Map generation method, electronic device, and robot

Abstract

This application discloses a map generation method, electronic device, and robot, relating to the field of artificial intelligence technology. The method includes: generating a structured prior map representing the occupancy state of the environmental space based on environmental geometry design data of the target physical task space; determining a simulated path for the intelligent mobile device to move within the traversable area of ​​the structured prior map, and performing a simulated scanning operation on a virtual physical model of the target physical task space along the simulated path to obtain laser point cloud simulation data and motion state simulation data; and generating a pose-based navigation map based on the laser point cloud simulation data and motion state simulation data. This application can solve the problems of related technologies in map generation, which are highly dependent on manual labor, time-consuming, and produce uncertain results. It can efficiently, cost-effectively, and automatically generate highly accurate maps, improving the task execution efficiency and accuracy of intelligent mobile devices.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence computing, and in particular to a map generation method, electronic device, and robot. Background Technology

[0002] When using smart mobile devices to perform tasks in a designated area, a corresponding task execution path needs to be generated for the smart mobile device based on a high-precision map that can be navigated in real time. Related technologies gradually build maps by conducting on-site, traversal data collection in the area. This on-site mapping method is not only time-consuming and labor-intensive but also susceptible to dynamic interference in the environment, and it also requires manual map post-processing and optimization.

[0003] Therefore, efficiently, cost-effectively, and automatically generating highly accurate maps to improve the task execution efficiency and accuracy of smart mobile devices is a technical problem that needs to be solved by those skilled in the art.

[0004] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of this application, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] This application provides a map generation method, electronic device, and robot that can efficiently, cost-effectively, and automatically generate highly accurate maps, thereby improving the task execution efficiency and accuracy of smart mobile devices.

[0006] To solve the above-mentioned technical problems, this application provides the following technical solution:

[0007] This application provides a map generation method, comprising: generating a structured prior map representing the occupancy state of the environmental space based on environmental geometry design data of the target physical task space; determining a simulated path for a smart mobile device to move within the passage area of ​​the structured prior map, and performing a simulated scanning operation on a virtual physical model of the target physical task space along the simulated path to obtain laser point cloud simulation data and motion state simulation data; and generating a pose-based navigation map based on the laser point cloud simulation data and the motion state simulation data.

[0008] This application also provides an electronic device, including a memory and a processor, wherein the processor is used to implement the steps of the map generation method described above when executing a computer program stored in the memory.

[0009] Finally, this application also provides a robot, including a data input terminal and a path planning processor; the data input terminal receives environmental geometry design data of the target physical task space and sends the environmental geometry design data to the path planning processor; the path planning processor executes computer programs / instructions to implement the steps of the above-described map generation method.

[0010] The advantages of the technical solution provided in this application are that it directly generates a structured prior map based on the existing environmental geometry design data of the target physical task space, replacing the step in related technologies that must rely on on-site environmental perception to construct an initial map, thus eliminating the time and computing power costs consumed by on-site data collection. Because the structured prior map accurately represents the environmental geometry and occupancy state, and plans simulated paths within the traversable area, subsequent scanning operations on the virtual physical model can directly relate to the geometric essence of the environmental space, enabling efficient and noise-free generation of high-fidelity simulation sensor data. This process avoids the problems of dynamic interference and incomplete data collection in the real environment. High-fidelity laser point clouds and motion state data are generated through lightweight scanning using a real-time physical simulation engine, and a pose graph construction algorithm is used to generate a high-precision map that can be directly used for navigation. This method achieves low-cost, rapid, and automated generation from existing design data to a usable navigation map, significantly shortening the deployment cycle, improving map accuracy and reliability, and enhancing the task execution efficiency and accuracy of intelligent mobile devices. Furthermore, this application also provides corresponding electronic devices and robots for the map generation method, further enhancing the practicality of the method. The electronic devices and robots possess corresponding advantages.

[0011] The technical features mentioned above, those to be mentioned below, and those shown individually in the accompanying drawings can be arbitrarily combined, as long as the combined technical features are not contradictory. All feasible combinations of features are the technical content explicitly described in this application. Any one of the multiple sub-features contained in the same statement can be applied independently, without necessarily being applied together with other sub-features.

[0012] It should be understood that the above general description and the following detailed description are merely exemplary and do not limit this application. Attached Figure Description

[0013] To more clearly illustrate the technical solutions of this application or related technologies, the drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a flowchart illustrating a map generation method provided in this application.

[0015] Figure 2 A flowchart illustrating another map generation method provided in this application.

[0016] Figure 3 This is a structural framework diagram of an exemplary embodiment of the map generation apparatus provided in this application.

[0017] Figure 4 This is a structural diagram of an exemplary embodiment of the electronic device provided in this application.

[0018] Figure 5 This is a structural diagram of an exemplary embodiment of the electronic device provided in this application. Detailed Implementation

[0019] To enable those skilled in the art to better understand the technical solutions of this application, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. The terms "first," "second," "third," "fourth," etc., used in the specification and the aforementioned drawings are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. The term "exemplary" means "serving as an example, embodiment, or illustration." Any embodiment described herein as "exemplary" is not necessarily to be construed as superior to or better than other embodiments.

[0020] With the rapid development of artificial intelligence technology, more and more fields will use smart mobile devices to perform tasks in designated areas. For example, robots will be used to perform corresponding tasks in application scenarios such as business guidance, industrial inspection and warehouse handling. Smart mobile devices need to be located and plan their movement paths in real time to perform tasks. This requires a map of the environment in which the smart mobile device is located. It can be seen that building a high-precision map that can be navigated in real time for the operating environment of mobile smart devices is a necessary step to realize their autonomous movement and operation.

[0021] Related technologies rely on perception devices such as LiDAR (Light Detection and Ranging) and visual sensors, using SLAM (Simultaneous Localization and Mapping) methods to collect data online in the operating environment and gradually build an occupancy grid map for navigation. However, in actual deployments, many structured scenarios (such as shopping malls, factories, and hospitals) already have complete CAD (Computer-Aided Design) drawings during the construction phase, containing precise geometric information of static structures such as walls, passageways, and fixed facilities. Related technologies have failed to effectively integrate this prior structured information, still relying on robots for full-coverage on-site exploration and data collection, leading to prolonged mapping cycles and susceptibility to dynamic environmental changes. Furthermore, the generated maps often require manual post-processing and optimization, with limited automation, hindering rapid and consistent deployment across multiple similar scenarios.

[0022] In view of this, this application generates a structured prior map representing the occupancy state of the environmental space based on the environmental geometry design data of the target physical task space where the intelligent mobile device performs the task. A simulated path for the intelligent mobile device to move within the traversable area of ​​the structured prior map is determined, and a simulated scanning operation is performed on the virtual physical model of the target physical task space along the simulated path to obtain laser point cloud simulation data and motion state simulation data. Based on the laser point cloud simulation data and motion state simulation data, an accurate, pose-map-based navigation map is constructed by optimizing the constraint relationships between pose nodes. According to the map generation method provided in this application, it can be implemented as a computer program product. The intelligent mobile device can be, for example, a robot. The computer program product can be installed and run on an edge computing device, which is then deployed on the robot body to generate an environmental navigation map for the robot in real time. The computer program product can also be installed and run on an industrial control terminal, combining the geometric data of production line equipment to generate a structured prior map, enabling automatic updates of the navigation map for factory AGVs (Automated Guided Vehicles, intelligent logistics equipment based on automatic navigation technology), and accurately planning inspection paths for industrial inspection robots. The computer program product can also be installed and run in intelligent warehousing and logistics hubs, generating structured prior maps based on CAD models of warehouse rack layouts, and automatically generating warehouse maps that can guide warehousing robots, thus generating corresponding task execution paths for the robots. The computer program product can also be installed and run on mobile workstations, generating corresponding structured prior maps based on CAD images of disaster sites, and then generating navigation maps for intelligent search devices such as drones to fly on-site based on these structured prior maps. Various non-limiting embodiments of this application are described in detail below with reference to the accompanying drawings and specific implementation details. Please refer to [link to relevant documentation] first. Figure 1 In some embodiments of this method, the following steps may be included:

[0023] S101: Based on the environmental geometry design data of the target physical task space, generate a structured prior map representing the occupancy state of the environmental space.

[0024] The target physical task space refers to the real environment or physical space where the mobile intelligent device will ultimately perform a specific task, such as robot navigation, mapping, or other operations. In other words, the target physical task space refers to the real physical environment area where intelligent mobile devices such as robots and drones need to perform tasks like movement, navigation, and mapping. Its scope is defined by the environmental geometry design data. Environmental geometry design data refers to the electronic design data generated as early as the design phase or as late as the construction phase of the target physical task space. It is a collection of data describing the geometric information of fixed structures (such as walls, columns, and fixed equipment bases) within the target physical task space in digital form. Environmental geometry design data can be any type of digital engineering drawing, such as CAD drawings or BIM (Building Information Modeling) data for the space, including the shape, dimensions, and location of the structures. For example, environmental geometry design data may include the following types of spatial and structural information: Building structure information: such as the location, dimensions, and outline of fixed structures like walls, columns, doors, windows, stairs, and elevator shafts. Functional area division: such as the boundaries of rooms, corridors, passageways, entrances and exits, office areas, warehouse areas, equipment areas, etc. Equipment and facility layout: such as the location of machine bases, workbenches, shelves, fixed containers, pipes, electrical equipment, etc. Passageways and passage spaces: such as main passageways, secondary passageways, turning areas, safety passageways, etc. Dimension and scale information: Drawings typically include scale bars, dimension annotations, coordinate systems, etc., which can be used for mapping real-world dimensions. Annotations and layer information: such as material specifications, elevations, structural numbers, layer separation (such as wall layers, equipment layers, annotation layers, etc.), facilitating information extraction. A structured prior map refers to an environmental geometric model extracted from environmental geometry design data. It represents the target physical task space environment by reflecting the existence of obstacles that prevent passage through certain areas. A structured prior map is a machine-interpretable discrete map that encodes the occupancy status of various locations in the environment (e.g., occupied by obstacles or vacant and passable areas), providing a structured prior environment model for subsequent navigation map generation. A common implementation is a processed occupancy grid map. This step utilizes existing data, such as construction CAD drawings, as prior structural information to quickly generate accurate maps, replacing the inefficient traditional SLAM model that relies entirely on manual operation of robots for on-site mapping.

[0025] S102: Determine the simulated path of the intelligent mobile device moving within the passage area of ​​the structured prior map, and perform a simulated scanning operation on the virtual physical model of the target physical task space along the simulated path to obtain laser point cloud simulation data and motion state simulation data.

[0026] Among them, intelligent mobile devices are automated devices with autonomous or semi-autonomous mobility capabilities, integrating environmental perception, positioning, and decision-making computing units. They are devices that use navigation maps to perform tasks, such as robots, autonomous vehicles, and drones. The simulated path is a continuous path planned within the traversable area of ​​the structured prior map, covering that area, used by the intelligent mobile device for simulated scanning operations. The virtual physical model is an environmental geometric modeling model constructed based on the structured prior map from the above steps, used for ray projection simulation. This model essentially reflects the geometry corresponding to the occupancy state represented by the map. Laser point cloud simulation data is point cloud data reflecting the position and contour of objects in the virtual environment, generated by projecting multiple virtual laser beams onto the virtual physical model and based on the intersection points of the laser beams and obstacles in the virtual physical model. Motion state simulation data is data synchronized with the simulated path, containing the displacement increment and heading angle increment of the intelligent mobile device in adjacent time steps, determined by the motion parameters and simulation parameters of the intelligent mobile device.

[0027] This step simulates the movement of a smart mobile device along a simulated path. At each path point, simulated scanning operations are performed using methods such as ray casting algorithms, ray stepping algorithms, ray tracing algorithms, beam models, or likelihood fields. Taking the ray casting algorithm as an example, multiple virtual laser signals are emitted into the virtual physical model, and intersection points are calculated to simulate lidar scanning, generating simulated laser point cloud data. Simultaneously, the displacement and orientation of the smart mobile device are calculated based on the path, generating simulated motion state data. This step proposes a lightweight sensor data generation method that does not rely on real-time physics simulation engines such as Gazebo (a 3D simulation simulator). It directly simulates the lidar scanning process in the environmental geometry model using a structured prior map, generating high-fidelity simulated laser point cloud data. The simulation process follows an automatically planned, fully covered navigation path, simultaneously generating simulated laser point cloud data and simulated motion state data.

[0028] S103: Generate a pose-based navigation map based on laser point cloud simulation data and motion state simulation data.

[0029] Pose graphs, a map representation method for navigation on smart mobile devices, consist of a series of nodes representing the historical poses (position and attitude) of the smart mobile device and edges representing spatial constraints between nodes (e.g., relative transformations calculated through sensor measurements). It is a sparse but efficient representation suitable for localization and navigation in large-scale scenarios. To facilitate the generation of pose graph-based navigation maps, the above steps can package laser point cloud simulation data and motion state simulation data into a standard bag format. The motion state simulation data and laser point cloud simulation data contain environmental landmarks (which can be represented as sub-mappings) and node associations (the poses from the observed landmark positions), which can be used as pose graph data. This step can directly utilize SLAM algorithms (such as Cartographer and SLAM Toolbox) to perform pose estimation using this pose graph data. Nodes are bound together by spatial constraints, thus realizing the construction of a pose graph-based navigation map.

[0030] In the technical solution provided in this application embodiment, a structured prior map is generated directly based on the existing environmental geometry design data of the target physical task space, replacing the step in related technologies that must rely on on-site environmental perception to construct an initial map, thus eliminating the time and computing power costs consumed by on-site data collection. Since the structured prior map accurately represents the environmental geometry and occupancy status and plans simulated paths within the passage area, subsequent scanning operations on the virtual physical model can be directly associated with the geometric essence of the environmental space, enabling the efficient and noise-free generation of high-fidelity simulation sensor data. This process avoids the problems of dynamic interference and incomplete data collection in the real environment. High-fidelity laser point cloud and motion state data are generated through lightweight scanning of the real-time physical simulation engine, and a high-precision map that can be directly used for navigation is generated using a pose graph construction algorithm. This method achieves low-cost, fast, and automated generation from existing design data to a usable navigation map, significantly shortening the deployment cycle, improving the accuracy and reliability of the map, and enhancing the task execution efficiency and accuracy of intelligent mobile devices.

[0031] Considering that the actual positions of some movable objects (such as shelves and furniture) during construction may deviate from existing design data, relying solely on the structured prior map of S101 cannot directly generate a high-precision navigation map that conforms to the actual environment. Figure 2 As shown, based on the above pose-based map, further mapping can be performed to generate an optimized navigation map, which may include the following:

[0032] Determine the global coverage path of the smart mobile device in the navigation map's travel area; acquire real-time environmental data collected by the smart mobile device during its movement along the global coverage path in the target physical task space, and update the navigation map based on the real-time environmental data.

[0033] The global coverage path is a movement path generated within the travel area of ​​the navigation map that covers the entire travel area. As the smart mobile device moves along this path, it can collect real-time environmental data. Real-time environmental data is data reflecting the actual environmental conditions collected by the smart mobile device as it moves along the global coverage path within the target physical task space.

[0034] In this embodiment, the global coverage path of the intelligent mobile device within the travel area of ​​the generated navigation map is first determined. Then, in the actual target physical task space, the intelligent mobile device equipped with sensors (such as LiDAR, vision sensors, etc.) is controlled to move along the global coverage path. During the movement, environmental data of the target physical task space is collected in real time, i.e., real-time environmental data. This data includes obstacle information, channel size changes, and other information in the real environment. Finally, based on this real-time environmental data collected on-site, channel data (such as channel width, travel height, etc.) or spatial information (such as area boundaries, obstacle distribution, etc.) in the navigation map that do not match the actual environment are corrected, and the navigation map is updated so that it can more accurately reflect the real environment.

[0035] As shown above, this embodiment solves the problem of discrepancies between design drawings and reality that may arise from relying solely on design drawings, such as temporary shelves not marked in the drawings or changes in passageways due to construction. By planning a global coverage path and controlling intelligent mobile devices to collect real-time environmental data along that path, comprehensive coverage detection of the target physical task space is achieved. By comparing the real-time environmental data with a structured prior map, obstacles not reflected in the design data can be accurately identified, and these discrepancies can be automatically discovered and corrected. This allows for the correction and updating of the navigation map, ensuring a high degree of consistency between the navigation map and the real environment, thus improving the map's reliability and usability.

[0036] Based on the above embodiments, this application also addresses the problem of how to efficiently plan a verification path in a complex environment. An exemplary process for determining a global coverage path may include:

[0037] Extract the environmental topology from the navigation map, construct a spatially connected main road network based on the environmental topology, determine the travel area based on the main road network, and generate a movement path covering the travel area within the travel area as a global coverage path.

[0038] The environmental topology reflects structural information such as connectivity and channel distribution between regions within the target physical task space, which is extracted from the navigation map for path planning. The trunk network, constructed based on the environmental topology, is a spatially connected network of major channels and forms the basis for determining travel areas and generating globally covered paths. The trunk network encompasses the main travel channels in the navigation map, ensuring full accessibility for smart mobile devices within this network.

[0039] In this step, a topology extraction algorithm (such as a graph-based topology extraction method) can be used to extract the environmental topology from the navigation map. First, target feature points in the map (such as passage intersections, area boundary points, obstacle vertices, etc.) are identified. Then, a topology graph is constructed based on the connectivity relationships between these feature points. Nodes in the topology graph represent key feature points, and edges represent travel paths between feature points, thus forming an environmental topology that reflects the connectivity relationships of the environment. The extracted environmental topology is analyzed to select the main channels that can achieve full-area connectivity. These channels are typically wide, frequently used, and connect various major functional areas. A spatially connected arterial road network is constructed based on these main channels, ensuring that the arterial road network covers all major travel areas in the navigation map and that all arterial roads are interconnected without isolated sections. Based on the arterial road network, the travel area range in the navigation map is determined. The travel area may include the arterial road network, secondary channels connected to the arterial roads, and passable spaces within functional areas. By identifying and excluding obstacle areas in the map, the boundaries of the travel area are clearly defined. Within a defined travel area, a global coverage path is generated using a coverage path planning algorithm (such as a grid-based coverage algorithm): the travel area is divided into uniform grids, and then a path that can traverse all grids is planned according to a preset coverage order (such as a serpentine traversal order), ensuring that the path can cover the entire travel area, and that the path has no repetition or omission, while minimizing the path length to improve the mobility efficiency of smart mobile devices.

[0040] As can be seen from the above, this embodiment generates a main road network based on map topology. Based on the constructed main road network, the scope of the travel area in the navigation map is clearly defined, ensuring that the path can effectively connect all key areas. Within this travel area, a movement path that can cover the entire travel area is generated and used as a global coverage path. This ensures that when the smart mobile device moves along the path, it can traverse the entire travel area, guaranteeing the integrity and completeness of the on-site data collection, and making the map correction process more comprehensive and reliable.

[0041] Based on the above embodiments, in order to avoid the problems of decreased sensor data quality and increased odometer error caused by irregular movement (such as frequent acceleration and deceleration, sharp turns) of intelligent mobile devices during autonomous exploration and mapping, this embodiment ensures the stability and reliability of the field-collected data by controlling the movement mode of the intelligent mobile device, which may include the following:

[0042] The global coverage path is discretized into multiple linear trajectory segments, each of which is identical or meets preset similarity conditions; the intelligent mobile device is controlled to move along each linear trajectory segment, and pose alignment is performed at each trajectory switching point.

[0043] Linear trajectory segments are path fragments formed by dividing the global coverage path according to preset rules. Each fragment has the same length or meets preset similarity conditions, such as the difference between them being within a minimum range. The minimum value can be flexibly selected according to the actual situation. Trajectory switching points are the connection points between two adjacent linear trajectory segments after the global coverage path is discretized into multiple linear trajectory segments, such as the intersections or endpoints of the segments. Pose alignment operations can be performed by adjusting the device's position and attitude through its own positioning module (such as an IMU (Inertial Measurement Unit), odometer, etc.) to ensure that the device can accurately and smoothly switch to the next linear trajectory segment to continue moving, avoiding positioning deviations caused by trajectory switching.

[0044] In this embodiment, a preset trajectory segment length threshold (e.g., 1 meter) is determined, and the global coverage path is discretized into multiple linear trajectory segments based on this threshold. If the total length of the global coverage path is L, the trajectory segment length threshold is... Then the number of linear trajectory segments `ceil` is a rounding function that ensures the length of each linear trajectory segment does not exceed a preset threshold, and that the segments are connected end-to-end to completely cover the global path. The length of each discretized linear trajectory segment is checked to determine if they are the same or meet preset similarity conditions (e.g., the length difference between any two segments does not exceed 0.1 meters). If any segment does not meet the conditions, it is adjusted (e.g., splitting or merging segments) to ensure all segments meet the requirements for uniform and smooth movement of the smart mobile device. The moving speed (e.g., 0.5 m / s) and acceleration (e.g., ...) of the smart mobile device are set. The system controls the intelligent mobile device to move along the first linear trajectory segment, starting from the beginning of the global coverage path. During movement, the device's position and speed are monitored in real-time by its odometer and IMU to ensure stable movement along the preset trajectory and speed. When the intelligent mobile device is about to reach the end of the first linear trajectory segment (i.e., the first trajectory switching point), its speed is reduced in advance. Upon reaching the trajectory switching point, movement stops and a pose alignment operation is performed. The device's visual sensors collect surrounding environmental features, and combined with a positioning algorithm, adjust the device's position and heading angle to align its attitude with the starting direction of the next linear trajectory segment, accurately aligning the position with the trajectory switching point. After pose alignment, the intelligent mobile device continues to move along the next linear trajectory segment at the preset speed, repeating the above steps until the device completes movement along all linear trajectory segments, achieving a complete traversal of the global coverage path. Throughout the entire movement process, the device continuously collects and stores real-time environmental data.

[0045] As can be seen from the above, this embodiment controls the intelligent mobile device to adopt a step-by-step consistent motion mode, following a parallel trajectory composed of approximately equidistant line segments, moving along each line segment, and performing in-situ rotation at the intersection point to align with the next line segment. By maintaining the stability of the mapping sensor data through constant motion intervals, and maintaining a uniform navigation distance to ensure the stability of continued mapping, the accuracy of map updates is further improved, providing a reliable guarantee for the safe and efficient navigation of intelligent mobile devices.

[0046] Based on the above embodiments, this embodiment also provides an exemplary implementation of how to update a navigation map according to real-time environmental data, which may include the following:

[0047] By comparing real-time environmental data with structured prior maps, target obstacles not represented by environmental geometry design data are identified; based on the target obstacles, channel data or spatial information in the navigation map that do not match the actual environment are corrected.

[0048] Among these, target obstacles are those not represented in the environmental geometry design data but identified by comparing real-time environmental data with a structured prior map. Channel data consists of parameters in the navigation map that characterize channel attributes, such as channel width, height, length, and direction. Spatial information is information in the navigation map that characterizes the spatial distribution of the environment, such as area boundaries, obstacle distribution, and functional zone divisions.

[0049] In this embodiment, preprocessing operations such as data denoising and coordinate calibration can be performed on the real-time environmental data. Data denoising can use filtering algorithms (such as Kalman filtering) to remove noise interference from the sensor-collected data. Coordinate calibration unifies the coordinate system of the real-time environmental data with the coordinate system of the structured prior map to ensure comparability. Then, a point-by-point comparison method can be used to compare and analyze the preprocessed real-time environmental data with the structured prior map: for areas present in the real-time environmental data but not marked in the structured prior map, it is further determined whether the area is an obstacle. For example, by analyzing the geometric features of the area (such as size, shape, positional stability, etc.), if it meets the characteristics of an obstacle (such as having a fixed outline, being impassable, etc.), it is identified as a target obstacle. These obstacles may be newly added facilities after construction, objects whose positions have changed, etc. Detailed information such as the position coordinates, size, and shape of the target obstacles is recorded, and the impact of these target obstacles on the navigation map is analyzed. If a target obstacle is located within a passage area on the original navigation map, causing the passage width to narrow or become completely blocked, or if the target obstacle alters the original area boundaries, the navigation map needs to be corrected accordingly. Based on the identified target obstacle's location and size, channel data (such as channel width and passage height) or spatial information (such as area boundaries and obstacle distribution) in the navigation map that do not match the actual environment are corrected, and the navigation map is updated to more accurately reflect the real environment. Based on the target obstacle information, relevant parameters of the channel data in the navigation map (such as channel width and passage height) are adjusted, area boundaries are redefined, and obstacle distribution information is updated. For example, if a passage's width is reduced from 3 meters to 2 meters due to a new obstacle, the width parameter of that passage is modified accordingly in the navigation map, and the position of the passage's boundary line is adjusted. A consistency check is performed on the updated navigation map to ensure that the connectivity, channel data, and spatial information of each area in the map are consistent and without contradictions or errors. After verification, the updated navigation map is saved for subsequent navigation on smart mobile devices.

[0050] As can be seen from the above, the map generation method of this embodiment, which uses prior guidance and incremental correction, does not simply use existing design data such as CAD drawings directly as the final map. Instead, it acknowledges and bridges the inherent gap between design drawings and physical reality. Using a structured prior map as the initial skeleton and navigation blueprint, it drives intelligent mobile devices to conduct structured on-site exploration. It uses real sensor data, such as LiDAR, to dynamically compare, verify, and refine the structured prior map, thereby automatically discovering and correcting real-world differences such as obstacles not marked in the design data, changes in passageways, and object placement. Finally, it generates an optimized map that highly matches the real environment and can be directly used for navigation.

[0051] The above embodiments do not limit how the structured prior map is generated. Based on the above embodiments, the present invention also provides an exemplary implementation, which may include the following:

[0052] By extracting entity geometry data from the environmental geometry design data of the target physical task space, an initial occupied grid map is generated; based on the connectivity analysis results of the initial occupied grid map, obstacle areas in the initial occupied grid map are determined, and the initial occupied grid map that identifies obstacle areas is used as a structured prior map.

[0053] The entity geometry data refers to the geometric structure information of the entities constituting the environment within the target physical task space's environmental geometry design data, such as the location, size, and outline of walls, pillars, and equipment. The initial occupancy grid map is a preliminary map reflecting the environmental occupancy status, generated by extracting entity geometry data from the target physical task space's environmental geometry design data. For example, the geometry in a CAD drawing can be converted into a grid map by adjusting the scale information, thus obtaining the initial occupancy grid map. The occupancy grid map discretizes the environment into a regular grid, with each grid (called a cell or voxel) storing the probability that a location is occupied by an obstacle. The connectivity analysis result is obtained after analyzing the initial occupancy grid map, showing whether the areas in the map are connected and the distribution of passable and obstacle areas.

[0054] For example, the process of determining obstacle areas in the initial occupied grid map may include: determining the passable and obstacle areas of the initial occupied grid map based on the connectivity analysis results; setting the obstacle areas as occupied state values ​​and the passable areas as vacant state values; and performing binarization processing on the initial occupied grid map based on the area attribute values. Here, the occupied state value is a numerical value used to identify obstacle areas in the initial occupied grid map, indicating that the area is occupied by obstacles. The vacant state value is a numerical value used to identify passable areas in the initial occupied grid map, indicating that the area is free space and passable.

[0055] In this embodiment, the environmental geometry design data is CAD drawings. The CAD drawings can be layered, and layers containing entity information such as walls, columns, equipment, and passageways can be selected. Then, the outline, position coordinates, and size parameters of each entity are identified. For example, the start and end coordinates, length, and thickness of walls are extracted, and the center coordinates and diameter of columns are extracted. This data is then organized into a standardized entity geometry dataset. The resolution of the raster map is determined (e.g., 0.05 meters / raster), and a coordinate system corresponding to the real environment is established. The entity geometry data is then mapped onto the raster map. Each raster is marked with its initial state based on whether it is occupied by an entity, thus forming an initial occupied raster map that can initially reflect the environmental occupancy situation. Connectivity analysis is performed on the initial occupied raster map. A graph search algorithm (such as breadth-first search) or a fast random tree algorithm is used to traverse the entire map to determine the connectivity relationships between each raster. Through analysis, the passable areas (i.e., the set of interconnected free raster cells) and obstacle areas (i.e., the set of raster cells occupied by entities) in the map are determined, clarifying the distribution range and boundaries of each area. The initial occupied grid map is binarized. One approach is to directly black out inaccessible obstacle areas. Another approach is to set the grid values ​​of obstacle areas to occupied (e.g., value 1) and the grid values ​​of passable areas to idle (e.g., value 0). After processing, a structured prior map is obtained, which clearly marks the obstacle and passable areas in the environment and can serve as the base map for subsequent path planning, simulated scanning, and other operations.

[0056] As can be seen from the above, this embodiment, by employing a professional drawing information extraction algorithm, can automatically and accurately extract entity geometric structure data from CAD drawings, avoiding the tediousness and errors of manual extraction and improving data extraction efficiency and accuracy. The initial occupancy grid map generated based on the extracted high-quality data can realistically reflect the geometric features of the environment. Through connectivity analysis, the passable area and obstacle area are clearly delineated and binarized, resulting in a concise and clear structured prior map that provides an accurate data foundation for subsequent operations such as path planning, simulated scanning, and pose graph construction, ensuring map generation accuracy.

[0057] Based on the above embodiments, this application also provides an implementation process for efficiently generating high-fidelity sensor simulation data directly on a geometric model, thereby solving the problems of complex configuration and high computational resource consumption caused by using large-scale simulation software, making the entire map generation process more lightweight, efficient, easy to integrate and automate, and may include the following:

[0058] The environmental topology is extracted from the structured prior map. The passable area is determined based on the environmental topology, and a continuous path covering the passable area is determined as the simulation path. Based on the motion parameters and simulation parameters of the smart mobile device, motion state simulation data synchronized with the simulation path is determined. The simulation scanning position is determined based on the motion state simulation data. At each simulation scanning position, multiple virtual laser beams are projected onto the virtual physical model. Based on the intersection points of each virtual laser beam and obstacles in the virtual physical model, laser point cloud simulation data is generated.

[0059] The simulated scanning position is the specific location where the smart mobile device performs virtual laser scanning, determined based on motion state simulation data. The virtual laser beam is the laser beam projected onto the virtual physical model during the simulated scanning operation to generate simulated laser point cloud data. The motion state simulation data includes the displacement increment and heading angle increment of the smart mobile device in adjacent time steps. Motion parameters are parameters related to the smart mobile device's own motion attributes, such as maximum speed, acceleration, and turning radius. Simulation parameters are relevant settings during the simulated scanning process, such as time step, scanning frequency, and number of laser beams.

[0060] In this embodiment, based on a structured prior map that clearly identifies at least the obstacles, a topology extraction algorithm can be used to extract the environmental topology from the structured prior map, identifying target nodes (such as passage intersections, area entrances / exits, etc.) and the paths connecting these nodes, thus constructing a topology map reflecting the connectivity of the environment. Based on the extracted environmental topology, the scope of the passable area is determined, and a path planning algorithm (such as a fast-exploration random tree algorithm) is used to plan a continuous simulated path within the passable area. This path must cover the entire passable area to ensure that the intelligent mobile device can perform a comprehensive simulated scan of the environment when moving along the path. Simultaneously, the path should be as smooth as possible, reducing the number of turns and sharp angles to lower the difficulty of the device's movement. The motion parameters and simulation parameters of the intelligent mobile device are determined. Motion parameters include the device's maximum moving speed, acceleration, turning radius, etc.; simulation parameters include the time step (e.g., 0.01 seconds), scanning frequency (e.g., 10Hz), etc. Based on these parameters, combined with the length and curvature changes of the simulated path, the displacement increment and heading angle increment of the device within each time step are calculated, forming motion state simulation data synchronized with the simulated path. For example, the motion state simulation data is virtual odometry data. Based on the linear velocity, angular velocity, simulation step time, and heading angle of the smart mobile device at the current moment, the virtual odometry data is determined by substituting these values ​​into the path kinematics model. Based on the motion state simulation data, the simulated scanning position of the smart mobile device at each time step is calculated sequentially, ensuring that the scanning position accurately corresponds to the device's motion trajectory on the simulated path. At each simulated scanning position, the control device projects multiple uniformly distributed virtual laser beams into the virtual physical model. Parameters such as the number of laser beams (e.g., 360 beams), the scanning angle range (e.g., 0-360 degrees), and the measurement distance range can be set according to actual needs. The emission direction and propagation path of each virtual laser beam are recorded. When a laser beam encounters an obstacle in the virtual physical model, such as a wall, the intersection coordinates are calculated and recorded. All intersection coordinates are organized and sorted according to the scanning order and laser beam number to generate laser point cloud simulation data. This data comprehensively reflects the position and contour information of obstacles in the virtual environment.

[0061] As shown above, this embodiment plans the simulated path based on the environmental topology, ensuring that the path fully covers the passage area and guarantees the generation of complete laser point cloud simulation data. The motion state simulation data generated based on the device motion parameters and simulation parameters is highly synchronized with the simulated path, accurately reflecting the device's motion state and ensuring the accuracy of the simulated scanning position. By reasonably setting the virtual laser scanning parameters, it is possible to ensure the generation of laser point cloud simulation data with appropriate resolution and complete information, which can clearly present the geometric features of the virtual environment without causing data redundancy, thus improving data processing efficiency. Using this high-quality simulation data to construct a pose graph navigation map can further improve the accuracy and reliability of the map, providing strong support for the navigation of intelligent mobile devices.

[0062] It should be noted that there is no strict order of execution for the steps in this application. As long as they conform to a logical order, these steps can be executed simultaneously or in a certain preset order. Figures 1-2 This is just an illustrative example and does not mean that this is the only possible execution order.

[0063] This application also provides a corresponding apparatus for the map generation method, further enhancing the practicality of the method. The apparatus can be described from both a functional module perspective and a hardware perspective. The map generation apparatus provided in this application is described below. This apparatus is used to implement the map generation method provided in this application. In this embodiment, the map generation apparatus may include or be divided into one or more program modules. These one or more program modules are stored in a storage medium and executed by one or more processors to complete the map generation method disclosed in Embodiment 1. The program module referred to in this embodiment is a series of computer program instruction segments capable of performing specific functions, which is more suitable than the program itself for describing the execution process of the map generation apparatus in the storage medium. The following description will specifically introduce the functions of each program module in this embodiment. The map generation apparatus described below and the map generation method described above can be referred to in correspondence.

[0064] From the perspective of functional modules, see Figure 3 , Figure 3 This is a structural diagram of the map generation apparatus provided in this embodiment under one specific implementation. The apparatus may include:

[0065] The prior map generation module 301 is used to generate a structured prior map representing the state of the environment space based on the environmental geometry design data of the target physical task space.

[0066] The simulation data generation module 302 is used to determine the simulated path of the intelligent mobile device moving within the passage area of ​​the structured prior map, and to perform a simulated scanning operation on the virtual physical model of the target physical task space along the simulated path to obtain laser point cloud simulation data and motion state simulation data.

[0067] The map generation module 303 is used to generate a pose-based navigation map based on laser point cloud simulation data and motion state simulation data.

[0068] For example, in some embodiments of this embodiment, the above-mentioned device may further include a map optimization and update module, which is used to: determine the global coverage path of the smart mobile device in the travel area of ​​the navigation map; acquire real-time environmental data collected by the smart mobile device during its movement along the global coverage path in the target physical task space, and update the navigation map according to the real-time environmental data.

[0069] As an exemplary implementation of the above embodiments, the map optimization and update module can be further used to: discretize the global coverage path into multiple linear trajectory segments, each linear trajectory segment being the same or satisfying a preset similarity condition; control the smart mobile device to move along each linear trajectory segment, and perform pose alignment operations at each trajectory switching point.

[0070] As another exemplary implementation of the above embodiments, the map optimization and update module can be further used to: extract the environmental topology from the navigation map, construct a spatially connected trunk road network based on the environmental topology; determine the travel area based on the trunk road network, and generate a movement path covering the travel area within the travel area as a global coverage path.

[0071] As another exemplary implementation of the above embodiments, the map optimization and update module can be further used to: compare real-time environmental data with structured prior maps to identify target obstacles not represented by environmental geometry design data; and correct channel data or spatial information in the navigation map that does not match the actual environment based on the target obstacles.

[0072] For example, in some other embodiments of this embodiment, the prior map generation module 301 can also be used to: generate an initial occupied grid map by extracting entity geometric structure data from the environmental geometry design data of the target physical task space; determine the obstacle area in the initial occupied grid map based on the connectivity analysis results of the initial occupied grid map, and use the initial occupied grid map that identifies the obstacle area as a structured prior map.

[0073] As an exemplary implementation of the above embodiments, the prior map generation module 301 can be further used to: determine the passable area and obstacle area of ​​the initial occupied grid map based on the connectivity analysis results of the initial occupied grid map; set the obstacle area as an occupied state value and the passable area as an idle state value; and perform binarization processing on the initial occupied grid map based on the area attribute values.

[0074] For example, in some other embodiments of this embodiment, the simulation data generation module 302 can also be used to: extract the environmental topology from the structured prior map, determine the passage area based on the environmental topology, and determine a continuous path covering the passage area as a simulation path; determine motion state simulation data synchronized with the simulation path based on the motion parameters and simulation parameters of the smart mobile device, wherein the motion state simulation data includes the displacement increment and heading angle increment of the smart mobile device in adjacent time steps; determine the simulation scanning position based on the motion state simulation data, and project multiple virtual laser beams onto the virtual physical model at each simulation scanning position; and generate laser point cloud simulation data based on the intersection points of each virtual laser beam and obstacles in the virtual physical model.

[0075] The map generation device mentioned above is described from the perspective of functional modules. Furthermore, this application also provides an electronic device, which is described from the perspective of hardware. Figure 4 This is a schematic diagram of the structure of an electronic device provided in one embodiment of the present application. The electronic device includes a memory and a processor. The memory stores a computer program, and the processor is configured to run the computer program to perform the steps in any of the map generation method embodiments described above.

[0076] like Figure 4As shown, the electronic device includes a memory 40 for storing computer programs and a processor 41 for executing the computer program to implement the steps of the map generation method mentioned in any of the above embodiments. The processor 41 may include one or more processing cores, such as a quad-core processor or an octa-core processor. The processor 41 may also be a controller, microcontroller, microprocessor, or other data processing chip. The processor 41 may be implemented using at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), or PLA (Programmable Logic Array). The processor 41 may also include a main processor and a coprocessor. The main processor is used to process data in the wake-up state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, the processor 41 may integrate a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content required to be displayed on the screen. In some embodiments, processor 41 may also include an AI (Artificial Intelligence) processor for processing computational operations related to machine learning.

[0077] The memory 40 may include one or more computer non-volatile storage media, which may be non-transitory. The memory 40 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices or flash memory devices. In some embodiments, the memory 40 may be an internal storage unit of an electronic device, such as a server hard drive. In other embodiments, the memory 40 may be an external storage device of an electronic device, such as a plug-in hard drive on a server, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. Furthermore, the memory 40 may include both internal and external storage units of the electronic device. The memory 40 can be used not only to store application software and various types of data installed on the electronic device, such as code in the process of executing the map generation method, but also to temporarily store data that has been output or will be output. In this embodiment, the memory 40 is used to store at least the following computer program 401, which, after being loaded and executed by the processor 41, can implement the relevant steps of the map generation method disclosed in any of the foregoing embodiments. In addition, the resources stored in the memory 40 may also include an operating system 402 and data 403, and the storage method may be temporary storage or permanent storage. The operating system 402 may include Windows, Unix, Linux, etc. The data 403 may include, but is not limited to, data corresponding to the map generation results.

[0078] In some embodiments, the aforementioned electronic device may further include a display screen 42, an input / output interface 43, a communication interface 44 (or network interface), a power supply 45, and a communication bus 46. The display screen 42 and input / output interface 43, such as a keyboard, are user interfaces. Exemplary user interfaces may also include standard wired interfaces, wireless interfaces, etc. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen, etc. The display may also be appropriately referred to as a display screen or display unit, used to display information processed in the electronic device and to display a visual user interface. The communication interface 44 may exemplary include wired and / or wireless interfaces, such as a Wi-Fi interface, a Bluetooth interface, etc., typically used to establish communication connections between the electronic device and other electronic devices. The communication bus 46 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. This bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0079] Those skilled in the art will understand that Figure 4 The structure shown does not constitute a limitation on the electronic device and may include more or fewer components than shown, such as sensors 47 that perform various functions.

[0080] It is understood that if the map generation method in the above embodiments is implemented as a software functional unit and sold or used as an independent product, it can be stored in a non-volatile storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the related technology, or all or part 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 executes all or part of the steps of the methods in the various embodiments of this application. The aforementioned storage medium includes, but is not limited to, various media capable of storing program code, such as: USB flash drive, mobile hard disk, read-only memory (ROM), random access memory (RAM), electrically erasable programmable ROM, register, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, removable disk, CD-ROM, magnetic disk, or optical disk. Based on this, this application also provides a non-volatile storage medium storing a computer program, which, when executed by a processor, performs the steps of the map generation method as described in any of the above embodiments.

[0081] It is understood that if the map generation method in the above embodiments is implemented as a software functional unit and sold or used as an independent product, the computer software product may not need to be stored in a physical storage medium. For example, it can be directly transmitted to a computer or other device with information processing capabilities via a wired or wireless network to execute all or part of the steps of the methods in the various embodiments of this application. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the related technology, or all or part of the technical solution, can be embodied in the form of a software product. Based on this, this application also provides a computer program product, which stores a computer program, and when the computer program is executed by a processor, it performs the steps of the map generation method as described in any of the above embodiments.

[0082] Finally, this application also provides a robot, which may include at least one image acquisition device 501. Figure 5 The diagram shows three cameras, and also includes a data input terminal and a path planning processor 502, which can be deployed in... Figure 5The data input terminal in the robot's base shown in the figure is not displayed. It can be set on the top, the base, or any other location. It is used to receive the environmental geometry design data of the target physical task space and send the environmental geometry design data to the path planning processor 502. The data input terminal can be a wired interface (such as a USB interface, Ethernet interface, etc.) or a wireless interface (such as a Wi-Fi module, Bluetooth module, etc.). For example, by importing CAD files through the USB interface or downloading BIM data from the network, the path planning processor 502 executes computer programs / instructions to implement the steps of the map generation method described in any of the above method embodiments. For example, the path planning processor 502 generates an occupation grid map by extracting structural information from CAD drawings, and further constructs a pose map for navigation. Then, it generates an efficient navigation map through path exploration and optimization algorithms, thereby greatly improving the robot's deployment speed and map quality. The steps include: (1) generating an occupation grid map for robot navigation based on the CAD drawings of the construction of areas such as shops and factories. (2) Construct a pose map for navigation based on the generated occupied grid map; combine image processing, overlay path planner and ray casting, and automatically generate pose map data from the occupied grid map. (3) Map optimization; combine fast exploration random tree and step-consistent motion strategy to automatically explore and map the main road of the map, and generate an optimized navigation map.

[0083] As can be seen from the above, this embodiment transforms robot map building from a task that is highly dependent on on-site manual labor, time-consuming, and with uncertain results into an efficient, low-cost, automated, and reliable industrial process. It is suitable for scenarios with CAD drawings (such as chain stores and standardized factories), and can realize rapid copying and deployment from drawing input to automatic map output. This greatly improves the feasibility and consistency of large-scale robot deployment and has significant practical value in fields such as warehousing and logistics, smart factories, and commercial retail.

[0084] The foregoing has provided a detailed description of a map generation method, electronic device, and robot provided in this application. The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. Whether the units and algorithm steps of the various examples described in the disclosed embodiments are executed by electronic hardware or computer software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, and such implementations should not be considered beyond the scope of this application. Several improvements and modifications can be made to this application without departing from the principles of this application, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A map generation method, characterized in that, include: Based on the environmental geometry design data of the target physical task space, a structured prior map representing the occupancy state of the environmental space is generated; The structured prior map generation process is as follows: by extracting entity geometric structure data from the environmental geometry design data of the target physical task space, an initial occupied grid map is generated; based on the connectivity analysis results of the initial occupied grid map, obstacle areas in the initial occupied grid map are determined, and the initial occupied grid map that identifies obstacle areas is used as the structured prior map. Determine the simulated path of the intelligent mobile device moving within the passage area of ​​the structured prior map, and perform a simulated scanning operation on the virtual physical model of the target physical task space along the simulated path to obtain laser point cloud simulation data and motion state simulation data. A pose-based navigation map is generated based on the laser point cloud simulation data and the motion state simulation data. The process of determining laser point cloud simulation data and motion state simulation data includes: determining motion state simulation data synchronized with the simulation path based on the motion parameters and simulation parameters of the intelligent mobile device; the motion state simulation data includes the displacement increment and heading angle increment of the intelligent mobile device in adjacent time steps; determining the simulation scanning position based on the motion state simulation data; projecting multiple virtual laser beams onto the virtual physical model at each simulation scanning position; generating laser point cloud simulation data based on the intersection points of each virtual laser beam and obstacles in the virtual physical model; the simulation path is a continuous path planned within the passage area of ​​the structured prior map and covering the passage area.

2. The map generation method according to claim 1, characterized in that, After generating the pose-based navigation map, the following steps are also included: Determine the global coverage path of the smart mobile device in the travel area of ​​the navigation map; The system acquires real-time environmental data collected by the smart mobile device during its movement along the global coverage path in the target physical task space, and updates the navigation map based on the real-time environmental data.

3. The map generation method according to claim 2, characterized in that, Real-time environmental data collected during movement along the global coverage path includes: The global coverage path is discretized into multiple linear trajectory segments, and each linear trajectory segment is the same or satisfies a preset similarity condition. The intelligent mobile device is controlled to move along each linear trajectory segment, and pose alignment is performed at each trajectory switching point.

4. The map generation method according to claim 2, characterized in that, Determine the global coverage path of the smart mobile device in the navigation map's travel area, including: Extract the environmental topology from the navigation map, and construct a spatially connected trunk road network based on the environmental topology; Based on the main road network, a travel area is determined, and a movement path covering the travel area is generated within the travel area as a global coverage path.

5. The map generation method according to claim 2, characterized in that, Updating the navigation map based on the real-time environmental data includes: By comparing the real-time environmental data with the structured prior map, target obstacles not represented by the environmental geometry design data are identified; Based on the target obstacle, correct the channel data or spatial information in the navigation map that does not match the actual environment.

6. The map generation method according to claim 1, characterized in that, Based on the connectivity analysis results of the initial occupied grid map, the obstacle regions in the initial occupied grid map are determined, including: Based on the connectivity analysis results of the initial occupied grid map, the passable area and obstacle area of ​​the initial occupied grid map are determined; The obstacle area is set to an occupied state value, the passable area is set to an idle state value, and the initial occupied grid map is binarized according to the area attribute value.

7. The map generation method according to any one of claims 1 to 6, characterized in that, Determining the simulated path of a smart mobile device moving within the access area of ​​the structured prior map includes: The environmental topology is extracted from the structured prior map, the passable area is determined based on the environmental topology, and a continuous path covering the passable area is determined as the simulated path.

8. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of the map generation method as described in any one of claims 1 to 7 when executing the computer program.

9. A robot, characterized in that, Includes data input terminals and path planning processors; The data input terminal receives environmental geometry design data of the target physical task space and sends the environmental geometry design data to the path planning processor. The path planning processor executes computer programs / instructions to implement the steps of the map generation method as described in any one of claims 1 to 7.

Images