Multi-scene adaptive bulk cargo yard intelligent planning system

Abstract

The present application provides a kind of multi-scene adaptive bulk cargo yard intelligent planning system, extracts the cargo type parameter from the collected image information automatically by data processing module, and accurately divides and morphologically analyzes the stacking by combining three-dimensional point cloud data, realizes the automatic identification and quantification of bulk cargo physical properties, and the planning engine module can dynamically generate the space layout and equipment path scheme matched with the current yard type and operation mode by fusing the automatically obtained cargo type attribute, stacking form parameter, yard topography and preset multi-dimensional scene configuration information, thereby solving the defects of insufficient pertinence of general planning method, realizing the whole process adaptation and automation from data acquisition to planning generation, and improving the planning accuracy, operation efficiency and management intelligent level of bulk cargo yard in different business scenarios.

Description

Technical Field

[0001] This invention relates to the field of warehousing and industrial automation technology, and in particular to a multi-scenario adaptive intelligent planning system for bulk cargo yards. Background Technology

[0002] With the continuous expansion of the logistics industry and industrial production scale, bulk cargo yards, as an important component of ports, mines, power plants, and large logistics hubs, directly impact operating costs and operational safety through their management efficiency and level of operational intelligence. Traditional yard management relies heavily on manual inspections, paper records, and experience-based judgment, which is no longer sufficient to meet the demands of real-time, accurate, and efficient modern management. Therefore, intelligent yard management systems based on drones, LiDAR, and 3D modeling technologies are gradually becoming an industry trend.

[0003] Currently, by using drones equipped with LiDAR to perform 3D scanning of the stockyard, and processing point cloud data to calculate the stack volume and digitally model the stockyard, the system can output the basic outline and volume information of the stacks and display the 3D scene of the stockyard in a visual interface. This type of system improves the automation level of data acquisition to a certain extent, replacing traditional manual measurement methods.

[0004] However, existing technologies typically rely on manual pre-input or configuration of cargo types and their physical parameters, failing to automatically identify and obtain cargo attributes from collected data. This results in insufficient adaptability, cumbersome configuration, and difficulty in coping with scenarios where cargo types change frequently during actual deployment. On the other hand, existing solutions are mostly designed for single or fixed scenarios, lacking the ability to adaptively plan based on multi-dimensional scenario information such as yard type and operation mode. The resulting layout and path schemes are often highly general but lack specificity, making it difficult to achieve optimal storage efficiency, operational smoothness, and safety compliance in different business scenarios. Summary of the Invention

[0005] This invention provides a multi-scenario adaptive intelligent planning system for bulk cargo yards, which solves the problem of low processing efficiency in existing bulk cargo yard planning when dealing with frequent changes in cargo types.

[0006] On the one hand, the present invention provides a multi-scenario adaptive intelligent planning system for bulk cargo yards, including: a data acquisition module, a data processing module, and a planning engine module;

[0007] The data acquisition module is configured to acquire three-dimensional point cloud data and image information of the bulk cargo yard;

[0008] The data processing module is configured to: acquire the three-dimensional point cloud data and the image information from the data acquisition module, and separate ground point cloud data and non-ground point cloud data from the three-dimensional point cloud data, and acquire cargo type parameters from the image information;

[0009] The non-ground point cloud data is stacked and segmented to obtain at least one single-stack point cloud dataset;

[0010] Based on the single-stack point cloud dataset and the cargo type parameters, the morphological parameters of the corresponding stack are determined;

[0011] The planning engine module is configured to generate a planning scheme that includes spatial layout and operation path based on preset scene configuration information, combined with the yard terrain represented by the ground point cloud data and the morphological parameters.

[0012] Optionally, the data acquisition module includes:

[0013] The drone equipped with lidar is configured to collect three-dimensional point cloud data and image information of bulk cargo yards;

[0014] The dynamic GNSS unit is configured to provide positioning data to the UAV in order to control the UAV to move along a preset route.

[0015] The 5G communication unit is configured to transmit the three-dimensional point cloud data and the image information to the data processing module.

[0016] Optionally, the data processing module includes a ground separation unit;

[0017] The ground separation unit is configured as follows:

[0018] Initial ground seed points are selected from the three-dimensional point cloud data, and an initial ground model is constructed based on the initial ground seed points;

[0019] Based on the initial ground model, calculate the vertical height difference between each point in the three-dimensional point cloud data and the initial ground model;

[0020] Based on a preset height difference threshold, points with a vertical height difference less than the height difference threshold are classified as candidate ground points, and points with a vertical height difference greater than the height difference threshold are classified as candidate non-ground points.

[0021] Based on the candidate ground points, the initial ground model is optimized and updated using a surface fitting method;

[0022] Repeat the process until the iteration termination condition is met, and output the candidate ground points and candidate non-ground points obtained in the final iteration as the ground point cloud data and the non-ground point cloud data, respectively.

[0023] Optionally, the data processing module includes a stacking segmentation unit;

[0024] The stacking segmentation unit, connected to the ground separation unit, is configured as follows:

[0025] Receive the non-ground point cloud data output by the ground separation unit;

[0026] The non-ground point cloud data is projected onto a two-dimensional plane, and a corresponding elevation image is generated based on the point cloud height information;

[0027] In the elevation image, local elevation minima are identified as initial seed points for the watershed algorithm;

[0028] The process of water flow spreading from the initial seed point to the surrounding area is simulated, and the watershed boundary is determined based on the elevation gradient.

[0029] Based on the watershed boundary, the non-ground point cloud data is divided into multiple disconnected point cloud subsets, each of which corresponds to a single stack, and a single stack point cloud dataset is output.

[0030] Optionally, the data processing module includes a parameter calculation unit;

[0031] The parameter calculation unit is connected to the stacking segmentation unit and is configured as follows:

[0032] Receive the single-stacking point cloud dataset output by the stacking segmentation unit;

[0033] Obtain the physical characteristic parameters of the goods, wherein the physical characteristic parameters include at least one of the angle of repose, density and angle of repose;

[0034] For each single stack point cloud dataset, based on the three-dimensional spatial distribution and combined with the stacking angle and natural repose angle in the physical characteristic parameters, the three-dimensional surface morphology model of the stack is calculated.

[0035] Based on the surface three-dimensional morphology model and the bottom surface of the storage yard represented by the ground point cloud data, the volume of space occupied by the stack is calculated.

[0036] The estimated weight of the stack is calculated based on the occupied space volume and the density in the physical property parameters.

[0037] The output includes the morphological parameters of the surface's three-dimensional morphological model, the volume it occupies, and the estimated weight.

[0038] Optionally, the planning engine module includes a scene recognition unit;

[0039] The scene recognition unit is configured as follows:

[0040] Receive user-inputted scenario raw parameters, which include at least one of yard type, cargo type attribute and operation mode;

[0041] The original parameters of the scene are standardized and encoded to generate a scene feature vector with a uniform format;

[0042] Based on predefined scene classification rules, the scene feature vectors are mapped to at least one standardized scene configuration label;

[0043] The mapped scene configuration tags are combined to output structured scene configuration information.

[0044] Optionally, the planning engine module includes a strategy matching unit;

[0045] The strategy matching unit is connected to the scene recognition unit and is configured as follows:

[0046] Receive the scene configuration information output by the scene recognition unit;

[0047] Access the pre-stored planning strategy knowledge base, which stores the mapping relationship between different scenario configuration information and planning strategy rule set;

[0048] Using the scenario configuration information as the query condition, a search and matching process is performed in the planning strategy knowledge base to obtain a set of planning strategy rules associated with the scenario configuration information; the set of planning strategy rules includes stacking strategy rules, path planning rules, and safety and environmental protection rules.

[0049] Output the set of planning strategy rules obtained through matching.

[0050] Optionally, the planning engine module further includes a scheme generation unit;

[0051] The scheme generation unit is connected to the strategy matching unit and the data processing module, and is configured as follows:

[0052] The system receives the set of planning strategy rules output by the strategy matching unit, as well as the yard terrain and morphological parameters represented by the ground point cloud data output by the data processing module.

[0053] Based on the set of planning strategy rules, combined with the yard topography and the morphological parameters, an analysis and adaptability assessment of the yard's idle areas are conducted to obtain the adaptability assessment results.

[0054] Based on the adaptive assessment results and the set of planning strategy rules, the spatial layout scheme of the stacking is adjusted;

[0055] Based on the spatial layout scheme, the morphological parameters of the stacking, and the path planning rules, a device operation path is generated;

[0056] The planning scheme is generated by integrating the spatial layout scheme with the equipment operation path.

[0057] Optionally, an output module may also be included;

[0058] The output module is configured as follows:

[0059] Receive the planning scheme;

[0060] The spatial layout scheme in the planning scheme is converted into a visual yard layout map, and the stacking location, outline, number and key passage information are marked on the yard layout map;

[0061] The equipment operation path in the planning scheme is converted into a sequence of instructions that the equipment can recognize and execute; the instruction sequence includes action instructions, coordinate point sequence and operation parameters.

[0062] Optionally, it also includes an adaptive optimization module; the adaptive optimization module is connected to the planning engine module;

[0063] The adaptive optimization module is configured as follows:

[0064] Collect historical operation data, perform statistical analysis on the set of planning strategy rules in the planning strategy knowledge base based on the historical operation data, calculate the deviation between operation efficiency and planning expectations, and generate deviation analysis results;

[0065] Based on the deviation analysis results, the thresholds, weights, or constraints in the planning strategy rule set are adjusted.

[0066] The multi-scenario adaptive bulk cargo yard intelligent planning system provided by this invention automatically extracts cargo type parameters from the collected image information through the data processing module, and performs precise segmentation and morphological analysis of the stacks by combining 3D point cloud data. This achieves automatic identification and quantification of the physical attributes of bulk cargo. The planning engine module can dynamically generate spatial layout and equipment path schemes that match the current yard type and operation mode by integrating automatically acquired cargo type attributes, stack morphological parameters, yard terrain and preset multi-dimensional scenario configuration information. This solves the deficiency of general planning methods in terms of specificity, and realizes full-process adaptation and automation from data collection to planning generation. It improves the planning accuracy, operation efficiency and management intelligence level of bulk cargo yards in different business scenarios. Attached Figure Description

[0067] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0068] Figure 1 This is one of the structural schematic diagrams of the multi-scenario adaptive bulk cargo yard intelligent planning system provided in this embodiment of the invention;

[0069] Figure 2 This is the second schematic diagram of the structure of the multi-scenario adaptive bulk cargo yard intelligent planning system provided in this embodiment of the invention;

[0070] Figure 3 This is a schematic diagram of the data acquisition module structure provided in an embodiment of the present invention;

[0071] Figure 4 This is a schematic diagram of the data processing module structure provided in an embodiment of the present invention;

[0072] Figure 5 This is a schematic diagram of the planning engine module structure provided in an embodiment of the present invention. Detailed Implementation

[0073] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0074] The following is combined with Figures 1-2 This invention describes a multi-scenario adaptive intelligent planning system for bulk cargo yards provided by an embodiment of the present invention.

[0075] Figure 1 This is one of the structural schematic diagrams of the multi-scenario adaptive bulk cargo yard intelligent planning system provided in the embodiments of the present invention.

[0076] like Figure 1 As shown, the multi-scenario adaptive bulk cargo yard intelligent planning system provided in this embodiment of the invention includes:

[0077] The system includes a data acquisition module 110, a data processing module 120, and a planning engine module 130.

[0078] The data acquisition module 110 is configured to acquire three-dimensional point cloud data and image information of the bulk cargo yard.

[0079] The 3D point cloud data reflects the topography and cargo stacking patterns of the bulk cargo yard. This data can be acquired using LiDAR. The 3D point cloud data contains a large number of point coordinates, each representing a specific location on the yard surface, thus constructing a 3D model of the yard. Image information records the bulk cargo yard from a visual perspective; this image information can be ordinary color images or infrared images. Ordinary color images can visually display the appearance of the yard, the color and texture features of the cargo. Infrared images can detect the temperature distribution of the cargo. The data acquisition module 110 can send the 3D point cloud data and image information to the data processing module 120 in real time, rather than sending them only after the data acquisition is complete.

[0080] The data processing module 120 is configured to: acquire three-dimensional point cloud data and image information from the data acquisition module 110, separate ground point cloud data and non-ground point cloud data from the three-dimensional point cloud data, and acquire cargo type parameters from the image information.

[0081] Stacking and segmenting of non-ground point cloud data yields at least one single-stack point cloud dataset.

[0082] Based on the single-stack point cloud dataset and cargo type parameters, the morphological parameters of the corresponding stack are determined.

[0083] Specifically, the data processing unit is configured to acquire 3D point cloud data containing stack space coordinates and height correlation information, as well as image information containing cargo appearance and texture features, from the data acquisition unit. Based on a preset ground elevation threshold, point cloud density features, and spatial distribution patterns, the 3D point cloud data is classified to separate ground point cloud data from the corresponding stack's non-ground point cloud data. Simultaneously, the image feature recognition algorithm is used to analyze the cargo appearance texture and color distribution features in the image information. The physical characteristic parameters of the corresponding cargo are extracted from a preset cargo feature database as cargo parameters. Then, a watershed boundary algorithm combined with point cloud spatial clustering rules is used to perform regional division and boundary definition processing on the non-ground point cloud data, dividing the continuously associated point cloud clusters into independent units to obtain at least one single-stack point cloud dataset. Each single-stack point cloud dataset corresponds to an independent stack. Finally, coordinate fitting and spatial modeling processing are performed on each single-stack point cloud dataset. Combined with the extracted cargo parameters, the length, width, height, volume, and stack outline boundary and other morphological parameters of the corresponding stack are calculated and determined.

[0084] The planning engine module 130 is configured to generate a planning scheme that includes spatial layout and operation path based on preset scene configuration information and combined with the yard terrain and morphological parameters represented by ground point cloud data.

[0085] The planning engine module 130 is configured to acquire preset scene configuration information, including stacking priority rules, cargo type zoning compatibility requirements, safety distance standards, operating equipment parameters, and cargo flow constraints. It also receives ground point cloud data and stacking morphology parameters output by the data processing unit. The ground point cloud data represents the elevation distribution, flatness, and obstacle location information of the yard terrain, while the morphology parameters include the length, width, height, volume, and outline boundary data of each stack.

[0086] Subsequently, a 3D terrain model of the yard is constructed based on ground point cloud data. Usable and restricted areas are identified through terrain elevation analysis. Combined with safety distance standards and cargo type zoning requirements in the scene configuration information, the usable areas are initially divided. Next, the morphological parameters of each stack are analyzed for compatibility with the divided usable areas. Based on storage priority rules and cargo flow constraints, the precise spatial coordinates, placement orientation, and spacing between adjacent stacks are determined, forming a yard spatial layout scheme. Finally, combining the operating equipment parameters in the scene configuration information, a path planning algorithm is used to calculate the path from the starting point of the operation, the operating points of each stack, and the ending point of the operation, avoiding terrain restrictions and obstacles, and planning a collision-free, highly efficient optimal operating path. Finally, the above spatial layout and operating path are integrated to generate a complete planning scheme.

[0087] In some embodiments, such as Figure 3 As shown, the data acquisition module 110 includes:

[0088] The UAV 1101, equipped with a lidar, is configured to collect three-dimensional point cloud data and image information of bulk cargo yards.

[0089] Among them, the UAV 1101, equipped with lidar, flies to the pre-set monitoring area of ​​the bulk cargo yard. It emits laser signals into the yard area and receives reflected signals through lidar, converting them into three-dimensional point cloud data containing the spatial coordinates and height information of the stack. At the same time, it captures image information of the yard and the stack through the airborne image acquisition component, and records the acquisition timestamps of the three-dimensional point cloud data and image information simultaneously.

[0090] The dynamic GNSS unit 1102 is configured to provide positioning data to the UAV 1101 in order to control the UAV 1101 to move along a preset route.

[0091] Among them, the dynamic GNSS unit 1102 can receive satellite positioning signals in real time, calculate and generate the real-time position coordinates, attitude data and movement trajectory information of the UAV 1101, and feed the positioning data back to the control module of the UAV 1101 in real time to support the UAV 1101 to move stably according to the preset planned cruise route, altitude and operation sequence, and ensure the spatial coverage integrity of data collection.

[0092] The 5G communication unit 1103 is configured to transmit 3D point cloud data and image information to the data processing module 120.

[0093] The 5G communication unit 1103 establishes a wireless communication link with the data processing module 120, receives 3D point cloud data and image information output by the UAV 1101 equipped with lidar, and adopts a simultaneous acquisition and transmission mode to synchronously transmit the 3D point cloud data with timestamps, image information and real-time status data of the UAV 1101 to the data processing module 120, ensuring the real-time and continuous transmission of data.

[0094] In some embodiments, such as Figure 4 As shown, the data processing module 120 includes a ground separation unit 1201;

[0095] Ground separation unit 1201 is configured as follows:

[0096] Initial ground seed points are selected from the 3D point cloud data, and an initial ground model is constructed based on the initial ground seed points;

[0097] Based on the initial ground model, calculate the vertical height difference between each point in the 3D point cloud data and the initial ground model;

[0098] Based on a preset height difference threshold, points with a vertical height difference less than the height difference threshold are classified as candidate ground points, and points with a vertical height difference greater than the height difference threshold are classified as candidate non-ground points.

[0099] Based on candidate ground points, the initial ground model is optimized and updated using a surface fitting method;

[0100] Repeat the process until the iteration termination condition is met, and output the candidate ground points and candidate non-ground points obtained in the final iteration as ground point cloud data and non-ground point cloud data, respectively.

[0101] Specifically, the ground separation unit 1201 selects representative initial ground seed points from the three-dimensional point cloud data transmitted by the data acquisition module 110 based on the point cloud elevation distribution characteristics and density distribution law, and builds an initial ground model through triangulation network construction or plane fitting method.

[0102] Subsequently, based on the initial ground model, the vertical height difference between each point in the 3D point cloud data and the initial ground model is calculated to clarify the spatial positional relationship of each point relative to the initial ground. According to the preset height difference threshold, points with a calculated vertical height difference less than the threshold are identified as candidate ground points, and points with a vertical height difference greater than the threshold are identified as candidate non-ground points. Then, based on the selected candidate ground points, the initial ground model is optimized, adjusted, and updated using the least squares method or irregular triangular mesh surface fitting method to improve the adaptability of the ground model to the actual yard terrain.

[0103] Repeat the above process of calculating the vertical height difference of each point, classifying candidate ground points and candidate non-ground points, and optimizing and updating the ground model until the preset iteration termination conditions are met, such as the elevation change of the ground model being less than a set threshold after two consecutive iterations, the number of iterations reaching a preset upper limit, or the number of candidate ground points stabilizing. Finally, the candidate ground points obtained after the iteration termination are output as ground point cloud data, and the candidate non-ground points are output as non-ground point cloud data.

[0104] In some embodiments, such as Figure 4 As shown, the data processing module 120 includes a stacking segmentation unit 1202;

[0105] The stacking segmentation unit 1202, connected to the ground separation unit 1201, is configured as follows:

[0106] Receive non-ground point cloud data output by ground separation unit 1201;

[0107] Non-ground point cloud data is projected onto a two-dimensional plane, and a corresponding elevation image is generated based on the point cloud height information;

[0108] In the elevation image, the local elevation minimum region is identified as the initial seed point for the watershed algorithm;

[0109] Simulate the process of water flow spreading from the initial seed point to the surrounding area, and determine the watershed boundary based on the elevation gradient;

[0110] Based on the watershed boundary, the non-ground point cloud data is divided into multiple disconnected point cloud subsets, each of which corresponds to a single stack, and the single stack point cloud dataset is output.

[0111] Specifically, the stacking segmentation unit first receives the non-ground point cloud data output by the ground separation unit 1201, projects the non-ground point cloud data onto a preset two-dimensional plane according to the horizontal coordinates, and maps the height information of the point cloud corresponding to each projection point to the pixel value of the corresponding position on the two-dimensional plane to generate an elevation image representing the elevation distribution of the stack.

[0112] Subsequently, in the elevation image, through neighborhood window traversal analysis, elevation value sorting and comparison, and local extremum detection algorithms, continuous regions with significantly lower elevation values ​​relative to the surrounding areas are identified, and these significantly lower continuous regions are determined as the initial seed points for the watershed algorithm.

[0113] Next, the elevation gradient is calculated based on the elevation values ​​of each pixel in the elevation image to simulate the process of water flow gradually spreading from the initial seed point to the surrounding areas with increasing elevation. When the diffusion areas corresponding to different initial seed points meet along the elevation gradient, the pixel boundary at the meeting point is determined as the watershed boundary.

[0114] Finally, based on the watershed boundary, the original non-ground point cloud data is back-correlated, and the point cloud within the same watershed boundary range is divided into multiple unconnected point cloud subsets. Each point cloud subset corresponds to a single stack in the bulk cargo yard, and then the single stack point cloud dataset corresponding to each independent stack is output.

[0115] In some embodiments, such as Figure 4 As shown, the data processing module 120 includes a parameter calculation unit 1203;

[0116] The parameter calculation unit 1203 is connected to the stacking segmentation unit 1202 and is configured as follows:

[0117] Receive the single-stack point cloud dataset output by the stacking segmentation unit 1202;

[0118] Obtain the physical property parameters of the cargo, including at least one of the angle of repose, density, and angle of repose.

[0119] For each single stack point cloud dataset, based on the three-dimensional spatial distribution and combined with the stacking angle and natural repose angle in the physical characteristic parameters, the three-dimensional surface morphology model of the stack is calculated.

[0120] The volume of space occupied by the stack is calculated based on the bottom surface of the storage yard represented by the three-dimensional surface morphology model and ground point cloud data.

[0121] Calculate the estimated weight of the stack based on the occupied space volume and density in the physical property parameters;

[0122] The output includes morphological parameters such as the surface 3D morphological model, the volume occupied, and the estimated weight.

[0123] Specifically, the parameter calculation unit 1203 receives the single-stack point cloud dataset corresponding to each independent stack output by the stacking segmentation unit 1202, and obtains the physical characteristic parameters of the corresponding cargo by querying the preset cargo physical characteristic database or receiving external input. The physical characteristic parameters include at least one or more of the following: angle of stacking, density, and angle of repose.

[0124] Subsequently, for each single-stack point cloud dataset, the three-dimensional coordinates of the three-dimensional point cloud data included in the single-stack point cloud dataset are extracted and the spatial distribution characteristics are analyzed. The stacking angle and natural repose angle in the obtained physical characteristic parameters are used as morphological constraints. Through three-dimensional coordinate fitting, surface reconstruction algorithm and morphological boundary constraint processing, a surface three-dimensional morphological model that can represent the actual appearance contour of the stack is constructed.

[0125] Next, using the actual bottom surface of the storage yard represented by the ground point cloud data output by the ground separation unit 1201 as the reference surface, the actual space volume occupied by the stack in the storage yard is calculated by spatial polygon mesh division, spatial integration of the area enclosed by the three-dimensional model and the reference surface, or three-dimensional modeling volume measurement algorithm. Then, the calculated occupied space volume is multiplied by the cargo density in the physical characteristic parameters, and combined with the preset volume-weight conversion coefficient, the estimated weight of the stack is obtained.

[0126] The final output includes a complete morphological parameter model of the stack's surface three-dimensional shape, the volume it occupies, and the estimated weight.

[0127] In some embodiments, such as Figure 5 As shown, the planning engine module 130 includes a scene recognition unit 1301;

[0128] The scene recognition unit 1301 is configured as follows:

[0129] Receive user input of original scenario parameters, which include at least one of the following: yard type, cargo type attribute, and operation mode;

[0130] The original parameters of the scene are standardized and encoded to generate scene feature vectors with a uniform format;

[0131] Based on predefined scene classification rules, scene feature vectors are mapped to at least one standardized scene configuration label;

[0132] The mapped scene configuration tags are combined to output structured scene configuration information.

[0133] Specifically, the scene recognition unit 1301 receives the original scene parameters input by the user through the interactive interface. The original scene parameters include at least one or more of the following: yard type, cargo type attribute, and operation mode. For example, the yard type is a dry bulk cargo yard or a mixed-use yard; the cargo type attribute is a specific cargo type identifier such as coal or ore; and the operation mode is any one of the following: continuous storage mode, batch turnover mode, or priority outbound mode.

[0134] Subsequently, according to the preset parameter encoding rules, the original parameters of different types of scenarios are split into fields, converted into data formats, and standardized and mapped to generate scenario feature vectors with unified field structure and standardized data format.

[0135] Next, the pre-stored scene classification rule library is invoked. The scene classification rule library contains a preset correspondence between scene feature vectors and standardized scene configuration labels. The generated scene feature vectors are compared with the classification conditions in the scene classification rule library through a feature matching algorithm, and then mapped to at least one standardized scene configuration label.

[0136] Finally, all the mapped scenario configuration tags are logically integrated and structurally encapsulated, and combined according to the preset data framework to form structured scenario configuration information containing dimensions such as scenario type identifier, cargo type adaptation requirements, and operational constraints.

[0137] In some embodiments, such as Figure 5 As shown, the planning engine module 130 includes a strategy matching unit 1302;

[0138] The strategy matching unit 1302 is connected to the scene recognition unit 1301 and is configured as follows:

[0139] Receive scene configuration information output by scene recognition unit 1301;

[0140] Access the pre-stored planning strategy knowledge base.

[0141] The planning strategy knowledge base stores the mapping relationship between different scenario configuration information and the set of planning strategy rules;

[0142] Using scenario configuration information as the query condition, the system searches and matches the planning strategy knowledge base to obtain a set of planning strategy rules associated with the scenario configuration information. The set of planning strategy rules includes storage strategy rules, path planning rules, and safety and environmental protection rules.

[0143] Output the set of planning strategy rules obtained from the matching.

[0144] Specifically, after receiving the scenario configuration information, the strategy matching unit 1302 accesses the pre-stored planning strategy knowledge base. This knowledge base pre-establishes and stores a one-to-one mapping relationship between different structured scenario configuration information and corresponding planning strategy rule sets. Each planning strategy rule set contains stacking strategy rules adapted to a specific scenario. These stacking strategy rules include stack allocation priority, stacking height limits, zoned stacking specifications, operating equipment travel route constraints, turning radius adaptation standards, path efficiency optimization criteria, stacking spacing safety thresholds, dust control operation requirements, and equipment operation safety range definitions.

[0145] Next, using the received scenario configuration information as the complete query condition, the rule matching algorithm is used to compare the mapping relationship between each dimension of the scenario configuration information and the planning strategy knowledge base field by field, perform feature association retrieval and full-dimensional adaptation analysis, and filter out the planning strategy rule set that is completely matched or optimally adapted to the scenario configuration information.

[0146] Finally, the complete set of planning strategy rules, including heaping strategy rules, path planning rules, and safety and environmental protection rules, will be output.

[0147] In some embodiments, such as Figure 5 As shown, the planning engine module 130 also includes a scheme generation unit 1303;

[0148] The scheme generation unit 1303 is connected to the strategy matching unit 1302 and the data processing module 120, and is configured as follows:

[0149] The receiving strategy matching unit 1302 outputs a set of planning strategy rules, and the data processing module 120 outputs ground point cloud data representing the yard terrain and morphological parameters.

[0150] Based on the set of planning strategy rules and combined with the topography and morphological parameters of the storage yard, an analysis and adaptability assessment of the storage yard's idle areas were conducted to obtain the adaptability assessment results.

[0151] Based on the adaptive assessment results and the set of planning strategy rules, the spatial layout scheme of the stacking is adjusted;

[0152] Based on the spatial layout plan, stacking morphology parameters, and path planning rules, generate equipment operation paths;

[0153] The planning scheme is generated by integrating spatial layout schemes and equipment operation paths.

[0154] Specifically, the scheme generation unit 1303 receives a set of planning strategy rules, and at the same time receives the yard terrain represented by ground point cloud data output by the data processing module 120, as well as the morphological parameters including the three-dimensional morphological model of each stack surface, the occupied space volume, and the estimated weight.

[0155] Subsequently, based on the safety distance standards, cargo type zoning compatibility requirements, and storage priority rules in the planning strategy rule set, combined with the elevation data of the yard terrain and the available area range, the currently unoccupied vacant areas of the yard are spatially divided, dimensionally calculated, and load-bearing conditions analyzed. At the same time, the morphological parameters of each stack are referenced to evaluate the suitability of the vacant areas and the stacks, resulting in a complete adaptability assessment.

[0156] Based on the results of the adaptive assessment, and combined with the stacking strategy rules in the planning strategy rule set, the preset spatial coordinates, placement orientation, stacking height limit and spacing between adjacent stacks of each stack are adjusted to form an optimized spatial layout scheme that takes into account both space utilization and operational convenience.

[0157] Next, based on the precise location, outline boundary, and morphological parameters of each stack in the final spatial layout scheme, and combined with the path planning rules in the planning strategy rule set, a path search algorithm is used to calculate the coherent path of the starting point of the operating equipment, the corresponding operating point of each stack, and the operating endpoint. This avoids terrain-restricted areas, obstacles, and planned stacking areas, generating a collision-free and highly efficient equipment operating path.

[0158] Finally, the optimized spatial layout scheme is integrated with the planned equipment operation path in a structured manner, and operation sequence guidance and equipment scheduling priority are added to generate and output a planning scheme that includes spatial layout details, operation path coordinate sequence and operation constraint description.

[0159] In some embodiments, such as Figure 2 As shown, the multi-scenario adaptive bulk cargo yard intelligent planning system provided in this embodiment of the invention also includes an output module 140;

[0160] Output module 140 is configured as follows:

[0161] Receive the planning proposal;

[0162] The spatial layout scheme in the planning scheme is converted into a visual yard layout map, and the stacking location, outline, number and key passage information are marked on the yard layout map;

[0163] The equipment operation path in the planning scheme is converted into a sequence of instructions that the equipment can recognize and execute; the instruction sequence includes action instructions, coordinate point sequence and operation parameters.

[0164] Specifically, the action commands specify the specific operations that the equipment needs to perform at different stages, such as forward, backward, turning, lifting, and lowering; the coordinate point sequence accurately provides the equipment's movement trajectory in the yard, ensuring that the equipment can accurately reach each work point; the operation parameters specify various technical indicators of the equipment during the operation process, such as speed, force, and time. The output module 140 optimizes and adjusts the command sequence according to the type and characteristics of the equipment to ensure the compatibility and adaptability of the command sequence with the equipment, enabling the equipment to execute the planning scheme efficiently and stably.

[0165] By providing operators and equipment with a visualized yard layout diagram and equipment-recognizable instruction sequences, the output module 140 provides intuitive and convenient operational guidelines for the intelligent operation of bulk cargo yards, which helps to improve the operational efficiency and management level of the yard, and reduce human error and resource waste.

[0166] In some embodiments, such as Figure 2 As shown, the multi-scenario adaptive bulk cargo yard intelligent planning system provided in this embodiment of the invention also includes an adaptive optimization module 150; the adaptive optimization module 150 is connected to the planning engine module 130;

[0167] The adaptive optimization module 150 is configured as follows:

[0168] Collect historical operation data, perform statistical analysis on the set of planning strategy rules in the planning strategy knowledge base based on the historical operation data, calculate the deviation between operation efficiency and planning expectations, and generate deviation analysis results;

[0169] Based on the results of the deviation analysis, the thresholds, weights, or constraints in the planning strategy rule set are adjusted.

[0170] The adaptive optimization module 150 periodically reviews the deviation analysis results. If the deviation analysis results show that the operating efficiency is consistently lower than the planned expectations for a certain period of time, and the deviation value exceeds the preset critical value, the adaptive optimization module 150 will mine historical operating data and, in combination with the yard terrain, stacking morphology parameters, and equipment operating paths, find out the specific reasons for the low operating efficiency.

[0171] For example, it might be discovered that an unreasonable layout of a certain type of stacking is causing the equipment to frequently adjust its path during operation, increasing operation time. In this case, the adaptive optimization module 150 will adjust the planning strategy rule set based on the analysis results. It may reduce the priority of this type of stacking in the spatial layout scheme or reset the layout rules to improve the equipment's operating efficiency.

[0172] Meanwhile, the adaptive optimization module 150 also takes into account the dynamic changes in the yard terrain. As stacks are continuously added, removed, and adjusted, the yard terrain changes, affecting the operating paths and efficiency of the equipment. The adaptive optimization module 150 collects information on changes in the yard terrain and incorporates this information into the deviation analysis and adjustment process of the planning strategy rule set.

[0173] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A multi-scenario adaptive intelligent planning system for bulk cargo yards, characterized in that, include: Data acquisition module, data processing module, and planning engine module; The data acquisition module is configured to acquire three-dimensional point cloud data and image information of the bulk cargo yard; The data processing module is configured to: acquire the three-dimensional point cloud data and the image information from the data acquisition module, and separate ground point cloud data and non-ground point cloud data from the three-dimensional point cloud data, and acquire cargo type parameters from the image information; The non-ground point cloud data is stacked and segmented to obtain at least one single-stack point cloud dataset; Based on the single-stack point cloud dataset and the cargo type parameters, the morphological parameters of the corresponding stack are determined; The planning engine module is configured to generate a planning scheme that includes spatial layout and operation path based on preset scene configuration information, combined with the yard terrain represented by the ground point cloud data and the morphological parameters.

2. The multi-scenario adaptive bulk cargo yard intelligent planning system according to claim 1, characterized in that, The data acquisition module includes: The drone equipped with lidar is configured to collect three-dimensional point cloud data and image information of bulk cargo yards; The dynamic GNSS unit is configured to provide positioning data to the UAV in order to control the UAV to move along a preset route. The 5G communication unit is configured to transmit the three-dimensional point cloud data and the image information to the data processing module.

3. The multi-scenario adaptive bulk cargo yard intelligent planning system according to claim 1, characterized in that, The data processing module includes a ground separation unit; The ground separation unit is configured as follows: Initial ground seed points are selected from the three-dimensional point cloud data, and an initial ground model is constructed based on the initial ground seed points; Based on the initial ground model, calculate the vertical height difference between each point in the three-dimensional point cloud data and the initial ground model; Based on a preset height difference threshold, points with a vertical height difference less than the height difference threshold are classified as candidate ground points, and points with a vertical height difference greater than the height difference threshold are classified as candidate non-ground points. Based on the candidate ground points, the initial ground model is optimized and updated using a surface fitting method; Repeat the process until the iteration termination condition is met, and output the candidate ground points and candidate non-ground points obtained in the final iteration as the ground point cloud data and the non-ground point cloud data, respectively.

4. The multi-scenario adaptive bulk cargo yard intelligent planning system according to claim 3, characterized in that, The data processing module includes a stacking segmentation unit; The stacking segmentation unit, connected to the ground separation unit, is configured as follows: Receive the non-ground point cloud data output by the ground separation unit; The non-ground point cloud data is projected onto a two-dimensional plane, and a corresponding elevation image is generated based on the point cloud height information; In the elevation image, local elevation minima are identified as initial seed points for the watershed algorithm; The process of water flow spreading from the initial seed point to the surrounding area is simulated, and the watershed boundary is determined based on the elevation gradient. Based on the watershed boundary, the non-ground point cloud data is divided into multiple disconnected point cloud subsets, each of which corresponds to a single stack, and a single stack point cloud dataset is output.

5. The multi-scenario adaptive bulk cargo yard intelligent planning system according to claim 4, characterized in that, The data processing module includes a parameter calculation unit; The parameter calculation unit is connected to the stacking segmentation unit and is configured as follows: Receive the single-stacking point cloud dataset output by the stacking segmentation unit; Obtain the physical characteristic parameters of the goods, wherein the physical characteristic parameters include at least one of the angle of repose, density and angle of repose; For each single stack point cloud dataset, based on the three-dimensional spatial distribution and combined with the stacking angle and natural repose angle in the physical characteristic parameters, the three-dimensional surface morphology model of the stack is calculated. Based on the surface three-dimensional morphology model and the bottom surface of the storage yard represented by the ground point cloud data, the volume of space occupied by the stack is calculated. The estimated weight of the stack is calculated based on the occupied space volume and the density in the physical property parameters. The output includes the morphological parameters of the surface's three-dimensional morphological model, the volume it occupies, and the estimated weight.

6. The multi-scenario adaptive bulk cargo yard intelligent planning system according to claim 1, characterized in that, The planning engine module includes a scene recognition unit; The scene recognition unit is configured as follows: Receive user-inputted scenario raw parameters, which include at least one of yard type, cargo type attribute and operation mode; The original parameters of the scene are standardized and encoded to generate a scene feature vector with a uniform format; Based on predefined scene classification rules, the scene feature vectors are mapped to at least one standardized scene configuration label; The mapped scene configuration tags are combined to output structured scene configuration information.

7. The multi-scenario adaptive bulk cargo yard intelligent planning system according to claim 6, characterized in that, The planning engine module includes a strategy matching unit; The strategy matching unit is connected to the scene recognition unit and is configured as follows: Receive the scene configuration information output by the scene recognition unit; Access the pre-stored planning strategy knowledge base, which stores the mapping relationship between different scenario configuration information and planning strategy rule set; Using the scenario configuration information as the query condition, a search and matching process is performed in the planning strategy knowledge base to obtain a set of planning strategy rules associated with the scenario configuration information; the set of planning strategy rules includes stacking strategy rules, path planning rules, and safety and environmental protection rules. Output the set of planning strategy rules obtained through matching.

8. The multi-scenario adaptive bulk cargo yard intelligent planning system according to claim 7, characterized in that, The planning engine module also includes a scheme generation unit; The scheme generation unit is connected to the strategy matching unit and the data processing module, and is configured as follows: The system receives the set of planning strategy rules output by the strategy matching unit, as well as the yard terrain and morphological parameters represented by the ground point cloud data output by the data processing module. Based on the set of planning strategy rules, combined with the yard topography and the morphological parameters, an analysis and adaptability assessment of the yard's idle areas are conducted to obtain the adaptability assessment results. Based on the adaptive assessment results and the set of planning strategy rules, the spatial layout scheme of the stacking is adjusted; Based on the spatial layout scheme, the morphological parameters of the stacking, and the path planning rules, a device operation path is generated; The planning scheme is generated by integrating the spatial layout scheme with the equipment operation path.

9. The multi-scenario adaptive bulk cargo yard intelligent planning system according to claim 8, characterized in that, It also includes an output module; The output module is configured as follows: Receive the planning scheme; The spatial layout scheme in the planning scheme is converted into a visual yard layout map, and the stacking location, outline, number and key passage information are marked on the yard layout map; The equipment operation path in the planning scheme is converted into a sequence of instructions that the equipment can recognize and execute; the instruction sequence includes action instructions, coordinate point sequence and operation parameters.

10. The multi-scenario adaptive bulk cargo yard intelligent planning system according to claim 7, characterized in that, It also includes an adaptive optimization module; the adaptive optimization module is connected to the planning engine module; The adaptive optimization module is configured as follows: Collect historical operation data, perform statistical analysis on the set of planning strategy rules in the planning strategy knowledge base based on the historical operation data, calculate the deviation between operation efficiency and planning expectations, and generate deviation analysis results; Based on the deviation analysis results, the thresholds, weights, or constraints in the planning strategy rule set are adjusted.

Images