Map model construction method and device
By acquiring and processing inertial navigation data, point cloud data and images on the map building mobile device, the problem of missing key environmental information and mistakenly adding non-environmental objects is solved, and a map model construction with higher accuracy is achieved.
Patent Information
- Application Number
- CN202411974556.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-31
- Publication Date
- 2025-04-15
AI Technical Summary
In the prior art, map models lack critical environmental information and errors increase objects that do not belong to the environment to the map models, especially in small and crowded environments.
By controlling the map building mobile device to travel in the target environment, obtaining inertial navigation data, point cloud data and images. The combined algorithm is used to convert point cloud data to the global coordinate system and integrate it, perform ground segmentation and plane fitting, and mark the ground. The image is detected by object, the images of non-static objects are filtered out, and their corresponding point cloud data is deleted to obtain the final map model.
It solves the problem of missing key environmental information and mistakenly adding non-environmental objects, and improves the accuracy of the map model.
Smart Images
Figure CN119826801A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of vehicle traffic technology, and in particular to a map model construction method and device. Background Art
[0002] At present, there are the following problems in constructing map models: because the environmental data is not perfect when it is acquired, the final constructed map model lacks key environmental information; pedestrians and vehicles in the environment are mistakenly presented in the map model, etc. The above problems are particularly obvious in the construction of map models in small and crowded environments. Summary of the invention
[0003] In view of this, embodiments of the present application provide a map model construction method, device, electronic device and computer-readable storage medium to solve the problem in the prior art that the map model lacks key environmental information and erroneously adds objects that do not belong to the environment to the map model.
[0004] In a first aspect of an embodiment of the present application, a method for constructing a map model is provided, comprising: controlling a mapping mobile device to travel along a set route in a target environment, and obtaining a plurality of inertial navigation data of the mapping mobile device, a plurality of point cloud data of the target environment, and a plurality of images during the movement of the mapping mobile device, wherein one point cloud data corresponds to one inertial navigation data and one image; processing each point cloud data using a combination algorithm based on the inertial navigation data corresponding to each point cloud data, so as to convert each point cloud data into a global coordinate system and integrate them to obtain an initial map model; performing ground segmentation on the initial map model to obtain a plurality of ground point cloud data; performing plane fitting on the plurality of ground point cloud data to obtain the ground, and marking the ground in the initial map model; performing target detection on each image to obtain a detection result of each image, and screening out images whose detection results are non-static objects from the plurality of images; deleting the point cloud data corresponding to the images whose detection results are non-static objects in the initial map model to obtain a final map model.
[0005] According to a second aspect of an embodiment of the present application, a map model construction device is provided, comprising: an acquisition module, configured to control a mapping mobile device to travel along a set route in a target environment, and during the travel of the mapping mobile device, multiple inertial navigation data of the mapping mobile device, multiple point cloud data of the target environment, and multiple images are acquired, wherein one point cloud data corresponds to one inertial navigation data and one image; a combination module, configured to process each point cloud data using a combination algorithm based on the inertial navigation data corresponding to each point cloud data, so as to convert each point cloud data into a global coordinate system and integrate them to obtain an initial map model; a ground segmentation module, configured to perform ground segmentation on the initial map model to obtain multiple ground point cloud data; a plane fitting module, configured to perform plane fitting on multiple ground point cloud data to obtain the ground, and mark the ground in the initial map model; a target detection module, configured to perform target detection on each image to obtain a detection result of each image, and filter out images whose detection results are non-static objects from the multiple images; a deletion module, configured to delete the point cloud data corresponding to the image whose detection result is non-static object in the initial map model to obtain a final map model.
[0006] According to a third aspect of an embodiment of the present application, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the computer program.
[0007] According to a fourth aspect of an embodiment of the present application, a computer-readable storage medium is provided, which stores a computer program, and when the computer program is executed by a processor, the steps of the above method are implemented.
[0008] Compared with the prior art, the embodiments of the present application have the following beneficial effects: controlling the mapping mobile device to travel along the set route in the target environment, obtaining multiple inertial navigation data of the mapping mobile device, multiple point cloud data of the target environment and multiple images during the movement of the mapping mobile device, wherein one point cloud data corresponds to one inertial navigation data and one image; processing each point cloud data using a combination algorithm according to the inertial navigation data corresponding to each point cloud data, so as to convert each point cloud data into a global coordinate system and integrate them to obtain an initial map model; performing ground segmentation on the initial map model to obtain multiple ground point cloud data; performing plane fitting on multiple ground point cloud data to obtain the ground, and marking the ground in the initial map model; performing target detection on each image to obtain the detection result of each image, and filtering out images with detection results of non-static objects from multiple images; deleting the point cloud data corresponding to the images with detection results of non-static objects in the initial map model to obtain the final map model. The above technical means can solve the problem that the map model lacks key environmental information and erroneously adds objects that do not belong to the environment to the map model in the prior art, thereby improving the accuracy of the map model. BRIEF DESCRIPTION OF THE DRAWINGS
[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0010] Figure 1 It is a flowchart of a method for constructing a map model provided in an embodiment of the present application;
[0011] Figure 2 It is a flowchart of another map model construction method provided in an embodiment of the present application;
[0012] Figure 3 is a structural schematic diagram of a map model building device provided in an embodiment of the present application;
[0013] Figure 4 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0014] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.
[0015] A map model construction method and device according to an embodiment of the present application will be described in detail below with reference to the accompanying drawings.
[0016] Figure 1 It is a flowchart of a map model construction method provided in an embodiment of the present application. Figure 1 The map model construction method can be executed by a computer or a server, or software on a computer or a server. Figure 1 As shown, the map model construction method includes:
[0017] S101, controlling the mapping mobile device to move along a set route in the target environment, and obtaining a plurality of inertial navigation data of the mapping mobile device, a plurality of point cloud data of the target environment, and a plurality of images during the movement of the mapping mobile device, wherein one point cloud data corresponds to one inertial navigation data and one image;
[0018] S102, processing each point cloud data using a combination algorithm according to the inertial navigation data corresponding to each point cloud data, so as to convert each point cloud data into a global coordinate system and integrate them to obtain an initial map model;
[0019] S103, performing ground segmentation on the initial map model to obtain a plurality of ground point cloud data;
[0020] S104, performing plane fitting on the plurality of ground point cloud data to obtain the ground, and marking the ground in the initial map model;
[0021] S105, performing target detection on each image to obtain detection results of each image, and screening out images whose detection results are non-static objects from the multiple images;
[0022] S106, deleting point cloud data corresponding to images with detection results of non-static objects in the initial map model to obtain a final map model.
[0023] Mobile mapping devices include robots and cars. Inertial navigation data includes the position data, velocity data, acceleration data and angular velocity data of the mobile mapping device. Each point cloud data is converted to a global coordinate system according to the inertial navigation data corresponding to each point cloud data. Taking into account the impact of the operating state of the mobile mapping device on the point cloud data, each point cloud data is converted to a global coordinate system to cover the target environment, which can reduce the lack of key environmental information in the map model. From the initial map model, the ground point cloud data representing the ground is determined, and multiple ground point cloud data are plane-fitted to obtain the ground, and the ground is marked in the initial map model. The plane fitting can use the RANSAC plane fitting algorithm in the PCL (Point Cloud Library) library to fit the extracted ground point cloud data. Target detection is performed on each image to filter out images with detection results of non-static objects from multiple images, and then the point cloud data corresponding to the images with detection results of non-static objects are deleted from the initial map model to avoid mistakenly presenting pedestrians and vehicles in the environment that do not belong to the environment in the map model.
[0024] According to the technical solution provided by the embodiment of the present application, the mapping mobile device is controlled to travel along the set route in the target environment. During the travel of the mapping mobile device, multiple inertial navigation data of the mapping mobile device, multiple point cloud data of the target environment and multiple images are obtained, wherein one point cloud data corresponds to one inertial navigation data and one image; based on the inertial navigation data corresponding to each point cloud data, each point cloud data is processed by a combination algorithm to convert each point cloud data into a global coordinate system and integrate to obtain an initial map model; the initial map model is ground segmented to obtain multiple ground point cloud data; the multiple ground point cloud data are plane fitted to obtain the ground, and the ground is marked in the initial map model; target detection is performed on each image to obtain the detection result of each image, and images with detection results of non-static objects are screened out from multiple images; point cloud data corresponding to images with detection results of non-static objects are deleted from the initial map model to obtain the final map model. The above technical means can solve the problem that the map model lacks key environmental information and erroneously adds objects that do not belong to the environment to the map model in the prior art, thereby improving the accuracy of the map model.
[0025] The embodiments of the present application particularly solve the problem of a map model lacking key environmental information in a small and crowded environment and erroneously adding objects that do not belong to the environment to the map model.
[0026] Furthermore, during the movement of the mapping mobile device, multiple inertial navigation data of the mapping mobile device, multiple point cloud data of the target environment, and multiple images are obtained, including: at a moment in the movement of the mapping mobile device: using a laser scanner provided on the mapping mobile device to obtain the point cloud data of the target environment at that moment; using a camera provided on the mapping mobile device to obtain the image of the target environment at that moment; obtaining the inertial navigation data of the mapping mobile device at that moment, wherein the inertial navigation data includes posture data, velocity data, acceleration data, and angular velocity data of the mapping mobile device.
[0027] One piece of inertial navigation data, one piece of point cloud data of the target environment and one image are obtained at each moment. In this way, by controlling the mapping mobile device to move along the set route in the target environment, multiple pieces of inertial navigation data, multiple pieces of point cloud data of the target environment and multiple images can be obtained.
[0028] Furthermore, based on the inertial navigation data corresponding to each piece of point cloud data, each piece of point cloud data is processed using a combined algorithm to convert each piece of point cloud data into a global coordinate system and integrate them to obtain an initial map model, including: determining a global coordinate system based on the inertial navigation data corresponding to each piece of point cloud data; converting each piece of point cloud data into a global coordinate system to determine a feature point of each piece of point cloud data; and integrating each piece of point cloud data according to its feature point to obtain an initial map model.
[0029] Each piece of point cloud data is in a local coordinate system, and the local coordinate system of each piece of point cloud data may be different. The local coordinate system of the point cloud data corresponding to the inertial navigation data that meets the preset standard can be used as the global coordinate system. This method can reduce the complexity of converting each piece of point cloud data to the global coordinate system. In order to improve the accuracy of integrating each piece of point cloud data, after converting each piece of point cloud data to the global coordinate system, several feature points can be determined from each piece of point cloud data, and then each piece of point cloud data can be integrated according to the feature points of each piece of point cloud data to obtain an initial map model. If two pieces of point cloud data have the same two feature points, the two pieces of point cloud data overlap at the two feature points.
[0030] Furthermore, the initial map model is segmented on the ground to obtain multiple ground point cloud data, including: determining multiple seed points with the lowest height values from the initial map model; creating a plane patch corresponding to each seed point with each seed point as the center; calculating the distance between other points in the initial map model and each plane patch; if the distance between other points and a plane patch is less than a preset value, merging the other points into the plane patch, and using each plane patch finally obtained as a ground point cloud data.
[0031] The preset range centered on each seed point can be used as the plane patch corresponding to the seed point. Other points are points not included in the plane patch.
[0032] Furthermore, the distances between other points in the initial map model and each plane patch are calculated, including: performing principal component analysis on the point cloud data of all points in each plane patch to obtain the fitting equation of each plane patch; and calculating the distances between each plane patch and other points based on the fitting equation of each plane patch and the point cloud data of other points.
[0033] Furthermore, after creating a plane patch corresponding to each seed point with each seed point as the center, the method also includes: for a plane patch, executing the following loop: calculating the distance between other points and the plane patch; merging the points among other points whose distance to the plane patch is less than a preset value into the plane patch; until there are no points among other points whose distance to the plane patch is less than the preset value, then exiting the loop.
[0034] For multiple plane patches, execute the following loop: calculate the distance between other points and each plane patch; merge the points whose distance to each plane patch is less than the preset value into each plane patch; continue to calculate the distance between other points and each plane patch; exit the loop until there are no points whose distance to each plane patch is less than the preset value among other points.
[0035] After each loop is executed, other points are updated, and other points are points that are not included in the plane patch after the update. If the distance between a point and multiple plane patches is less than the preset value, the point will be merged into the plane patch with the closest distance by default.
[0036] Furthermore, target detection is performed on each image to obtain a detection result of each image, and images whose detection results are non-static objects are screened out from multiple images, including: inputting each image into a target detection model to obtain a detection result of each image, wherein the target detection model has been trained to classify objects in each image; marking whether the object in each image is a non-static object according to the detection result of the image; and screening out images whose detection results are non-static objects from multiple images.
[0037] The map model only needs to present static objects such as buildings, roads, and signs, and does not need to present non-static objects. Therefore, the embodiment of the present application requires the target detection model to delete the point cloud data corresponding to the non-static objects from the initial map model of the object. Non-static objects include pedestrians, animals, vehicles, etc. Even a vehicle that has been turned off is considered a non-static object by default. The target detection model can be any commonly used target detection algorithm that can classify objects in an image. The classification result, that is, the detection result, is the category of the object in the image.
[0038] Figure 2 FIG. 1 is a flow chart of another method for constructing a map model provided in an embodiment of the present application. Figure 2 As shown, the method includes:
[0039] S201, determining a global coordinate system based on inertial navigation data corresponding to each point cloud data;
[0040] S202, converting each piece of point cloud data into a global coordinate system, and determining a feature point of each piece of point cloud data;
[0041] S203, integrating each piece of point cloud data according to its characteristic points to obtain an initial map model;
[0042] S204, determining a plurality of seed points with the lowest height values from the initial map model;
[0043] S205, creating a plane patch corresponding to each seed point with each seed point as the center;
[0044] S206, calculating the distances between other points in the initial map model and each plane patch;
[0045] S207, if the distance between other points and a plane patch is less than a preset value, then merge the other points into the plane patch, and use each plane patch finally obtained as a piece of ground point cloud data;
[0046] S208, performing plane fitting on the plurality of ground point cloud data to obtain the ground, and marking the ground in the initial map model;
[0047] S209, performing target detection on each image to obtain detection results of each image, and screening out images whose detection results are non-static objects from the multiple images;
[0048] S210 , deleting point cloud data corresponding to images with detection results of non-static objects in the initial map model to obtain a final map model.
[0049] All the above optional technical solutions can be arbitrarily combined to form optional embodiments of the present application, which will not be described one by one here.
[0050] The following is an embodiment of the device of the present application, which can be used to execute the embodiment of the method of the present application. For details not disclosed in the embodiment of the device of the present application, please refer to the embodiment of the method of the present application.
[0051] Figure 3 Schematic diagram of a map model construction device provided in an embodiment of the present application. Figure 3 As shown, the map model building device includes:
[0052] The acquisition module 301 is configured to control the mapping mobile device to travel along a set route in the target environment, and during the movement of the mapping mobile device, obtain multiple inertial navigation data of the mapping mobile device, multiple point cloud data of the target environment, and multiple images, wherein one point cloud data corresponds to one inertial navigation data and one image;
[0053] The combination module 302 is configured to process each point cloud data using a combination algorithm according to the inertial navigation data corresponding to each point cloud data, so as to convert each point cloud data into a global coordinate system and integrate them to obtain an initial map model;
[0054] The ground segmentation module 303 is configured to perform ground segmentation on the initial map model to obtain a plurality of ground point cloud data;
[0055] A plane fitting module 304 is configured to perform plane fitting on the plurality of ground point cloud data to obtain the ground, and mark the ground in the initial map model;
[0056] The object detection module 305 is configured to perform object detection on each image to obtain detection results of each image, and filter out images whose detection results are non-static objects from multiple images;
[0057] The deletion module 306 is configured to delete the point cloud data corresponding to the image whose detection result is a non-static object in the initial map model to obtain a final map model.
[0058] Mobile mapping devices include robots and cars. Inertial navigation data includes the position data, velocity data, acceleration data and angular velocity data of the mobile mapping device. Each point cloud data is converted to a global coordinate system according to the inertial navigation data corresponding to each point cloud data. Taking into account the impact of the operating state of the mobile mapping device on the point cloud data, each point cloud data is converted to a global coordinate system to cover the target environment, which can reduce the lack of key environmental information in the map model. From the initial map model, the ground point cloud data representing the ground is determined, and multiple ground point cloud data are plane-fitted to obtain the ground, and the ground is marked in the initial map model. The plane fitting can use the RANSAC plane fitting algorithm in the PCL (Point Cloud Library) library to fit the extracted ground point cloud data. Target detection is performed on each image to filter out images with detection results of non-static objects from multiple images, and then the point cloud data corresponding to the images with detection results of non-static objects are deleted from the initial map model to avoid mistakenly presenting pedestrians and vehicles in the environment that do not belong to the environment in the map model.
[0059] According to the technical solution provided by the embodiment of the present application, the mapping mobile device is controlled to travel along the set route in the target environment. During the travel of the mapping mobile device, multiple inertial navigation data of the mapping mobile device, multiple point cloud data of the target environment and multiple images are obtained, wherein one point cloud data corresponds to one inertial navigation data and one image; based on the inertial navigation data corresponding to each point cloud data, each point cloud data is processed by a combination algorithm to convert each point cloud data into a global coordinate system and integrate to obtain an initial map model; the initial map model is ground segmented to obtain multiple ground point cloud data; the multiple ground point cloud data are plane fitted to obtain the ground, and the ground is marked in the initial map model; target detection is performed on each image to obtain the detection result of each image, and images with detection results of non-static objects are screened out from multiple images; point cloud data corresponding to images with detection results of non-static objects are deleted from the initial map model to obtain the final map model. The above technical means can solve the problem that the map model lacks key environmental information and erroneously adds objects that do not belong to the environment to the map model in the prior art, thereby improving the accuracy of the map model.
[0060] In some embodiments, the acquisition module 301 is further configured to, at a moment in the movement of the mapping mobile device: use a laser scanner provided on the mapping mobile device to acquire point cloud data of the target environment at that moment; use a camera provided on the mapping mobile device to acquire an image of the target environment at that moment; and acquire inertial navigation data of the mapping mobile device at that moment, wherein the inertial navigation data includes posture data, velocity data, acceleration data, and angular velocity data of the mapping mobile device.
[0061] One piece of inertial navigation data, one piece of point cloud data of the target environment and one image are obtained at each moment. In this way, by controlling the mapping mobile device to move along the set route in the target environment, multiple pieces of inertial navigation data, multiple pieces of point cloud data of the target environment and multiple images can be obtained.
[0062] In some embodiments, the combination module 302 is also configured to determine a global coordinate system based on the inertial navigation data corresponding to each point cloud data; convert each point cloud data into the global coordinate system to determine the feature points of each point cloud data; integrate each point cloud data according to the feature points of each point cloud data to obtain an initial map model.
[0063] Each piece of point cloud data is in a local coordinate system, and the local coordinate system of each piece of point cloud data may be different. The local coordinate system of the point cloud data corresponding to the inertial navigation data that meets the preset standard can be used as the global coordinate system. This method can reduce the complexity of converting each piece of point cloud data to the global coordinate system. In order to improve the accuracy of integrating each piece of point cloud data, after converting each piece of point cloud data to the global coordinate system, several feature points can be determined from each piece of point cloud data, and then each piece of point cloud data can be integrated according to the feature points of each piece of point cloud data to obtain an initial map model. If two pieces of point cloud data have the same two feature points, the two pieces of point cloud data overlap at the two feature points.
[0064] In some embodiments, the ground segmentation module 303 is also configured to determine multiple seed points with the lowest height values from the initial map model; create a plane patch corresponding to each seed point with the seed point as the center; calculate the distance between other points in the initial map model and each plane patch; if the distance between other points and a plane patch is less than a preset value, merge the other points into the plane patch, and use the final plane patches as a ground point cloud data.
[0065] The preset range centered on each seed point can be used as the plane patch corresponding to the seed point. Other points are points not included in the plane patch.
[0066] In some embodiments, the ground segmentation module 303 is also configured to perform principal component analysis on the point cloud data of all points in each plane patch to obtain a fitting equation for each plane patch; and calculate the distance between each plane patch and other points based on the fitting equation of each plane patch and the point cloud data of other points.
[0067] In some embodiments, the ground segmentation module 303 is also configured to execute the following loop for a plane patch: calculate the distance between other points and the plane patch; merge the points among other points whose distance to the plane patch is less than a preset value into the plane patch; and exit the loop until there are no points among other points whose distance to the plane patch is less than a preset value.
[0068] For multiple plane patches, execute the following loop: calculate the distance between other points and each plane patch; merge the points whose distance to each plane patch is less than the preset value into each plane patch; continue to calculate the distance between other points and each plane patch; exit the loop until there are no points whose distance to each plane patch is less than the preset value among other points.
[0069] After each loop is executed, other points are updated, and other points are points that are not included in the plane patch after the update. If the distance between a point and multiple plane patches is less than the preset value, the point will be merged into the plane patch with the closest distance by default.
[0070] In some embodiments, the target detection module 305 is further configured to input each image into a target detection model to obtain a detection result for each image, wherein the target detection model has been trained to classify objects in each image; mark whether the object in the image is a non-static object according to the detection result of each image; and filter out images whose detection results are non-static objects from multiple images.
[0071] The map model only needs to present static objects such as buildings, roads, and signs, and does not need to present non-static objects. Therefore, the embodiment of the present application requires the target detection model to delete the point cloud data corresponding to the non-static objects from the initial map model of the object. Non-static objects include pedestrians, animals, vehicles, etc. Even a vehicle that has been turned off is considered a non-static object by default. The target detection model can be any commonly used target detection algorithm that can classify objects in an image. The classification result, that is, the detection result, is the category of the object in the image.
[0072] In some embodiments, the deletion module 306 is further configured to determine a global coordinate system based on the inertial navigation data corresponding to each point cloud data; convert each point cloud data into a global coordinate system and determine the feature points of each point cloud data; integrate each point cloud data according to the feature points of each point cloud data to obtain an initial map model; determine multiple seed points with the lowest height values from the initial map model; create a plane patch corresponding to each seed point with each seed point as the center; calculate the distance between other points in the initial map model and each plane patch; if the distance between other points and a plane patch is less than a preset value, merge the other points into the plane patch, and use the final plane patches as a ground point cloud data; perform plane fitting on multiple ground point cloud data to obtain the ground, and mark the ground in the initial map model; perform target detection on each image to obtain the detection results of each image, and filter out images with detection results of non-static objects from multiple images; delete the point cloud data corresponding to the images with detection results of non-static objects in the initial map model to obtain the final map model.
[0073] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.
[0074] Figure 4 Schematic diagram of an electronic device 4 provided in an embodiment of the present application. Figure 4As shown, the electronic device 4 of this embodiment includes: a processor 401, a memory 402, and a computer program 403 stored in the memory 402 and executable on the processor 401. When the processor 401 executes the computer program 403, the steps in the above-mentioned method embodiments are implemented. Alternatively, when the processor 401 executes the computer program 403, the functions of the modules / units in the above-mentioned device embodiments are implemented.
[0075] The electronic device 4 may be a desktop computer, a notebook, a PDA, a cloud server, or other electronic device. The electronic device 4 may include, but is not limited to, a processor 401 and a memory 402. Those skilled in the art will appreciate that Figure 4 The electronic device 4 is merely an example and does not limit the electronic device 4 , and may include more or less components than those shown in the figure, or different components.
[0076] The processor 401 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.
[0077] The memory 402 may be an internal storage unit of the electronic device 4, for example, a hard disk or memory of the electronic device 4. The memory 402 may also be an external storage device of the electronic device 4, for example, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device 4. The memory 402 may also include both an internal storage unit and an external storage device of the electronic device 4. The memory 402 is used to store computer programs and other programs and data required by the electronic device.
[0078] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units.
[0079] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. The computer program may include computer program code, and the computer program code may be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electric carrier signals and telecommunication signals.
[0080] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.
Claims
1. A method for constructing a map model, characterized in that: include: Controlling the mapping mobile device to move along a set route in the target environment, and obtaining a plurality of inertial navigation data of the mapping mobile device, a plurality of point cloud data of the target environment, and a plurality of images during the movement of the mapping mobile device, wherein one point cloud data corresponds to one inertial navigation data and one image; According to the inertial navigation data corresponding to each point cloud data, each point cloud data is processed by using a combination algorithm to convert each point cloud data into a global coordinate system and integrate them to obtain an initial map model; Performing ground segmentation on the initial map model to obtain a plurality of ground point cloud data; Performing plane fitting on multiple ground point cloud data to obtain ground, and marking the ground in the initial map model; Performing target detection on each image to obtain detection results of each image, and filtering out images whose detection results are non-static objects from multiple images; The point cloud data corresponding to the image whose detection result is a non-static object is deleted from the initial map model to obtain a final map model.
2. The method according to claim 1, characterized in that During the movement of the mapping mobile device, a plurality of inertial navigation data of the mapping mobile device, a plurality of point cloud data of the target environment and a plurality of images are acquired, including: At a moment in the process of the mobile mapping device moving: Using a laser scanner provided on the mobile mapping device to acquire point cloud data of the target environment at that moment; Using a camera provided on the mobile mapping device to acquire an image of the target environment at that moment; Acquire inertial navigation data of the mapping mobile device at this moment, wherein the inertial navigation data includes position data, velocity data, acceleration data, and angular velocity data of the mapping mobile device.
3. The method according to claim 1, characterized in that According to the inertial navigation data corresponding to each point cloud data, each point cloud data is processed by a combination algorithm to convert each point cloud data into a global coordinate system and integrate them to obtain an initial map model, including: Determine the global coordinate system based on the inertial navigation data corresponding to each point cloud data; Convert each point cloud data into a global coordinate system and determine the characteristic points of each point cloud data; The individual point cloud data are integrated according to their feature points to obtain the initial map model.
4. The method according to claim 1, characterized in that The initial map model is segmented to obtain a plurality of ground point cloud data, including: Determine a plurality of seed points with the lowest height values from the initial map model; Create a plane patch corresponding to each seed point with each seed point as the center; Calculating the distances between other points in the initial map model and each plane patch; If the distance between other points and a plane patch is less than a preset value, the other points are merged into the plane patch, and the plane patches finally obtained are used as a piece of ground point cloud data.
5. The method according to claim 4, characterized in that Calculating the distances between other points in the initial map model and each plane patch includes: Perform principal component analysis on the point cloud data of all points in each plane patch to obtain the fitting equation of each plane patch; According to the fitting equation of each plane patch and the point cloud data of other points, the distance between each plane patch and other points is calculated.
6. The method according to claim 4, characterized in that After creating a plane patch corresponding to each seed point with each seed point as the center, the method further includes: For a flat patch, the following loop is performed: Calculate the distances between other points and the plane patch; Merge points whose distance from the plane patch is less than a preset value among other points into the plane patch; The loop is exited until there is no point in other points whose distance from the plane patch is less than a preset value.
7. The method according to claim 1, characterized in that Perform target detection on each image to obtain the detection results of each image, and filter out images whose detection results are non-static objects from multiple images, including: Inputting each image into a target detection model to obtain a detection result for each image, wherein the target detection model has been trained to classify objects in each image; Mark whether an object in each image is a non-static object according to the detection result of the image; Filter out images with detection results of non-static objects from multiple images.
8. A map model construction device, characterized in that: include: an acquisition module, configured to control the mapping mobile device to travel along a set route in the target environment, and to acquire a plurality of inertial navigation data of the mapping mobile device, a plurality of point cloud data of the target environment, and a plurality of images during the movement of the mapping mobile device, wherein one point cloud data corresponds to one inertial navigation data and one image; The combination module is configured to process each point cloud data by using a combination algorithm according to the inertial navigation data corresponding to each point cloud data, so as to convert each point cloud data into a global coordinate system and integrate the data to obtain an initial map model; A ground segmentation module is configured to perform ground segmentation on the initial map model to obtain a plurality of ground point cloud data; A plane fitting module is configured to perform plane fitting on a plurality of ground point cloud data to obtain a ground surface, and mark the ground surface in the initial map model; The target detection module is configured to perform target detection on each image to obtain detection results of each image, and filter out images whose detection results are non-static objects from multiple images; The deletion module is configured to delete the point cloud data corresponding to the image with the detection result of a non-static object in the initial map model to obtain a final map model.
9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.
10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.