Stock inventory method and system for stockyard stockpile, terminal and medium

Abstract

The invention belongs to the technical field of stockyard management, and particularly discloses a stockyard stockpile inventory method and system, a terminal and a medium, and the stockyard stockpile inventory method comprises the steps: collecting original spatial data; performing point cloud registration, filtering and noise reduction processing on the original data, and constructing a three-dimensional model of the stockpile; and based on the three-dimensional model, calculating the volume of the stockpile by adopting an integral algorithm, and estimating the weight of the stockpile in combination with a material density parameter. A preset flight task is executed through the unmanned aerial vehicle to automatically collect three-dimensional data of a stockpiling area, manual measurement and climbing operation procedures are omitted, the inventory cycle is greatly shortened, and the overall inventory efficiency of a stockyard is remarkably improved; an air scanning mode is adopted to replace manual short-distance measurement, so that operators are prevented from approaching high heap or dangerous areas, and the operation risk is effectively reduced; meanwhile, full coverage of the material pile is achieved through path planning and multi-view scanning, the space density and uniformity of point cloud data are improved, and then the volume calculation accuracy is improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of stockyard management, and particularly relates to a stocktaking method and system for stockyard stacking, a terminal and a medium. BACKGROUND

[0002] A stockyard is an important area for material centralized stacking in industrial production, mineral transportation, port operation and the like. Accurate stocktaking of the stacking volume and weight is a key link for realizing fine management, inventory checking, loss control and scheduling optimization. At present, stockyard stocktaking work is mainly completed by manual or semi-automatic methods. A surveyor generally carries a total station instrument, a handheld laser range finder and the like to manually measure the stacking boundary, and then estimates the volume and calculates the weight by combining the material density.

[0003] However, the existing manual stocktaking method has the following problems. First, manual measurement requires personnel to walk around the stacking area, which has obvious operation safety risks, especially in high stacking and high slope stockyards. Second, manual measurement has low efficiency and limited coverage, and is difficult to meet the rapid inventorying needs of large-area and multi-region stockyards. Third, due to limited and unevenly distributed measurement points, the estimation result is easily affected by human experience, resulting in poor volume calculation accuracy and affecting the accuracy of inventory management. SUMMARY

[0004] The application provides a stocktaking method, system, terminal and medium for stockyard stacking to solve the problems of low efficiency, high safety risk and poor calculation accuracy in the prior art.

[0005] The technical scheme adopted by the application is as follows: In a first aspect, the application provides a stocktaking method for stockyard stacking, which comprises the following steps: Step S1, obtaining a scanning task instruction of a stockyard stacking area; Step S2, controlling a UAV to scan above the stacking area according to a preset flight path to obtain original spatial data in the scanning process; Step S3, performing registration processing on the original spatial data; Step S4, performing filtering and noise reduction processing on the registered spatial data; Step S5, constructing a three-dimensional model of the stacking based on the processed spatial data; Step S6, calculating the stacking volume according to the three-dimensional model, and generating weight information in combination with a preset density parameter.

[0006] Further, in step S2, the preset flight path comprises a scanning trajectory automatically generated based on the stacking boundary contour, and the scanning trajectory is determined by the following method: A two-dimensional contour map of the heap boundary projection is obtained, a minimum enclosing rectangular region is constructed using a contour bounding box algorithm, and parallel equidistant scanning paths are generated according to the long side direction of the rectangle; The spacing Δd between the parallel scanning paths is calculated according to the following formula:

[0007] where h is a preset unmanned aerial vehicle operation flight height, θ is an angle of view of a scanning device on the unmanned aerial vehicle, k is a sampling overlap coefficient, and k>1.

[0008] Further, in step S2, the unmanned aerial vehicle dynamically adjusts the flight path by real-time sensing of the position of environmental obstacles while performing the scanning path flight, in combination with an A* pathfinding algorithm.

[0009] Further, in step S3, the registration processing of the original spatial data includes inter-frame registration of point cloud data under different scanning angles, and the inter-frame registration uses an iterative closest point algorithm; Before performing the iterative closest point registration, initial registration is performed based on unmanned aerial vehicle trajectory information and pose data provided by an inertial navigation system.

[0010] Further, a registration error threshold ε is set during the registration process, and the iterative closest point algorithm terminates iteration when the following convergence conditions are met:

[0011] N is the total number of point cloud registration point pairs, is the distance between the source point cloud and the coordinate after registration transformation, is the nearest neighbor point in the target point cloud that matches .

[0012] Further, in step S4, the filtering and noise reduction processing of the registered spatial data includes: performing voxel grid filtering on the target point cloud; After voxel filtering, a statistical outlier rejection algorithm is used to determine whether each point is an outlier based on the average distance of the point from points in its neighborhood, and noise points with a distance exceeding a set standard deviation threshold are rejected.

[0013] Further, in step S5, triangular mesh reconstruction is performed based on the point cloud data after filtering and noise reduction processing, and a Poisson surface reconstruction algorithm is used to generate a continuous surface model; In step S6, a volume integral algorithm is used to calculate the volume V, which is determined by the following formula:

[0014] wherein an area of the i-th triangular mesh in a horizontal projection plane, a height of a barycenter of the triangular facet to a reference datum plane; obtain a material density ρ corresponding to the type of the stockpile, and calculate a weight W of the stockpile according to a formula:

[0015] In a second aspect, the present application provides a stocktaking system for a stockpile, which is used to implement the stocktaking method for a stockpile as described in the first aspect. The system comprises: a scanning instruction obtaining module, configured to obtain a scanning task instruction of a stockpile area; a UAV flight control module, configured to control a UAV to scan above the stockpile area according to a preset flight path, and obtain original spatial data in the scanning process; a point cloud registration module, configured to perform registration processing on the original spatial data, including initial registration estimation based on UAV trajectory information and pose data provided by an inertial navigation system, and frame-to-frame point cloud alignment using an iterative closest point algorithm; a point cloud processing module, configured to perform voxel grid filtering and statistical outlier rejection processing on the registered spatial data; a three-dimensional modeling module, configured to construct a three-dimensional model of the stockpile based on the processed spatial data, and the modeling module generates a continuous curved surface model using a Poisson surface reconstruction algorithm; a volume calculation and weight estimation module, configured to calculate a volume of the stockpile using an integral algorithm based on the three-dimensional model, and generate corresponding weight information in combination with a preset material density parameter; a data output module, configured to output the volume and weight information of the stockpile for calling by a stocktaking management system.

[0016] In a third aspect, the present application provides a terminal, comprising: a memory, configured to store a stocktaking program for a stockpile; a processor, configured to execute the stocktaking program for a stockpile to implement the steps of the stocktaking method for a stockpile as described in the first aspect.

[0017] In a fourth aspect, the present application provides a computer readable storage medium, which stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes the stocktaking method for a stockpile as described in the first aspect.

[0018] It can be seen from the above technical solutions that the present application has the following advantages: ​(1) The unmanned aerial vehicle is used to automatically collect three-dimensional data of the stacking area according to a preset flight task, manual measurement and climbing operation processes are saved, the inventory cycle is greatly shortened, and the overall efficiency of the stockpile inventory is significantly improved; the manual close-range measurement is replaced by the aerial scanning mode, the operation personnel are prevented from approaching the high pile or dangerous area, and the operation risk is effectively reduced; meanwhile, the path planning and multi-view scanning are used to realize full coverage of the pile, the spatial density and uniformity of the point cloud data are improved, and the volume calculation accuracy is improved.

[0019] (2) The point cloud registration, filtering, triangular mesh reconstruction and volume calculation algorithm are combined, the system automatically completes the whole process from the original data to the inventory result, the human intervention and experience dependence are reduced, and the consistency and traceability of the inventory result are ensured.

[0020] (3) The system can be adapted to stacking areas with different heights, angles and irregular shapes, a curved surface reconstruction algorithm is used to construct a real three-dimensional topographic model, and the system is suitable for various application scenarios such as ports, mining areas and power plants, and has wide applicability. BRIEF DESCRIPTION OF DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the present application, the drawings needed in the description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained without creative labor on the basis of these drawings.

[0022] Figure 1 The flowchart of the inventory method for the stockpile provided by the embodiments of the present application is shown. DETAILED DESCRIPTION

[0023] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. 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.

[0024] Please refer to Figure 1 The present application provides an inventory method for a stockpile, which comprises the following steps: Step S1, obtaining a scanning task instruction of a stockpile stacking area; The purpose of this step is to determine the target area and task range of the inventory operation, and to provide a task basis for subsequent unmanned aerial vehicle flight and data collection. The scanning task instruction can be generated by a background scheduling system according to a time plan, a stacking in-out warehouse record or manual manual issuance.

[0025] In practical applications, the system management platform automatically generates a scanning task instruction according to the stockyard operation plan, and the instruction contains the geographic boundary coordinates of the stockyard area, the scanning start and end time, the stockyard type information and the required flight accuracy requirement. The operator can also manually select the area to be scanned through the interface and issue the corresponding task. After the task is issued, the system uploads the task data to the unmanned aerial vehicle control system for subsequent execution.

[0026] Step S2, controlling the unmanned aerial vehicle to scan above the stockyard area according to the preset flight path to obtain original spatial data in the scanning process; This step performs aerial scanning on the stockyard area by the laser radar or photographic equipment carried by the unmanned aerial vehicle to collect three-dimensional spatial information. The flight path needs to be planned in advance to ensure complete coverage of the area and reasonable scanning angle.

[0027] After receiving the task instruction, the unmanned aerial vehicle automatically generates a scanning path according to the stockyard boundary profile, and the path is generally an equidistant parallel flight path, and sets the take-off height, flight speed, flight path interval and other parameters. The unmanned aerial vehicle automatically flies according to the path, continuously collects spatial data during flight, and stores the original point cloud or image data in the on-board storage device in real time, and part of the data can be transmitted to the ground station through the wireless link for backup.

[0028] Step S3, registration processing of the original spatial data; Since the unmanned aerial vehicle collects data from different angles and positions during flight, the data collected in different batches are inconsistent in space, so registration processing is needed to align the multi-angle data to a continuous three-dimensional structure in a unified coordinate system.

[0029] The registration processing is divided into two stages of initial registration and fine registration. First, according to the trajectory data and attitude information recorded during the flight of the unmanned aerial vehicle, the collected data is subjected to preliminary spatial transformation to roughly align it. Then, a point cloud inter-frame comparison algorithm is used to finely align the multiple batches of point clouds, gradually optimize the relative position relationship, eliminate position errors, and unify all data in the same reference coordinate system to form a continuous three-dimensional point cloud data set.

[0030] Step S4, filtering and noise reduction processing of the registered spatial data; The original point cloud may contain noise data such as floating points, edge burrs, and abnormal points caused by device jitter, which need to be cleaned through filtering and noise reduction to improve the accuracy and stability of three-dimensional modeling.

[0031] First, the point cloud data is subjected to voxel filtering, i.e. the point cloud space is divided into a plurality of cubic grids, and a representative point in each grid is retained, thereby reducing the data amount and homogenizing the distribution. Then, statistical methods are used to detect and eliminate outliers. The system analyzes the distance distribution of each point and the points in its neighborhood, identifies abnormal points deviating from the normal density distribution, and deletes them, thereby obtaining a clean and continuous point cloud data set.

[0032] Step S5, constructing a three-dimensional model of the stockpile based on the processed spatial data; By geometric reconstruction of the processed point cloud data, a three-dimensional surface model is formed which can be used for volume calculation. The model should accurately reflect the true shape of the stockpile, supporting subsequent volume analysis.

[0033] The system performs grid processing on the cleaned point cloud to construct a triangular facet structure, converting the point cloud into a geometric model with a closed surface topology. Preferably, a reconstruction algorithm based on implicit function field solving is used to automatically fill in the small gaps in the point cloud and generate a continuous and smooth model surface. The model can support three-dimensional rotation, scaling and cross-section analysis, providing a geometric basis for inventory calculation.

[0034] Step S6, calculating the volume of the stockpile according to the three-dimensional model, and generating weight information in combination with the pre-set density parameters; Using the built three-dimensional model, the volume of the stockpile is obtained through spatial geometric calculation, and the weight is estimated in combination with the material density information, thereby completing the automatic inventory.

[0035] The system adds the volume of each small unit by the height difference between the model surface and the reference plane, in combination with the boundary range of the model bottom, to finally obtain the total volume of the entire stockpile. Subsequently, the pre-set material density data in the stockyard database is called, the corresponding density parameters are matched according to the type of the stockpile, the weight of the stockpile is calculated, and a complete inventory report is generated, which can be exported to an enterprise resource management system or a visualization platform.

[0036] In some embodiments, in step S2, the pre-set flight path includes a scanning trajectory automatically generated based on the stockpile boundary contour, which is determined by the following method: A two-dimensional contour map of the stockpile boundary projection is obtained, a minimum bounding rectangle region is constructed using a contour bounding box algorithm, and a parallel equidistant scanning path is generated according to the long side direction of the rectangle; In order to improve the efficiency and accuracy of the unmanned aerial vehicle scanning, the flight path planning needs to be automatically generated according to the spatial distribution of the stockpile. Especially in the case of irregular shape and scattered distribution of the stockpile in the stockyard, by automatically identifying the stockpile boundary and generating a reasonable flight scanning trajectory, the coverage rate can be significantly improved, the redundant flight can be reduced, and the completeness of the point cloud data can be ensured.

[0037] To this end, the system first projects the boundary of the stockpile to obtain a two-dimensional contour map of the stockpile on the horizontal plane. Then, the contour map is processed using geometric analysis to construct a minimum rectangular surrounding area that closely adheres to the boundary of the stockpile. After determining the surrounding area, the system takes the long side of the rectangle as the reference direction and plans several equally spaced parallel flight paths in this direction as the scanning path of the UAV. This method not only ensures the continuity and integrity of the scanning, but also improves the uniformity of spatial data coverage, facilitating subsequent point cloud modeling.

[0038] In a specific implementation, the system first calls remote sensing images or rough point cloud data collected in previous flights to perform boundary extraction processing on the stockpile in the stockyard. The extraction method can use a contour extraction algorithm to identify the edge lines of the stockpile on the projection plane and form a closed contour map.

[0039] Subsequently, the system calculates the minimum enclosing rectangle based on the boundary contour. The rectangle refers to a rectangular area that closely adheres to the boundary and covers the entire stockpile area, and its direction is automatically determined based on the main direction of the contour. The system takes the long side direction of the rectangle as the flight main direction and generates multiple parallel and evenly spaced scanning paths within the rectangle. The start and end points of these paths are located between the two side boundaries of the rectangle and extend to cover the entire stockpile boundary area.

[0040] The UAV flies along each generated scanning trajectory in turn and turns on the scanning device during flight to collect spatial data in real time. After scanning is complete, the system evaluates the sampling completeness based on the flight path record and data coverage, and if necessary, can supplement local paths to improve edge accuracy.

[0041] The spacing Δd between the parallel scanning paths is calculated according to the following formula:

[0042] where h is the preset UAV operation flight height, θ is the viewing angle of the scanning device on the UAV, and k is the sampling overlap coefficient, k > 1.

[0043] In some embodiments, in step S2, the UAV adjusts the flight path dynamically by real-time sensing of the environment obstacle position while performing the scanning path flight, combined with the A* pathfinding algorithm.

[0044] In the stockyard, due to the presence of mobile equipment, temporary storage, structures and other obstacles in the stockpile area, in order to ensure the safety of the UAV flight and the continuity of data collection, it is necessary to have real-time obstacle avoidance capability during flight. By dynamically sensing the surrounding environment and introducing a pathfinding algorithm in path planning, adaptive adjustment of the UAV path can be achieved.

[0045] During flight execution, the obstacle avoidance sensor carried by the UAV monitors the obstacle information in the space in front of the flight in real time. When the system detects that there is an obstacle on the scanning path, the flight control unit calls the pathfinding algorithm to re-plan an obstacle-free path in the stockpile area according to the accessibility between the current flight position and the target path node. The path ensures that the flight trajectory avoids all detected physical obstacles while meeting the scanning coverage requirements, thereby ensuring the stability and safety of the flight process.

[0046] In some embodiments, in step S3, the registration processing on the original spatial data includes inter-frame registration on the point cloud data under different scanning angles, and the inter-frame registration adopts an iterative closest point algorithm. Before performing the iterative closest point registration, initial registration is performed based on the UAV trajectory information and the pose data provided by the inertial navigation system.

[0047] In the spatial data obtained by the UAV at multiple angles and in different time periods, there is a relative position deviation between each frame of point cloud. In order to integrate the data in a unified coordinate system, the point cloud needs to be registered between frames. Meanwhile, in order to improve the registration accuracy and algorithm efficiency, the approximate pose of each frame of data needs to be obtained before registration as the starting point of registration.

[0048] The system first extracts the flight trajectory and attitude information corresponding to each frame of scanning data from the UAV flight control system and the inertial navigation device, establishes the initial spatial reference position of each frame of point cloud, and then uses a spatial point cloud comparison algorithm to register the point cloud between two or more frames, so that the point cloud is maximally coincident in the three-dimensional space, and a complete and continuous three-dimensional data set is constructed. The whole process is a trade-off between processing accuracy and efficiency to ensure the accuracy of the final model.

[0049] In some embodiments, a registration error threshold ε is set during the registration process, and the iterative closest point algorithm terminates iteration when the following convergence conditions are met:

[0050] N is the total number of point cloud registration point pairs, is the distance between the source point cloud and the coordinate after registration transformation, is the distance between the target point cloud and the nearest neighbor point matched.

[0051] In some embodiments, in step S4, the filtering and noise reduction processing on the registered spatial data includes: performing voxel grid filtering on the target point cloud; After voxel filtering, a statistical outlier rejection algorithm is used to determine whether each point is an outlier according to the average distance between the point and the points in its neighborhood, and the noise points with a distance exceeding a set standard deviation threshold are removed.

[0052] The original point cloud data collected by the UAV may have problems such as uneven sampling density and noisy points, which will affect the modeling quality. Therefore, the point cloud data needs to be structurally optimized and cleaned, including point density normalization and abnormal point elimination.

[0053] Firstly, the system divides the entire point cloud space into several regular cubic cells, samples the data points in each cell, and selects representative points for retention, thereby realizing data compression and uniform distribution. Subsequently, the system analyzes the neighborhood of each point and determines whether it is an isolated abnormal point based on the distance distribution between the point and other points in the neighborhood. For points with a distance significantly deviating from the average distribution, the system identifies them as outliers and eliminates them, thereby improving the coherence and reliability of the point cloud data used for subsequent modeling.

[0054] In some embodiments, in step S5, triangular mesh reconstruction is performed based on the filtered and denoised point cloud data, and a Poisson surface reconstruction algorithm is used to generate a continuous surface model; After completing the point cloud cleaning, in order to further generate a continuous and closed stockpile surface model for volume calculation and visualization analysis, it is necessary to convert the sparse point cloud data into a structured surface mesh model. Poisson reconstruction is a surface patch generation method with high precision and good smoothness, which is suitable for processing unstructured point cloud data.

[0055] When the system reconstructs the processed point cloud data, the Poisson reconstruction method is preferably used. This method calculates a smooth surface covering the entire point set by analyzing the normal vectors and spatial distribution in the point cloud. The completed surface model has good continuity and geometric integrity, can effectively fill the holes caused by occlusion or data loss, and generate a closed three-dimensional stockpile shape for subsequent volume calculation module.

[0056] In step S6, the volume integral algorithm is used to calculate the volume V, which is determined by the following formula:

[0057] wherein, is the area of the i-th triangular mesh on the horizontal projection plane, is the height from the center of the triangular patch to the reference datum plane; The material density p corresponding to the type of the stockpile is obtained, and the weight W of the stockpile is calculated according to the formula:

[0058] In some embodiments, the present application provides a stocktaking system for stockpiles, which comprises: A scanning instruction acquisition module is configured to acquire a scanning task instruction of a stockpile area. ​The UAV flight control module is configured to control the UAV to fly above the stockpile area according to a preset flight path to obtain original spatial data during scanning; The point cloud registration module is configured to perform registration processing on the original spatial data, including initial registration estimation based on UAV trajectory information and pose data provided by an inertial navigation system, and inter-frame point cloud alignment using an iterative closest point algorithm; The point cloud processing module is configured to perform voxel grid filtering and statistical outlier removal processing on the registered spatial data; The three-dimensional modeling module is configured to construct a three-dimensional model of the stockpile based on the processed spatial data, and the modeling module generates a continuous curved surface model using a Poisson surface reconstruction algorithm; The volume calculation and weight estimation module is configured to calculate the volume of the stockpile based on the three-dimensional model using an integral algorithm, and generate corresponding weight information in combination with a preset material density parameter; The data output module is configured to output the volume and weight information of the stockpile for calling by a stockpile management system.

[0059] In some embodiments, the present application provides a terminal, comprising: A memory is configured to store a stockpile program for a stockpile in a stockyard; A processor is configured to execute the stockpile program for the stockpile in the stockyard to implement the steps of the stockpile method for the stockpile in the stockyard.

[0060] In some embodiments, the present application provides a computer readable storage medium storing computer instructions, when a computer reads the computer instructions in the storage medium, the computer executes the stockpile method for the stockpile in the stockyard.

[0061] It can be understood that the system, device, module or unit illustrated in the above embodiments can be specifically implemented by a computer chip or entity, or by a product having certain functions. A typical implementation device is a computer, and the specific form of the computer can be a personal computer, a laptop computer, a personal digital assistant, a tablet computer, a wearable device, or a combination of any of these devices.

[0062] In a typical configuration, the computer includes one or more processors (CPU), input / output interfaces, network interfaces, and memories.

[0063] The memory can include a non-persistent memory in a computer readable medium, a random access memory (RAM), and / or a non-volatile memory such as a read-only memory (ROM) or a flash memory (flash RAM). The memory is an example of a computer readable medium.

[0064] Computer-readable media includes permanent and non-permanent, movable and non-movable media that can implement information storage by any method or technology. Information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette, disk storage, quantum memory, graphene-based storage medium or other magnetic storage device, or any other non-transmission medium that can be used to store information accessible to a computing device. According to the definition herein, computer-readable media does not include transitory media such as modulated data signals and carriers.

[0065] It should also be noted that the terms "comprising," "including," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements recited, but can also include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without more limitations, an element defined by the phrase "comprising a" does not exclude the existence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0066] It should be understood that although one or more embodiments of the specification can employ the terms first, second, third, etc. to describe various information, these information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of one or more embodiments of the specification, the first information can also be referred to as the second information, and similarly, the second information can also be referred to as the first information. Depending on the context, the word "if" as used herein can be interpreted as "when" or "when" or "in response to determining".

[0067] The above description is only the preferred embodiment of one or more embodiments of the specification and does not limit one or more embodiments of the specification. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of one or more embodiments of the specification shall be included in the scope of protection of one or more embodiments of the specification.

Claims

1. A method for inventorying materials in a stockyard, characterized in that, Includes the following steps: Step S1: Obtain the scanning task instruction for the material storage area; Step S2: Control the drone to scan above the material stacking area according to the preset flight path to obtain the raw spatial data during the scanning process; Step S3: Perform registration processing on the original spatial data; Step S4: Filter and denoise the registered spatial data; Step S5: Construct a three-dimensional model of the stockpile based on the processed spatial data; Step S6: Calculate the stockpile volume based on the three-dimensional model and generate weight information by combining it with preset density parameters.

2. The inventory method for stockpiling materials in a material yard according to claim 1, characterized in that, In step S2, the preset flight path includes a scanning trajectory automatically generated based on the material stack boundary profile, which is determined by the following method: Obtain the two-dimensional contour map of the material stack boundary projection, construct the minimum bounding rectangle region using the contour bounding box algorithm, and generate parallel and equally spaced scanning paths according to the direction of the long side of the rectangle; The spacing Δd between parallel scan paths is calculated according to the following formula: Where h is the preset flight altitude of the UAV, θ is the angle of view of the scanning device on the UAV, and k is the sampling overlap coefficient, k>1.

3. The inventory method for stockpiling materials in a material yard according to claim 2, characterized in that, In step S2, when the UAV is performing a scanning path flight, it dynamically adjusts its flight path by sensing the location of environmental obstacles in real time and combining the A* pathfinding algorithm.

4. The inventory method for stockpiling materials in a material yard according to claim 1, characterized in that, In step S3, the registration process for the original spatial data includes inter-frame registration of point cloud data from different scanning perspectives, wherein the inter-frame registration adopts the iterative nearest point algorithm. Before performing iterative nearest point registration, initial registration is performed based on the UAV trajectory information and the pose data provided by the inertial navigation system.

5. The inventory method for stockpiling materials in a material yard according to claim 4, characterized in that, During the registration process, a registration error threshold ε is set, and the iterative nearest point algorithm terminates its iteration when the following convergence condition is met: N is the total number of point cloud registration point pairs. Source Cloud Coordinates after registration transformation In the target point cloud The nearest neighbor that is matched.

6. The inventory method for stockpiling materials in a material yard according to claim 5, characterized in that, In step S4, the filtering and noise reduction processing of the registered spatial data includes: Perform voxel mesh filtering on the target point cloud; After voxel filtering, a statistical outlier removal algorithm is used to determine whether each point is an outlier based on the average distance between each point and its neighboring points, and to remove noise points whose distance exceeds a set standard deviation threshold.

7. The inventory method for stockpiling materials in a material yard according to claim 6, characterized in that, In step S5, triangular mesh reconstruction is performed based on the point cloud data after filtering and noise reduction, and a continuous surface model is generated using the Poisson surface reconstruction algorithm. In step S6, the volume V is calculated using a volume integration algorithm, and the volume is determined by the following formula: in, Let be the area of ​​the i-th triangular mesh on the horizontal projection plane. The height of the centroid of the triangular facet from the reference plane; To obtain the material density ρ corresponding to the stockpile type, use the formula: Calculate the weight W of the stockpile.

8. An inventory system for stockpiling materials in a material yard, characterized in that, For implementing the inventory method for stockpiling materials in a material yard as described in claim 1, the system comprises: The scan instruction acquisition module is used to acquire scan task instructions for the stockpiling area of ​​the material yard; The drone flight control module is used to control the drone to scan above the material stacking area according to a preset flight path and acquire raw spatial data during the scanning process. The point cloud registration module is used to perform registration processing on the original spatial data, including initial registration estimation based on UAV trajectory information and pose data provided by the inertial navigation system, and inter-frame point cloud alignment using an iterative nearest point algorithm. The point cloud processing module is used to perform voxel grid filtering and statistical outlier removal on the registered spatial data. The 3D modeling module is used to construct a 3D model of the stockpile based on the processed spatial data. The modeling module uses the Poisson surface reconstruction algorithm to generate a continuous surface model. The volume calculation and weight estimation module is used to calculate the stockpile volume based on the three-dimensional model using an integral algorithm, and generate corresponding weight information by combining preset material density parameters. The data output module is used to output the volume and weight information of the stockpiled materials for use by the inventory management system.

9. A terminal, characterized in that, include: Memory, used to store inventory procedures for stockpiling materials in the material yard; A processor is configured to implement the steps of the inventory method for stockpiling as described in claim 1 when executing the inventory system for stockpiling in a stockyard.

10. A computer-readable storage medium, characterized in that, The storage medium stores computer instructions. When the computer reads the computer instructions from the storage medium, the computer executes the inventory method for stockpiling materials in the material yard as described in claim 1.

Images