Scanning method and system for embodied agent

Abstract

This application provides a scanning method and system for embodied intelligent agents, relating to the field of 3D scanner automation technology. The method includes: dividing multiple scanning viewpoints on the workpiece to be scanned into multiple scanning partitions, each partition including multiple neighboring scanning viewpoints; determining the chassis docking stations corresponding to each scanning partition; determining the traversal order of each scanning partition based on the coordinate positions of the chassis docking stations corresponding to each scanning partition; controlling a mobile chassis to navigate sequentially to the chassis docking stations corresponding to each scanning partition according to the traversal order, and controlling the 3D scanning device to acquire 3D data of each scanning partition according to pre-identified target shutter parameters; generating and outputting a 3D point cloud model of the workpiece to be scanned based on the 3D data of all scanning partitions. This means the scanning coverage space is unrestricted, meeting the requirement for unobstructed full-surface scanning of large, ultra-long, or irregularly shaped workpieces, thus improving scanning efficiency.

Description

Technical Field

[0001] This application relates to the field of 3D scanner automation technology, and more specifically, to a scanning method and system for an embodied intelligent agent. Background Technology

[0002] Currently, in the field of industrial-grade 3D scanning automation, most solutions still rely on a combination of fixed robotic arms and customized tooling fixtures. Specifically, this solution typically involves rigidly fixing a six-axis industrial robotic arm to a ground base, controlling a 3D scanner mounted at its end effector via a preset trajectory to acquire single-view or multi-view images of a stationary workpiece, and then generating a complete 3D model of the workpiece through point cloud registration.

[0003] However, this scanning method has the problem of limited spatial coverage, making it difficult to meet the needs of full-surface unobstructed scanning of large-sized (such as car bodies and wind turbine blades), ultra-long (such as track beams and ship sections) or irregularly shaped workpieces (such as those with deep cavities, inner curved surfaces, and bottom suspended structures). Summary of the Invention

[0004] The purpose of this application is to provide a scanning method and system for embodied intelligent agents to address the shortcomings of the prior art and solve the problems existing in the prior art.

[0005] To achieve the above objectives, the technical solution adopted in this application is as follows: In a first aspect, embodiments of this application provide a scanning method for an embodied intelligent agent, applied to a computational control module within the embodied intelligent agent. The embodied intelligent agent includes at least: a mobile chassis, a multi-degree-of-freedom robotic arm, and the computational control module. A three-dimensional scanning device is mounted on the end effector of the robotic arm of the embodied intelligent agent. The method includes: Multiple scanning viewpoints on a pre-determined workpiece to be scanned are obtained, and the multiple scanning viewpoints are divided into intervals to obtain multiple scanning partitions, each of which includes multiple neighboring scanning viewpoints. Determine the chassis docking station corresponding to each scanning partition so that after the mobile chassis moves to the chassis docking station corresponding to the scanning partition, the end effector of the robotic arm will not collide and can reach all scanning viewpoints within the scanning partition; Based on the coordinates of the chassis docking stations corresponding to each scanning zone, the order in which each scanning zone is traversed is determined. According to the described sequence, the mobile chassis is controlled to navigate to the chassis docking station corresponding to each of the scanning partitions in turn, and the 3D scanning device is controlled to collect 3D data of each of the scanning partitions according to the preset target shutter parameters; Based on the 3D data of all scanned partitions, a 3D point cloud model of the workpiece to be scanned is generated and output.

[0006] Optionally, acquiring multiple scanning viewpoints on the predetermined workpiece to be scanned includes: Obtain the rough model data of the first workpiece stage and the rough model data of the second workpiece stage. The rough model data of the first workpiece stage is the baseline environment model before the workpiece is placed, and the rough model data of the second workpiece stage is the environment model containing the workpiece obtained by the embodied intelligent agent surrounding and modeling after the workpiece is placed. The second workpiece stage rough model data and the first workpiece stage rough model data are precisely registered and differentially calculated to obtain workpiece rough model data containing only the workpiece. The workpiece rough model data is used to calculate the scanning viewpoints to obtain multiple scanning viewpoints on the workpiece to be scanned, and each scanning viewpoint contains camera pose data.

[0007] Optionally, the step of dividing the plurality of scanning viewpoints into intervals to obtain a plurality of scanning partitions includes: Based on the rough model data of the workpiece and the preset safety buffer distance, the movable working area of ​​the embodied intelligent agent is constructed. Based on the attribute parameters of the embodied intelligent agent and the movable working area, the reachable space range of the end effector of the robotic arm is determined. The attribute parameters include at least: chassis size, maximum working radius of the robotic arm, and safe collision buffer distance of the body. Based on the reachable space range of the robotic arm end effector and the preset spherical clustering algorithm, the multiple scanning viewpoints are iteratively clustered to obtain multiple viewpoint clusters, and each viewpoint cluster is used as an initial scanning partition. The multiple initial scan partitions are deduplicated to obtain the multiple scan partitions.

[0008] Optionally, the step of deduplicating multiple initial scan partitions to obtain the multiple scan partitions includes: Determine the target scan viewpoint located in multiple initial scan partitions; Determine the distance between the target scanning viewpoint and the chassis docking station corresponding to each initial scanning partition; Based on the distance between the target scanning viewpoint and the chassis docking station corresponding to each initial scanning partition, the target scanning viewpoint is located in the target scanning partition, and the target scanning viewpoint is removed from each initial scanning partition except the target scanning partition.

[0009] Optionally, determining the chassis docking station corresponding to each of the scanned partitions includes: The center point of the current scan partition is determined based on the viewpoint partition cluster corresponding to the current scan partition; Based on the reachable space range of the robotic arm end and the center point of the current scanning partition, the ground area corresponding to the current scanning partition is determined, and uniform sampling is performed in the ground area corresponding to the current scanning partition to obtain multiple candidate chassis docking points within the current scanning partition. Verify whether each of the candidate chassis docking points meets the preset collision-free and accessibility conditions when scanning all scanning viewpoints within the current scanning partition; If satisfied, then from at least one candidate chassis docking point that has passed the verification, a target candidate chassis docking point that meets the preset optimal site conditions is selected, and the target candidate chassis docking point is used as the chassis docking site corresponding to the current scanning partition. The optimal site conditions include: the total stroke of the robotic arm covering all viewpoints of the current scanning partition is minimized by the target candidate chassis docking point, and the moving chassis of the embodied intelligent agent does not interfere with the workpiece or the worktable, and the robotic arm of the embodied intelligent agent is in a non-singular configuration.

[0010] Optionally, determining the traversal order of each scanning partition based on the coordinates of the chassis docking stations corresponding to each scanning partition includes: Based on the coordinates of the chassis docking stations corresponding to each scanning partition, a distance matrix is ​​constructed for each scanning partition. Using a pre-defined traveling salesman algorithm, the station traversal order is planned for the distance matrix of each scanned partition to obtain the traversal order of each scanned partition.

[0011] Optionally, the step of controlling the mobile chassis to navigate sequentially to the chassis docking station corresponding to each of the scanning partitions according to the described passing order, and controlling the 3D scanning device to acquire 3D data of each of the scanning partitions according to preset target shutter parameters, includes: Based on the scanning order of adjacent scanning partitions, the obstacle avoidance path between the chassis docking stations corresponding to the adjacent scanning partitions is obtained; The mobile chassis is controlled to navigate sequentially to the chassis docking stations of each of the adjacent scanning zones according to the obstacle avoidance path between the chassis docking stations corresponding to the adjacent scanning zones.

[0012] Optionally, controlling the 3D scanning device to acquire 3D data of each scanning zone according to preset target shutter parameters includes: After controlling the mobile chassis to move to the chassis docking station of the current scanning partition, positioning calibration is performed based on the pre-built scene map, and the mobile chassis is locked after positioning calibration; A start scanning command is sent to the 3D scanning device, and the robotic arm is controlled to scan all the scanning viewpoints in the current scanning partition according to a pre-set continuous scanning trajectory to collect the 3D data of the current scanning partition. The continuous scanning trajectory is generated based on the scanning tilt angle and height of all viewpoints in the current scanning partition. After scanning all the scanning viewpoints in the current scanning partition, a pause scanning command is sent to the 3D scanning device, the mobile chassis is unlocked, and the mobile chassis is controlled to move along the pre-planned obstacle avoidance path to the chassis docking station corresponding to the next scanning partition of the current scanning partition.

[0013] Optionally, before acquiring the predetermined multiple scanning viewpoints on the workpiece to be scanned, the method further includes: In the first scenario, the mobile chassis is controlled to perform scene pre-modeling according to a preset walking path to obtain the initial scene model. The first scenario is an empty scene without any workpieces or 3D scanning equipment placed on it. In the second scenario, based on the current scenario modeling and the initial scenario modeling, the positioning status of the 3D scanning device and the material type of the workpiece to be scanned are identified. The second scenario is when the workpiece to be scanned and the 3D scanning device have been placed. Based on the material type of the workpiece to be scanned, the preset scanning parameters corresponding to the material type are called, and the workpiece to be scanned is pre-scanned based on the preset scanning parameters in order to automatically adjust and determine the target shutter parameters.

[0014] Secondly, embodiments of this application also provide a scanning system, the system comprising: an embodied intelligent agent and a three-dimensional scanning device, wherein the three-dimensional scanning device is communicatively connected to a computing control module in the embodied intelligent agent; The three-dimensional scanning device is used to receive scanning commands issued by the computing control module and send the collected three-dimensional data of each scanning partition to the computing control module. The computational control module is used to execute the embodied intelligent agent scanning method described above.

[0015] Thirdly, embodiments of this application provide an electronic device, which includes a memory for storing one or more programs and a processor. When the one or more programs are executed by the processor, the above-described scanning method for embodied intelligent agents is implemented.

[0016] Fourthly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described scanning method for embodied intelligent agents.

[0017] The beneficial effects of this application are: This application provides a scanning method and system for an embodied intelligent agent. In this application, multiple scanning viewpoints can be divided into intervals to form multiple regions, i.e., scanning partitions. This allows a mobile chassis to move to each scanning partition for full-coverage scanning. A chassis docking station corresponding to each scanning partition is determined, ensuring that after the mobile chassis moves to the docking station, the robotic arm end effector can reach all scanning viewpoints within the scanning partition without collision, thus balancing motion efficiency, obstacle avoidance, and full scanning coverage. Based on the coordinate positions of the docking stations corresponding to each scanning partition, the traversal order of each scanning partition is determined to minimize the overall travel distance of the mobile chassis, enabling global path planning between multiple docking stations for the embodied intelligent agent. Following the traversal order, the mobile chassis is controlled to navigate sequentially to the docking stations corresponding to each scanning partition, and a 3D scanning device is controlled to collect 3D data from each scanning partition according to preset target shutter parameters. Based on the 3D data from all scanning partitions, a 3D point cloud model of the workpiece to be scanned is generated and output, achieving unobstructed scanning of the entire surface of the workpiece. Therefore, the scanning method of the embodied intelligent agent provided in this application can autonomously surround the workpiece, move to multiple positions such as the back of the workpiece and the bottom at a long distance, and complete the overall surround scan in one go. There is no need to use tooling to transfer the workpiece, that is, the scanning coverage space is not limited. It can meet the requirements of full-surface unobstructed scanning of large-sized (such as automobile bodies and wind turbine blades), ultra-long (such as track beams and ship sections) or irregularly shaped workpieces (such as those with deep cavities, inner curved surfaces, and bottom suspended structures), thus improving scanning efficiency.

[0018] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram illustrating an application scenario of a scanning system provided in an embodiment of this application; Figure 2 A flowchart illustrating a scanning method for an embodied intelligent agent provided in an embodiment of this application; Figure 3 A flowchart illustrating another scanning method for an embodied intelligent agent provided in an embodiment of this application; Figure 4A flowchart illustrating another scanning method for an embodied intelligent agent provided in an embodiment of this application; Figure 5 A flowchart illustrating another scanning method for an embodied intelligent agent provided in an embodiment of this application; Figure 6 A flowchart illustrating another scanning method for an embodied intelligent agent provided in an embodiment of this application; Figure 7 A flowchart illustrating another scanning method for an embodied intelligent agent provided in an embodiment of this application; Figure 8 A flowchart illustrating another scanning method for an embodied intelligent agent provided in an embodiment of this application; Figure 9 A flowchart illustrating another scanning method for an embodied intelligent agent provided in an embodiment of this application; Figure 10 A flowchart illustrating another scanning method for an embodied intelligent agent provided in an embodiment of this application; Figure 11 A flowchart illustrating the determination of target shutter parameters in another scanning method for an embodied intelligent agent provided in this application embodiment; Figure 12 A schematic diagram of the structure of a scanning device for an embodied intelligent agent provided in an embodiment of this application; Figure 13 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0022] Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0023] First, the technical terms used in this application will be introduced.

[0024] 1. In this application, "embodied intelligent agent" refers to a mobile robot with its own mobile chassis.

[0025] Among them, embodied intelligent agents have the following advantages: (1) Wide-range scanning, not limited by the mounting base: Ordinary robotic arms are fixed on the ground with a fixed working radius, and large / multi-station workpieces can only be loaded and unloaded in batches; the embodied intelligent body has its own mobile chassis, which can autonomously surround the workpiece, move to the back of the workpiece, and stand at a long distance at the bottom, completing the overall surround scanning in one go, without the need for tooling to transfer the workpiece.

[0026] (2) Multi-site partition scanning, adaptable to ultra-large / irregular workpieces: Multiple docking stations can be planned around the workpiece, and the robotic arm completes local scanning at each station, solving the problems of insufficient single arm extension and blind spot observation in deep cavity; the fixed robotic arm can only change the angle of a single arm, and scanning loopholes are very likely to occur on the back and inside cavity.

[0027] (3) Autonomous environmental perception without the need for pre-deployment of external cameras: It has built-in global depth vision and can autonomously construct on-site environmental maps and identify workpiece position shifts; (4) Adaptable to flexible production of workpieces of various sizes: When changing workpieces of different sizes, only the surrounding station needs to be replanned, and there is no need to modify the robot arm mounting base; for large and extra-long parts, fixed robot arms often need to be replaced with extended arms or modified bases.

[0028] (5) High degree of freedom of multiple joints: An embodied intelligent agent has human-like joints and can freely and flexibly complete more complex movements.

[0029] The following embodiments will specifically illustrate how this application utilizes an embodied intelligent agent, combined with the actual scenario of automated 3D scanning, to achieve an intelligent scanning method that replaces manual labor and enables autonomous scanning.

[0030] Please see Figure 1 , Figure 1 This is a schematic diagram of an application scenario for a scanning system provided in an embodiment of this application. The system includes: an embodied intelligent agent and a 3D scanning device, wherein the 3D scanning device is communicatively connected to the computing control module in the embodied intelligent agent.

[0031] The embodied intelligent agent is equipped with multiple functional modules, as follows: (1) The motion mechanism includes: a mobile chassis, a multi-degree-of-freedom robotic arm, and a dexterous gripper configured at the end of the robotic arm; (2) 3D photography module: realizes 3D data acquisition and obtains a rough model; (3) Two-dimensional image capture module: realizes two-dimensional image capture; (4) Calculation control module: realizes functions such as calculation, communication, and command execution.

[0032] For example, the 3D scanning device can be a high-precision 3D data acquisition device used to acquire 3D data of the workpiece; wherein, the 3D scanning device includes various types of scanners such as tracking type, point-to-point type, and fixed type. This application takes a tracking type scanner as an example, but other types of scanners can also be used in this application.

[0033] It is understandable that the workpiece is the object being scanned and modeled, and is placed on the workpiece stage after the scene pre-modeling is completed.

[0034] It should be noted that, Figure 1 The application scenario shown is before the 3D scanning equipment is assembled to the end of the robotic arm, before scanning begins.

[0035] Optionally, the scanning method of the embodied intelligent agent provided in this application can autonomously surround the workpiece, move to multiple positions such as the back of the workpiece, and the bottom at a long distance, and complete the overall surround scan in one go. There is no need to use tooling to transfer the workpiece, that is, the scanning coverage space is not limited. It can meet the requirements of full-surface unobstructed scanning of large-sized (such as automobile bodies, wind turbine blades), ultra-long (such as track beams, ship sections) or irregularly shaped workpieces (such as those with deep cavities, inner curved surfaces, and bottom suspended structures), thus improving scanning efficiency.

[0036] Figure 2 This application provides a flowchart illustrating a scanning method for an embodied intelligent agent; the executing entity of this method is the aforementioned... Figure 1 The computational control module in an embodied intelligent agent can be an electronic device with data processing capabilities. For example... Figure 2 As shown, the method includes: S101. Obtain multiple scanning viewpoints on the workpiece to be scanned that have been predetermined in advance, and divide the multiple scanning viewpoints into intervals to obtain multiple scanning partitions.

[0037] The workpiece to be scanned is an industrial part that is being processed or inspected. For example, the workpiece to be scanned may be a gear, an engine housing, etc.

[0038] Each scan partition includes multiple neighboring scan viewpoints. For example, scan partition 1 includes 10 neighboring scan viewpoints.

[0039] Among them, the scanning viewpoint is the camera pose of each position point on the workpiece scanned by the 3D scanning equipment. The camera pose includes: camera position (X,Y,Z) and camera angle.

[0040] It is understandable that scanning a complex 3D workpiece requires more than one scanning viewpoint and often requires multiple scans around the workpiece. Therefore, a large number of scanning viewpoints can be calculated based on the rough workpiece model data that only contains the workpiece.

[0041] Optionally, in order to improve scanning efficiency, multiple scanning viewpoints can be divided into intervals according to the feasible standing area of ​​the embodied intelligent agent's body, forming multiple regions, i.e., scanning partitions. In this way, the mobile chassis can move to each scanning partition to perform full-coverage scanning.

[0042] S102. Determine the chassis docking station corresponding to each scanning zone so that after the mobile chassis moves to the chassis docking station corresponding to the scanning zone, the end of the robotic arm will not collide and can reach all scanning viewpoints within the scanning zone.

[0043] A chassis docking station refers to the ground coordinates where the mobile chassis of the embodied intelligent agent is stationary. Within the same station, the robotic arm of the embodied intelligent agent can reach all viewpoints in the area. That is, each scanning zone corresponds to one chassis docking station for the embodied intelligent agent.

[0044] The scan partition is the set of all scan viewpoints corresponding to a single chassis docking station.

[0045] Optionally, for each scanning zone, the optimal ground docking pose, i.e., the chassis docking station, can be obtained based on the center point of each scanning zone and the reachable space of the robotic arm's end effector. In this way, after the mobile chassis moves to the chassis docking station corresponding to each scanning zone, only the robotic arm needs to be moved to complete the entire scanning area. Then, the mobile chassis can be controlled to move to the next station, balancing motion efficiency, obstacle avoidance, and full scanning coverage.

[0046] S103. Determine the traversal order of each scanning zone based on the coordinates of the chassis docking stations corresponding to each scanning zone.

[0047] For example, the coordinates of the chassis docking station corresponding to scan partition 1 (P1), the coordinates of the chassis docking station corresponding to scan partition 2 (P2), and the coordinates of the chassis docking station corresponding to scan partition 3 (P3) are calculated. The order of passing through each scan partition is scan partition 1 -> scan partition 3 -> scan partition 2. In this way, the overall travel distance of the mobile chassis when passing through all scan partitions is minimized.

[0048] In one feasible approach, an optimal movement path can be planned based on the coordinates of the chassis docking stations corresponding to each scanning partition, i.e., the order in which each scanning partition is traversed, so as to minimize the overall movement distance of the mobile chassis, thus enabling the embodied intelligent agent to perform global path planning between multiple docking stations.

[0049] S104. Following the sequence of travel, control the mobile chassis to navigate sequentially to the chassis docking station corresponding to each scanning zone, and control the 3D scanning equipment to collect 3D data of each scanning zone according to the pre-identified target shutter parameters.

[0050] The target shutter parameter is obtained by pre-scanning the workpiece multiple times, based on the acquired scanning data corresponding to the shutter parameter of each scan. In other words, the acquired scanning data corresponding to the target shutter parameter is the maximum.

[0051] It should be noted that the computing control module in the embodied intelligent agent is connected to the 3D scanning device. During the scanning process, the computing control module can send commands such as pause scanning, adjust shutter parameters, and switch modes to the 3D scanning device as needed, and automatically adjust the parameters of the 3D scanning device.

[0052] S105. Based on the 3D data of all scanning partitions, generate and output a 3D point cloud model of the workpiece to be scanned.

[0053] In this embodiment, obstacle avoidance path planning can be performed based on the traversal order of each scanning partition to obtain the obstacle avoidance path between the chassis docking stations corresponding to each scanning partition. The mobile chassis is controlled to navigate to the chassis docking stations corresponding to each scanning partition in sequence according to the obstacle avoidance path between the chassis docking stations corresponding to each scanning partition. The 3D scanning device is controlled to collect the 3D data of each scanning partition according to the preset target shutter parameters to achieve efficient scanning of the workpiece. Then, after scanning all scanning partitions, the 3D data of all scanning partitions are registered as point clouds to generate and output the 3D point cloud model of the workpiece to be scanned, realizing full-surface unobstructed scanning of the workpiece.

[0054] In summary, this application provides a scanning method for an embodied intelligent agent. To improve scanning efficiency, this application proposes dividing multiple scanning viewpoints into intervals to form multiple regions, i.e., scanning partitions. This allows a mobile chassis to move to each scanning partition for full-coverage scanning. Furthermore, it determines the chassis docking station corresponding to each scanning partition, ensuring that after the mobile chassis moves to the docking station, the robotic arm end effector avoids collisions and can reach all scanning viewpoints within the scanning partition, thus balancing motion efficiency, obstacle avoidance, and full scanning coverage. Based on the coordinates of the docking stations corresponding to each scanning partition, the traversal order of each scanning partition is determined to minimize the overall travel distance of the mobile chassis, enabling global path planning between multiple docking stations for the embodied intelligent agent. Following the traversal order, the mobile chassis is controlled to navigate sequentially to the docking stations corresponding to each scanning partition, and the 3D scanning device is controlled to collect 3D data from each scanning partition according to pre-identified target shutter parameters. Based on the 3D data from all scanning partitions, a 3D point cloud model of the workpiece to be scanned is generated and output, achieving unobstructed scanning of the entire surface of the workpiece. Therefore, the scanning method of the embodied intelligent agent provided in this application can autonomously surround the workpiece, move to multiple positions such as the back of the workpiece and the bottom at a long distance, and complete the overall surround scan in one go. There is no need to use tooling to transfer the workpiece, that is, the scanning coverage space is not limited. It can meet the requirements of full-surface unobstructed scanning of large-sized (such as automobile bodies and wind turbine blades), ultra-long (such as track beams and ship sections) or irregularly shaped workpieces (such as those with deep cavities, inner curved surfaces, and bottom suspended structures), thus improving scanning efficiency.

[0055] Optionally, refer to Figure 3 As shown, obtaining multiple scanning viewpoints on the workpiece to be scanned in step S101 above includes: S201. Obtain the rough model data of the first workpiece stage and the rough model data of the second workpiece stage.

[0056] Among them, the first workpiece stage rough model data is the baseline environment model before the workpiece is placed, and the second workpiece stage rough model data is the environment model containing the workpiece obtained by surrounding modeling by the embodied intelligent agent after the workpiece is placed.

[0057] In one feasible approach, before placing the workpiece or the 3D scanning equipment, the mobile chassis is controlled to move around the scanning scene based on a preset walking path, and scene data modeling is performed. Specifically, this includes: (1) Walk to the "preset workpiece photo position", use the two-dimensional photo module in the embodied intelligent agent to take a two-dimensional photo of the workbench, and record it as "workpiece table two-dimensional image A"; (2) During the process of walking around the scanned scene, the three-dimensional photography module in the embodied intelligent agent is used to collect the scanned scene in real time and perform scene modeling to obtain a "scene map"; (3) Model around the workpiece stage to obtain the "workpiece stage rough model A" data, that is, the first workpiece stage rough model data; (4) Take a photo at the preset position on the instrument table to obtain "Instrument Table Photo A"; The coarse model obtained in (3) is already in the "scene map" coordinate system in (2).

[0058] After placing the 3D scanning equipment and the workpiece, the system enters the pre-identification module, as follows: The mobile chassis is controlled to move around the workpiece stage, and a model is created using the low-precision 3D sensor built into the embodied intelligent agent to obtain the "workpiece stage rough model B" data containing the workpiece, which is the second workpiece stage rough model data.

[0059] S202. Perform precise registration and difference calculation on the rough model data of the second workpiece stage and the rough model data of the first workpiece stage to obtain rough model data of the workpiece containing only the workpiece.

[0060] S203. Perform scanning viewpoint calculation on the rough model data of the workpiece to obtain multiple scanning viewpoints on the workpiece to be scanned.

[0061] Each scanning viewpoint contains camera pose data.

[0062] In this embodiment, by precisely aligning the second workpiece stage rough model data with the first workpiece stage rough model data and finding the nearest neighbor point, the nearest corresponding point in the second workpiece stage rough model data is found in the first workpiece stage rough model data. If no corresponding point is found in the second workpiece stage rough model data, it is the workpiece's own data, which is recorded as the workpiece rough model data. Then, based on the workpiece rough model data, the scanning viewpoint is calculated to obtain multiple scanning viewpoints on the workpiece to be scanned, thus realizing the scanning viewpoint calculation based on the rough model.

[0063] Optionally, refer to Figure 4 As shown, in step S101 above, multiple scanning viewpoints are divided into intervals to obtain multiple scanning partitions, including: S301. Based on the rough model data of the workpiece and the preset safety buffer distance, construct the movable working area of ​​the embodied intelligent agent.

[0064] In one feasible approach, the axis-aligned global bounding box (AABB) of the workpiece to be scanned can be calculated based on the coarse model data of the workpiece. A safe buffer distance is then expanded outward along the AABB global bounding box to construct the movable working area of ​​the embodied intelligent agent, which defines the annular feasible motion corridor of the mobile chassis. This constrains the motion boundary of the chassis, prevents the robotic arm from colliding with the workpiece to be scanned during the scanning process, and improves the safety of the scanning operation.

[0065] S302. Based on the attribute parameters and movable working area of ​​the embodied intelligent agent, determine the reachable space range of the robotic arm end effector.

[0066] Among them, the attribute parameters include at least: chassis size, maximum working radius of the robotic arm, and safe collision buffer distance of the machine body.

[0067] In this embodiment, a cylindrical coordinate system is established with the geometric center of the workpiece as the origin. The cylindrical coordinate system includes a radial component R, a circumferential component θ, and an axial component H, which are used to characterize and adapt the scanning operation space layout of the embodied intelligent agent surrounding the workpiece. Based on the attribute parameters of the embodied intelligent agent and the movable working area, the reachable space range of the robotic arm end effector is calculated. That is, given the chassis docking point, the maximum reachable sphere that the robotic arm end effector can reach and the center point of this range are solved. Viewpoints outside the range are divided into other partitions.

[0068] S303. Based on the reachable space range of the robotic arm end effector and the preset spherical clustering algorithm, iterative clustering is performed on multiple scanning viewpoints to obtain multiple viewpoint clusters.

[0069] Each viewpoint cluster is used as an initial scan partition.

[0070] For example, the spherical clustering algorithm is the distance-based K-means spherical clustering method.

[0071] In one feasible approach, spatial clustering is performed on all planned scanning viewpoints, grouping spatially adjacent viewpoints that can be covered by the same chassis location into the same cluster (a partition). This is done using distance-based K-means spherical clustering, specifically: (1) Use the camera's 3D coordinates at each scanning viewpoint as clustering feature points; (2) Set clustering distance threshold: If the distance between two scanning viewpoints is less than the robotic arm's range of motion threshold, they can be classified into the same partition; (3) Iterative clustering: Based on the reachable space range of the robotic arm end and the K-means spherical clustering method, multiple scanning viewpoints are iteratively clustered to obtain multiple viewpoint clusters, and each viewpoint cluster corresponds to an initial scanning partition.

[0072] S304. Perform deduplication on multiple initial scan partitions to obtain multiple scan partitions.

[0073] In one feasible approach, to ensure the rationality of each initial scan partition division, multiple initial scan partitions can be deduplicated to obtain multiple scan partitions. For example, if a certain scan viewpoint is close to the center of multiple viewpoint clusters, the scan partition corresponding to that scan viewpoint is eliminated across cluster viewpoints by a secondary determination based on the reachability of the robotic arm.

[0074] Optionally, refer to Figure 5 As shown, step S304 above includes: S401. Determine the target scan viewpoint located in multiple initial scan partitions.

[0075] S402. Determine the distance between the target scanning viewpoint and the chassis docking station corresponding to each initial scanning partition.

[0076] S403. Based on the distance between the target scanning viewpoint and the chassis docking station corresponding to each initial scanning partition, determine the target scanning partition where the target scanning viewpoint is located, and remove the target scanning viewpoint from each initial scanning partition except the target scanning partition.

[0077] In one feasible approach, to deduplicate multiple initial scan partitions, a method is proposed to find target scan viewpoints located in multiple initial scan partitions based on all scan viewpoints in each initial scan partition. For example, the target scan viewpoints are located in initial scan partition 1 and initial scan partition 2, respectively. Then, the distance between the target scan viewpoint and the chassis docking station corresponding to its initial scan partition 1 is calculated as L1, and the distance between the target scan viewpoint and the chassis docking station corresponding to its initial scan partition 2 is calculated as L2. If L2 is greater than L1, that is, the distance from the target scan viewpoint to the chassis docking station corresponding to initial scan partition 1 is the smallest, then initial scan partition 1 can be used as the target scan partition where the target scan viewpoint is located, and the target scan viewpoint can be removed from initial scan partition 2. In other words, when a scan viewpoint can be covered by two scan partition stations simultaneously, the scan viewpoint can be assigned to the scan partition with lower robotic arm movement cost, reducing the number of chassis movements.

[0078] Optionally, blind zones can be merged and adjusted. Specifically, areas with concentrated, difficult-to-scan viewpoints, such as deep cavities and bottoms of the workpiece, can be divided into separate small zones to avoid long-distance cross-zone scanning. Furthermore, the multiple scanning zones can be balanced. Specifically, the upper limit of the number of scanning viewpoints in each scanning zone can be limited to prevent excessive movement of the robotic arm at a single station and low scanning efficiency. If there are too many scanning viewpoints, the zone can be split into two sub-scanning zones to optimize the zone boundaries and redistribute viewpoints, ensuring the rationality of the scanning zone division.

[0079] Optionally, refer to Figure 6 As shown, step S102 above includes: S501. Determine the center point of the current scan partition based on the viewpoint partition cluster corresponding to the current scan partition.

[0080] S502. Based on the reachable space range of the robotic arm end and the center point of the current scanning partition, determine the ground area corresponding to the current scanning partition, and perform uniform sampling in the ground area corresponding to the current scanning partition to obtain multiple candidate chassis docking points in the current scanning partition.

[0081] In one feasible approach, for each scanning partition, the optimal ground docking pose is solved. Specifically, based on the viewpoint partition cluster corresponding to the current scanning partition, the center of the viewpoint partition cluster, i.e., the center point of the current scanning partition, can be calculated. Then, the center of the reachable range of the robotic arm is aligned with the center point of the current scanning partition to obtain the ground area corresponding to the current scanning partition. The ground area corresponding to the current scanning partition is then uniformly sampled to obtain multiple candidate chassis docking points within the current scanning partition. S503. Verify whether each candidate chassis docking point meets the preset collision-free and accessibility conditions when scanning all scanning viewpoints within the current scanning partition.

[0082] S504. If satisfied, select the target candidate chassis docking point that meets the preset optimal site conditions from at least one candidate chassis docking point that has passed the verification, and use the target candidate chassis docking point as the chassis docking site corresponding to the current scanning partition.

[0083] The optimal site conditions include: minimizing the total travel distance of the robotic arm at the target candidate chassis docking point covering all viewpoints in the current scanning partition, ensuring no interference between the moving chassis of the embodied intelligent agent and the workpiece or worktable, and ensuring that the robotic arm of the embodied intelligent agent is in a non-singular configuration.

[0084] Optionally, accessibility verification is performed on each candidate chassis docking point. Specifically, based on the current candidate chassis docking point, it is determined whether the robotic arm can reach the current candidate chassis docking point without collision when scanning each scanning viewpoint. If so, a target candidate chassis docking point that meets the preset optimal site conditions is selected from at least one candidate chassis docking point that has passed the verification, and the target candidate chassis docking point is used as the chassis docking site corresponding to the current scanning partition. The chassis docking site includes: chassis site coordinates and chassis rotation angle.

[0085] Optionally, refer to Figure 7 As shown, step S103 above includes: S601. Based on the coordinates of the chassis parking stations corresponding to each scanning partition, construct the distance matrix for each scanning partition.

[0086] S602. Using the preset traveling salesman algorithm, the station traversal order is planned for the distance matrix of each scanning partition to obtain the traversal order of each scanning partition.

[0087] In one feasible approach, all scanning partition sites are abstracted as path nodes, and a distance matrix for each scanning partition is constructed based on the coordinates of the chassis docking sites corresponding to each scanning partition. Then, the Traveling Salesman Program (TSP) algorithm is used to plan the site traversal order of each scanning partition's distance matrix to obtain the optimal site traversal order between each scanning partition, i.e., the traversal order, in order to minimize the overall travel distance of the mobile chassis.

[0088] Optionally, refer to Figure 8 As shown, in step S104 above, the mobile chassis is controlled to navigate sequentially to the chassis docking stations corresponding to each scanning partition according to the order of travel, including: S701. Based on the scanning order of adjacent scanning partitions, obtain the obstacle avoidance path between the chassis docking stations corresponding to adjacent scanning partitions.

[0089] S702. Control the mobile chassis to navigate sequentially to the chassis docking stations of each scanning partition according to the obstacle avoidance path between the chassis docking stations corresponding to adjacent scanning partitions.

[0090] In one feasible approach, ground obstacle avoidance paths are added between the chassis docking stations corresponding to each scanning partition, and the chassis movement trajectory of the intelligent agent is planned. Specifically, based on the scanning order of adjacent scanning partitions, an obstacle avoidance path is obtained between the chassis docking stations corresponding to adjacent scanning partitions, and the mobile chassis is controlled to navigate to the chassis docking stations of each scanning partition in sequence according to the obstacle avoidance path between the chassis docking stations corresponding to adjacent scanning partitions, thus realizing the planning of obstacle avoidance paths for partition traversal.

[0091] Optionally, refer to Figure 9 As shown, step S104 above, controlling the 3D scanning device to acquire 3D data of each scanning zone according to the preset target shutter parameters, includes: S801. After controlling the mobile chassis to move to the chassis docking station of the current scanning partition, perform positioning calibration based on the pre-built scene map, and lock the mobile chassis after positioning calibration.

[0092] S802. Send a start scanning command to the 3D scanning device and control the robotic arm to scan all scanning viewpoints in the current scanning partition according to the preset continuous scanning trajectory, so as to collect the 3D data of the current scanning partition.

[0093] The continuous scanning trajectory is generated by sorting the scanning tilt angle and height of all viewpoints within the current scanning partition.

[0094] S803. After scanning all the scanning viewpoints in the current scanning partition, send a pause scanning command to the 3D scanning device, unlock the mobile chassis, and control the mobile chassis to move along the pre-planned obstacle avoidance path to the chassis docking station corresponding to the next scanning partition of the current scanning partition.

[0095] In this embodiment, after controlling the mobile chassis to move to the chassis docking station of the current scanning partition 1, positioning calibration is performed based on the pre-built scene map, and the mobile chassis is locked after positioning calibration, that is, the mobile chassis is locked and stationary; a start scanning command is sent to the 3D scanning device, and the robotic arm is controlled to scan all scanning viewpoints in the current scanning partition in sequence according to the pre-set continuous scanning trajectory to complete the 3D data acquisition of the current scanning partition; after scanning all scanning viewpoints in the current scanning partition, a pause scanning command is sent to the 3D scanning device, the mobile chassis is unlocked, and the mobile chassis is controlled to move along the pre-planned obstacle avoidance path to the chassis docking station corresponding to the next scanning partition of the current scanning partition.

[0096] Optionally, after all scanning partitions have been scanned, a stop scanning command is sent to the 3D scanning device. The 3D scanning device stops scanning and performs post-processing work, performing unified point cloud registration and fusion of the 3D data collected from all scanning partition sites to obtain a complete 3D model of the workpiece.

[0097] Optionally, refer to Figure 10 As shown, before step S101 above, the procedure also includes: S901. In the first scenario, control the mobile chassis to perform scenario pre-modeling according to the preset walking path to obtain the initial scenario model.

[0098] The first scenario is an empty scene without any workpieces or 3D scanning equipment.

[0099] S902. In the second scenario, based on the current scenario modeling and the initial scenario modeling, identify the positioning status of the 3D scanning device and the material type of the workpiece to be scanned.

[0100] The second scenario involves a workpiece to be scanned and a 3D scanning device already in place. S903. Based on the material type of the workpiece to be scanned, call the preset scanning parameters corresponding to the material type, and pre-scan the workpiece to be scanned based on the preset scanning parameters to automatically adjust and determine the target shutter parameters.

[0101] In one feasible approach, before the workpiece and 3D scanning equipment are placed, i.e. in the first scenario, the mobile chassis is controlled to perform scene pre-modeling according to a preset walking path to obtain the initial scene model. The initial scene model includes: "2D image A of the workpiece stage", "scene map", "coarse model A of the workpiece stage" data (i.e., the first coarse model data of the workpiece stage) and "photo A of the instrument stage". After placing the 3D scanning equipment and the workpiece, in the second scene, the pre-recognition module is entered, as follows: (1) Walk to the preset position of the instrument table and take a picture to obtain "Instrument Table Photo B". Compare Instrument Table Photo B with Instrument Table Photo A to identify whether the instrument table has a 3D scanning device. If the equipment is available, the robotic arm will grasp the 3D scanning equipment according to the preset grasping path; if the equipment is unavailable, an alarm will be triggered.

[0102] (2) Pre-identification of workpiece material: walk to the "preset workpiece photo position", use the two-dimensional photo module to take a two-dimensional photo of the worktable, and obtain "two-dimensional image B of workpiece table". Based on the two-dimensional image A of workpiece table and the two-dimensional image B of workpiece table, identify the material type of the workpiece to be scanned, call the preset parameters corresponding to the material type of the workpiece to be scanned, and perform multiple pre-scans on the workpiece to be scanned. Based on the pre-scan results, filter to obtain the target shutter parameters.

[0103] Optionally, based on the two-dimensional image A and the two-dimensional image B of the workpiece stage, the material type of the workpiece to be scanned is identified, including: Noise reduction and illumination compensation were performed on the 2D images A and B of the worktable, respectively, to obtain compensated 2D images A and B. Brightness, contrast, and hue consistency were then processed on the compensated 2D images A and B. Mask segmentation was performed on the processed results to extract the target ROI, and texture, reflection, and depth features were extracted from the target ROI. The probability of various materials was calculated by a classification model, and the material type of the workpiece to be scanned was obtained by confidence screening.

[0104] Optionally, refer to Figure 11 As shown, the workpiece to be scanned is pre-scanned multiple times. Based on the pre-scan results, the target shutter parameters are selected, including: Step 1: Set the preset parameters corresponding to the material type as the current shutter speed parameters; The second step is to send a start scanning command to the 3D scanning device, start scanning with the current shutter parameters, control the moving chassis to quickly sweep the workpiece to be scanned along the preset path, and obtain the scanning data volume N corresponding to the current shutter parameters S. The third step is to add a preset step size L to the current shutter speed parameters to obtain new shutter speed parameters, and determine whether the new shutter speed parameters are within the preset shutter speed range. If yes, use the new shutter speed parameters as the current shutter speed parameters and jump to the second step above. If no, proceed to the fourth step below. Step 4: Initialize the current shutter speed parameters; Step 5: Based on the current shutter speed parameters, decrease the current shutter speed parameters by a preset step size L to obtain the updated current shutter speed parameters; Step 6: Send a start scanning command to the 3D scanning device to start scanning with the updated current shutter parameters, and control the moving chassis to quickly sweep the workpiece to be scanned along the preset path, and obtain the amount of scanning data corresponding to the updated current shutter parameters. Step 7: Determine whether the updated current shutter speed parameters are within the preset shutter speed range; if yes, use the updated current shutter speed parameters as the current shutter speed parameters and jump to step 5 above; if no, proceed to step 8 below. Step 8: Sort the scan data volume corresponding to each of the above shutter parameters, and take the shutter parameter with the largest scan data volume as the target shutter parameter.

[0105] Optionally, refer to Figure 12 As shown, this application also provides a scanning device for an embodied intelligent agent, applied to a computational control module within the embodied intelligent agent. The embodied intelligent agent includes at least: a mobile chassis, a multi-degree-of-freedom robotic arm, and the computational control module. A three-dimensional scanning device is mounted on the end effector of the robotic arm of the embodied intelligent agent. The device includes: The acquisition module 1201 is used to acquire multiple scanning viewpoints on a pre-determined workpiece to be scanned; The partitioning module 1202 is used to divide the multiple scanning viewpoints into intervals to obtain multiple scanning partitions, each scanning partition including multiple neighboring scanning viewpoints; The determining module 1203 is used to determine the chassis docking station corresponding to each of the scanning partitions, so that after the mobile chassis moves to the chassis docking station corresponding to the scanning partition, the end effector of the robotic arm will not collide and can reach all scanning viewpoints within the scanning partition; and the passing order of each scanning partition is determined based on the coordinate position of the chassis docking station corresponding to each of the scanning partitions. The control module 1204 is used to control the mobile chassis to navigate sequentially to the chassis docking station corresponding to each of the scanning partitions according to the passing order, and to control the three-dimensional scanning device to collect three-dimensional data of each of the scanning partitions according to the pre-identified target shutter parameters; The generation module 1205 is used to generate and output a three-dimensional point cloud model of the workpiece to be scanned based on the three-dimensional data of all scanning partitions.

[0106] Optionally, the determining module 1203 is specifically used for: Obtain the rough model data of the first workpiece stage and the rough model data of the second workpiece stage. The rough model data of the first workpiece stage is the baseline environment model before the workpiece is placed, and the rough model data of the second workpiece stage is the environment model containing the workpiece obtained by the embodied intelligent agent surrounding and modeling after the workpiece is placed. The second workpiece stage rough model data and the first workpiece stage rough model data are precisely registered and differentially calculated to obtain workpiece rough model data containing only the workpiece. The workpiece rough model data is used to calculate the scanning viewpoints to obtain multiple scanning viewpoints on the workpiece to be scanned, and each scanning viewpoint contains camera pose data.

[0107] Optionally, the partitioning module 1202 is specifically used for: Based on the rough model data of the workpiece and the preset safety buffer distance, the movable working area of ​​the embodied intelligent agent is constructed. Based on the attribute parameters of the embodied intelligent agent and the movable working area, the reachable space range of the end effector of the robotic arm is determined. The attribute parameters include at least: chassis size, maximum working radius of the robotic arm, and safe collision buffer distance of the body. Based on the reachable space range of the robotic arm end effector and the preset spherical clustering algorithm, the multiple scanning viewpoints are iteratively clustered to obtain multiple viewpoint clusters, and each viewpoint cluster is used as an initial scanning partition. The multiple initial scan partitions are deduplicated to obtain the multiple scan partitions.

[0108] Optionally, the partitioning module 1202 is specifically used for: Determine the target scan viewpoint located in multiple initial scan partitions; Determine the distance between the target scanning viewpoint and the chassis docking station corresponding to each initial scanning partition; Based on the distance between the target scanning viewpoint and the chassis docking station corresponding to each initial scanning partition, the target scanning viewpoint is located in the target scanning partition, and the target scanning viewpoint is removed from each initial scanning partition except the target scanning partition.

[0109] Optionally, the determining module 1203 is specifically used for: The center point of the current scan partition is determined based on the viewpoint partition cluster corresponding to the current scan partition; Based on the reachable space range of the robotic arm end and the center point of the current scanning partition, the ground area corresponding to the current scanning partition is determined, and uniform sampling is performed in the ground area corresponding to the current scanning partition to obtain multiple candidate chassis docking points within the current scanning partition. Verify whether each of the candidate chassis docking points meets the preset collision-free and accessibility conditions when scanning all scanning viewpoints within the current scanning partition; If satisfied, then from at least one candidate chassis docking point that has passed the verification, a target candidate chassis docking point that meets the preset optimal site conditions is selected, and the target candidate chassis docking point is used as the chassis docking site corresponding to the current scanning partition. The optimal site conditions include: the total stroke of the robotic arm covering all viewpoints of the current scanning partition is minimized by the target candidate chassis docking point, and the moving chassis of the embodied intelligent agent does not interfere with the workpiece or the worktable, and the robotic arm of the embodied intelligent agent is in a non-singular configuration.

[0110] Optionally, the determining module 1203 is specifically used for: Based on the coordinates of the chassis docking stations corresponding to each scanning partition, a distance matrix is ​​constructed for each scanning partition. Using a pre-defined traveling salesman algorithm, the station traversal order is planned for the distance matrix of each scanned partition to obtain the traversal order of each scanned partition.

[0111] Optionally, the control module 1204 is specifically used for: Based on the scanning order of adjacent scanning partitions, the obstacle avoidance path between the chassis docking stations corresponding to the adjacent scanning partitions is obtained; The mobile chassis is controlled to navigate sequentially to the chassis docking stations of each of the adjacent scanning zones according to the obstacle avoidance path between the chassis docking stations corresponding to the adjacent scanning zones.

[0112] Optionally, the control module 1204 is specifically used for: After controlling the mobile chassis to move to the chassis docking station of the current scanning partition, positioning calibration is performed based on the pre-built scene map, and the mobile chassis is locked after positioning calibration; A start scanning command is sent to the 3D scanning device, and the robotic arm is controlled to scan all the scanning viewpoints in the current scanning partition according to a pre-set continuous scanning trajectory to collect the 3D data of the current scanning partition. The continuous scanning trajectory is generated based on the scanning tilt angle and height of all viewpoints in the current scanning partition. After scanning all the scanning viewpoints in the current scanning partition, a pause scanning command is sent to the 3D scanning device, the mobile chassis is unlocked, and the mobile chassis is controlled to move along the pre-planned obstacle avoidance path to the chassis docking station corresponding to the next scanning partition of the current scanning partition.

[0113] Optionally, the device further includes: The modeling module is used to control the mobile chassis to perform scene pre-modeling according to a preset walking path in the first scene to obtain the initial scene model. The first scene is an empty scene without any workpieces or 3D scanning equipment. The identification module is used to identify the positioning status of the 3D scanning device and the material type of the workpiece to be scanned in the second scenario, based on the current scenario modeling and the initial scenario modeling. The second scenario is when the workpiece to be scanned and the 3D scanning device have been placed. The pre-scanning module is used to call preset scanning parameters corresponding to the material type of the workpiece to be scanned, and to pre-scan the workpiece based on the preset scanning parameters, so as to automatically adjust and determine the target shutter parameters.

[0114] The above-described device is used to execute the method provided in the foregoing embodiments, and its implementation principle and technical effect are similar, so they will not be described again here.

[0115] These modules can be one or more integrated circuits configured to implement the above methods, such as one or more Application Specific Integrated Circuits (ASICs), one or more digital signal processors (DSPs), or one or more Field Programmable Gate Arrays (FPGAs). Alternatively, when a module is implemented using processing element scheduler code, the processing element can be a general-purpose processor, such as a Central Processing Unit (CPU) or other processor capable of calling program code. Furthermore, these modules can be integrated together as a system-on-a-chip (SOC).

[0116] Figure 13 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 1 The computing control module in the embodied intelligent agent mentioned above can be integrated into an electronic device, which can be a computing device with data processing capabilities.

[0117] The electronic device includes: processor 1301 and memory 1302.

[0118] The memory 1302 is used to store programs, and the processor 1301 calls the programs stored in the memory 1302 to execute the above method embodiments. The specific implementation and technical effects are similar, and will not be described again here.

[0119] Optionally, this application also provides a program product, such as a computer-readable storage medium, including a program that, when executed by a processor, performs the above-described method embodiments.

[0120] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0121] In addition, the functional modules in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0122] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0123] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0124] It will be apparent to those skilled in the art that this application is not limited to the details of the exemplary embodiments described above, and that this application can be implemented in other specific forms without departing from the spirit or essential characteristics of this application. Therefore, the embodiments should be considered illustrative and non-limiting in all respects, and the scope of this application is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within this application. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A scanning method for an embodied intelligent agent, characterized in that, A computational control module applied in an embodied intelligent agent, the embodied intelligent agent comprising at least: a mobile chassis, a multi-degree-of-freedom robotic arm, and the computational control module, wherein a 3D scanning device is mounted on the end effector of the robotic arm of the embodied intelligent agent; the method includes: Multiple scanning viewpoints on a pre-determined workpiece to be scanned are obtained, and the multiple scanning viewpoints are divided into intervals to obtain multiple scanning partitions, each of which includes multiple neighboring scanning viewpoints. Determine the chassis docking station corresponding to each scanning partition so that after the mobile chassis moves to the chassis docking station corresponding to the scanning partition, the end effector of the robotic arm will not collide and can reach all scanning viewpoints within the scanning partition; Based on the coordinates of the chassis docking stations corresponding to each scanning zone, the order in which each scanning zone is traversed is determined. According to the described sequence, the mobile chassis is controlled to navigate sequentially to the chassis docking station corresponding to each of the scanning partitions, and the 3D scanning device is controlled to collect 3D data of each of the scanning partitions according to the pre-identified target shutter parameters; Based on the 3D data of all scanned partitions, a 3D point cloud model of the workpiece to be scanned is generated and output.

2. The method according to claim 1, characterized in that, The acquisition of multiple scanning viewpoints on the pre-determined workpiece to be scanned includes: Obtain the rough model data of the first workpiece stage and the rough model data of the second workpiece stage. The rough model data of the first workpiece stage is the baseline environment model before the workpiece is placed, and the rough model data of the second workpiece stage is the environment model containing the workpiece obtained by the embodied intelligent agent surrounding and modeling after the workpiece is placed. The second workpiece stage rough model data and the first workpiece stage rough model data are precisely registered and differentially calculated to obtain workpiece rough model data containing only the workpiece. The workpiece rough model data is used to calculate the scanning viewpoints to obtain multiple scanning viewpoints on the workpiece to be scanned, and each scanning viewpoint contains camera pose data.

3. The method according to claim 1, characterized in that, The process of dividing the multiple scanning viewpoints into intervals to obtain multiple scanning partitions includes: Based on the rough model data of the workpiece and the preset safety buffer distance, the movable working area of ​​the embodied intelligent agent is constructed. Based on the attribute parameters of the embodied intelligent agent and the movable working area, the reachable space range of the end effector of the robotic arm is determined. The attribute parameters include at least: chassis size, maximum working radius of the robotic arm, and safe collision buffer distance of the body. Based on the reachable space range of the robotic arm end effector and the preset spherical clustering algorithm, the multiple scanning viewpoints are iteratively clustered to obtain multiple viewpoint clusters, and each viewpoint cluster is used as an initial scanning partition. The multiple initial scan partitions are deduplicated to obtain the multiple scan partitions.

4. The method according to claim 3, characterized in that, The process of deduplicating multiple initial scan partitions to obtain the multiple scan partitions includes: Determine the target scan viewpoint located in multiple initial scan partitions; Determine the distance between the target scanning viewpoint and the chassis docking station corresponding to each initial scanning partition; Based on the distance between the target scanning viewpoint and the chassis docking station corresponding to each initial scanning partition, the target scanning viewpoint is located in the target scanning partition, and the target scanning viewpoint is removed from each initial scanning partition except the target scanning partition.

5. The method according to claim 1, characterized in that, The determination of the chassis docking station corresponding to each of the scanned partitions includes: The center point of the current scan partition is determined based on the viewpoint partition cluster corresponding to the current scan partition; Based on the reachable space range of the robotic arm end and the center point of the current scanning partition, the ground area corresponding to the current scanning partition is determined, and uniform sampling is performed in the ground area corresponding to the current scanning partition to obtain multiple candidate chassis docking points within the current scanning partition. Verify whether each of the candidate chassis docking points meets the preset collision-free and accessibility conditions when scanning all scanning viewpoints within the current scanning partition; If satisfied, then from at least one candidate chassis docking point that has passed the verification, a target candidate chassis docking point that meets the preset optimal site conditions is selected, and the target candidate chassis docking point is used as the chassis docking site corresponding to the current scanning partition. The optimal site conditions include: the total stroke of the robotic arm covering all viewpoints of the current scanning partition is minimized by the target candidate chassis docking point, and the moving chassis of the embodied intelligent agent does not interfere with the workpiece or the worktable, and the robotic arm of the embodied intelligent agent is in a non-singular configuration.

6. The method according to claim 1, characterized in that, The step of determining the traversal order of each scanning partition based on the coordinates of the chassis docking stations corresponding to each scanning partition includes: Based on the coordinates of the chassis docking stations corresponding to each scanning partition, a distance matrix is ​​constructed for each scanning partition. Using a pre-defined traveling salesman algorithm, the station traversal order is planned for the distance matrix of each scanned partition to obtain the traversal order of each scanned partition.

7. The method according to claim 1, characterized in that, The step of controlling the mobile chassis to navigate sequentially to the chassis docking station corresponding to each of the scanning partitions according to the stated sequence, and controlling the 3D scanning device to acquire 3D data of each of the scanning partitions according to preset target shutter parameters, includes: Based on the scanning order of adjacent scanning partitions, the obstacle avoidance path between the chassis docking stations corresponding to the adjacent scanning partitions is obtained; The mobile chassis is controlled to navigate sequentially to the chassis docking stations of each of the adjacent scanning zones according to the obstacle avoidance path between the chassis docking stations corresponding to the adjacent scanning zones.

8. The method according to claim 1, characterized in that, The control of the 3D scanning device to acquire 3D data of each scanning zone according to preset target shutter parameters includes: After controlling the mobile chassis to move to the chassis docking station of the current scanning partition, positioning calibration is performed based on the pre-built scene map, and the mobile chassis is locked after positioning calibration; A start scanning command is sent to the 3D scanning device, and the robotic arm is controlled to scan all the scanning viewpoints in the current scanning partition according to a pre-set continuous scanning trajectory to collect the 3D data of the current scanning partition. The continuous scanning trajectory is generated based on the scanning tilt angle and height of all viewpoints in the current scanning partition. After scanning all the scanning viewpoints in the current scanning partition, a pause scanning command is sent to the 3D scanning device, the mobile chassis is unlocked, and the mobile chassis is controlled to move along the pre-planned obstacle avoidance path to the chassis docking station corresponding to the next scanning partition of the current scanning partition.

9. The method according to any one of claims 1-8, characterized in that, Before acquiring multiple predetermined scanning viewpoints on the workpiece to be scanned, the method further includes: In the first scenario, the mobile chassis is controlled to perform scene pre-modeling according to a preset walking path to obtain the initial scene model. The first scenario is an empty scene without any workpieces or 3D scanning equipment placed on it. In the second scenario, based on the current scenario modeling and the initial scenario modeling, the positioning status of the 3D scanning device and the material type of the workpiece to be scanned are identified. The second scenario is when the workpiece to be scanned and the 3D scanning device have been placed. Based on the material type of the workpiece to be scanned, the preset scanning parameters corresponding to the material type are called, and the workpiece to be scanned is pre-scanned based on the preset scanning parameters in order to automatically adjust and determine the target shutter parameters.

10. A scanning system, characterized in that, The system includes: an embodied intelligent agent and a 3D scanning device, wherein the 3D scanning device is communicatively connected to the computing and control module in the embodied intelligent agent; The three-dimensional scanning device is used to receive scanning commands issued by the computing control module and send the collected three-dimensional data of each scanning partition to the computing control module. The computing control module is used to execute the method described in any one of claims 1 to 9.

Images