as possible in space, to avoid candidate points being highly concentrated in a certain area leading to redundant deployment;

[0107] After generating the candidate points, the system will evaluate the visibility of each point, taking the point as the virtual layout position of the device, combining the device's view range and the collection radius, to detect how many effective sampling voxels can be covered from this position. This number is the visibility coverage index of the point. During the visibility detection process, the occlusion removal mechanism described earlier is also applied to ensure that all statistical voxels are unobstructed, reachable, and fall within the device's effective field of view. The detection method can be ray casting or directional sampling beam method, depending on the device's performance. After obtaining the visibility coverage of all candidate points, the system will sort the indexes and select a group of points with the highest coverage as the initial layout points. The goal is to make the initial layout have greater global field of view support capability and improve the search quality of the entire scheduling system.

[0108] Embodiment Two:

[0109] With reference to Figure 2 In the second embodiment of the present application, the present application provides a three-dimensional data collection management system, the system comprising:

[0110] a space modeling module for grid discrete modeling of the collection area and identifying obstacle regions to form a collection space;

[0111] a layout generation module for generating initial layout points according to device parameters and establishing device path control functions;

[0112] a visibility modeling module for constructing the visibility coverage relationship between the device and the space sampling unit and constructing a support graph structure;

[0113] an optimization solving module for constructing a joint optimization target of coverage rate, path energy consumption and occlusion penalty, and solving the optimal device layout scheme and path;

[0114] a scheduling control module for issuing the optimal result to the three-dimensional data collection device, scheduling and controlling the device layout and trajectory execution tasks.

[0115] Specifically, the space modeling module: this module is responsible for grid discrete modeling of the collection area and identifying obstacle regions, and finally forming a complete collection space. The grid space is convenient for subsequent path optimization and visibility analysis. The identification of obstacle regions can be based on known obstacle positions or realized through real-time feedback information from sensors;

[0116] the layout generation module: according to device parameters (such as the maximum collection radius, sensing ability, etc. of the device), this module generates multiple initial layout points and establishes the path control functions of the device. The generation of initial layout points is through a heuristic sampling method to ensure that the selected layout points can maximize the coverage area and avoid obstacles;

[0117] View modeling module: this module is responsible for constructing the view coverage relationship between the device and the space sampling unit. The view coverage relationship takes into account the perception range, directionality of the device and the influence of the occlusion. Through view modeling, the system can determine which areas each device can cover, and construct a support graph structure between devices, representing the overlap of the field of view between devices and the possibility of information exchange;

[0118] Optimization solving module: this module considers multiple factors such as coverage rate, path energy consumption and occlusion penalty by constructing a joint optimization objective function. Through optimization of the objective function, the optimal device layout scheme and path are solved, so as to realize the path planning with minimum energy consumption, maximum coverage rate and best obstacle avoidance performance;

[0119] Scheduling control module: finally, the optimal result output by the optimization solving module will be issued to the three-dimensional data acquisition device through the scheduling control module. The scheduling control module is responsible for scheduling and controlling the device layout and trajectory execution tasks, ensuring that the device performs the path according to the optimal scheme and completes the data acquisition task;

[0120] Embodiment three

[0121] The third embodiment of the application is based on the same inventive concept, and the application provides a computer readable storage medium storing a computer program, which is executed by a processor to implement the steps of the three-dimensional data acquisition management method of the above-mentioned embodiment.

[0122] Embodiment four

[0123] The fourth embodiment of the application is based on the same inventive concept, and the application provides a computer device, which comprises a processor and a memory. The processor and the memory communicate with each other. The memory is used to store instructions, and the processor is used to execute the instructions in the memory to implement the three-dimensional data acquisition management method of the above-mentioned embodiment.

[0124] It should be understood that the parts of the application can be realized by hardware, software, firmware or their combination. In the above-mentioned embodiments, multiple steps or methods can be realized by software or firmware stored in the memory and executed by a suitable instruction execution system. For example, if realized by hardware, and as in another embodiment, it can be realized by any one or their combination of the following technologies known in the art: discrete logic circuit with logic gate circuit for implementing logic function on data signal, application specific integrated circuit with suitable combination logic gate circuit, programmable gate array (PGA), field programmable gate array (FPGA) and the like.

[0125] Finally, it should be noted that the above only describes the preferred embodiments of the present application and is not intended to limit the present application. Although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art will appreciate that the technical solutions described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalent features, and any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for three-dimensional data acquisition and management, characterized in that, Includes the following steps: S1. Obtain the three-dimensional spatial model of the area to be collected, and discretize the space into multiple sampling units to form a voxel grid; S2. Identify the occlusion areas in the three-dimensional spatial model and construct an effectively accessible three-dimensional acquisition space; In S2, the effectively accessible space is obtained by voxel modeling the occlusion areas and excluding points that are occluded in any acquisition direction based on ray analysis, thereby obtaining the area where the device can be deployed. S3. Generate initial deployment points based on the equipment field of view parameters and establish the acquisition path control function for each equipment. In S3, the equipment path control function is established through a kinematic model. The kinematic model is a time-continuous function, and its velocity input is constrained by the maximum velocity and acceleration of the equipment, with the initial deployment point as the starting point of the path. S4. Based on the device's acquisition parameters and sampling unit positions, establish a view coverage relationship to determine whether each sampling unit is effectively covered by the device; the established view coverage relationship in S4 is based on the following conditions: The distance between the device and the sampling unit shall not exceed the effective acquisition radius of the device, and the device shall be oriented towards the unit, with no obstruction between the line of sight and the unit; It also includes constructing a support graph structure between device view domains, where each node in the support graph corresponds to a device view domain. If the view domains of two devices overlap spatially, a connecting edge is established in the graph to form a continuously nested view domain chain structure. S5. Construct a joint optimization objective function, which simultaneously considers acquisition coverage, path energy consumption cost, and spatial conflict penalty between the path and the occluded area; the constructed joint optimization objective function includes the following three sub-objectives: S501, The proportion of sampling units not covered by any device in the total sampling units; S502, the total energy consumption cost of all device acquisition paths, wherein the energy consumption cost is represented by the integral of the path velocity function over the path interval; S503, The penalty value generated when all device paths cross the occlusion area during execution, the penalty value is determined by the spatial intersection measure of the path trajectory and the occlusion voxel set; S6. Under constraints, the optimal equipment placement points and path control functions are solved by optimization methods, and the optimal results are output to guide the deployment and scheduling of actual 3D data acquisition equipment. In S6, the optimization method includes the alternating direction multiplier method or the mixed integer programming method, which minimizes the joint objective function value by jointly optimizing the layout point variables, path variables, and coverage relationship variables. Path optimization is based on optimal control theory modeling. By constructing a Hamiltonian function and combining it with the Lagrange multiplier method, the optimal control variables are determined, thereby minimizing the energy consumption of the equipment path. The initial deployment point generation in the method adopts a heuristic sampling method. The method generates multiple candidate points based on spatial boundary conditions and obstacle distribution, and selects the point with the highest visibility coverage as the initial deployment location.

2. A three-dimensional data acquisition and management system, characterized in that, A method for three-dimensional data acquisition and management as described in claim 1, the system comprising: The spatial modeling module is used to perform grid discretization modeling of the acquisition area and identify obstacle areas to form the acquisition space; The deployment generation module is used to generate initial deployment points based on equipment parameters and establish equipment path control functions; The view modeling module is used to construct the view coverage relationship between the device and the spatial sampling unit and to build a supporting graph structure. The optimization solution module is used to construct a joint optimization objective of coverage, path energy consumption and occlusion penalty, and solve for the optimal device deployment scheme and path; The scheduling and control module is used to send the optimal results to the three-dimensional data acquisition equipment and to schedule and control the equipment deployment and trajectory execution tasks.

3. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements a method for three-dimensional data acquisition and management as described in claim 1.

4. A readable storage medium, characterized in that, The readable storage medium stores a computer program, which, when executed by a processor, implements the three-dimensional data acquisition and management method as described in claim 1.

Images