Ground object identification method and device, storage medium and electronic equipment

Abstract

This invention provides a method, apparatus, storage medium, and electronic device for ground object recognition. The method involves performing planar detection on a 3D point cloud image to obtain multiple 3D planar objects; clustering the pixel grayscale values ​​of the projected image of the 3D point cloud image to obtain multiple 2D grayscale datasets in the projected image; identifying at least one non-ground object among the multiple 3D planar objects based on these datasets; and filtering out non-ground planar objects from the multiple 3D planar objects to obtain the ground objects. This method combines 2D grayscale data with 3D planar objects, further subdividing the 3D planar objects based on the 2D grayscale data to exclude non-ground objects, thereby achieving ground object recognition and improving the accuracy of ground object recognition.

Description

Technical Field

[0001] This application relates to the field of virtual reality technology, and in particular to a ground object recognition method, device, storage medium and electronic device. Background Technology

[0002] Virtual Reality (VR) technology is a combination of simulation technology and various other technologies, including computer graphics, human-computer interface technology, multimedia technology, sensing technology, and network technology. VR technology simulates virtual environments using computers, thus creating a sense of immersion. This technology is widely used across various industries.

[0003] In the initial stage of VR tracking and positioning, ground recognition is required to determine the ground height. In existing technologies, three-dimensional planar recognition technology is usually used to detect planes within the field of view. Since there may be multiple planes within the field of view, when determining the ground from multiple planes, some non-ground planes, such as tabletops, chair surfaces, and carpets, may be identified as the ground, resulting in low accuracy of ground recognition. Summary of the Invention

[0004] In view of the above problems, embodiments of this application are proposed, which provide a ground object recognition method, apparatus, storage medium and electronic device to at least solve the above problems.

[0005] One or more embodiments of this application provide a ground object recognition method, including: performing plane detection on a three-dimensional point cloud image to obtain multiple three-dimensional plane objects; clustering the pixel grayscale values ​​of the projection image of the three-dimensional point cloud image to obtain multiple two-dimensional grayscale datasets in the projection image; determining at least one non-ground object among the multiple three-dimensional plane objects based on the multiple two-dimensional grayscale datasets; and filtering out non-ground plane objects from the multiple three-dimensional plane objects to obtain ground objects.

[0006] Optionally, based on multiple two-dimensional grayscale datasets, at least one non-ground object among multiple three-dimensional planar objects is determined, including: establishing a mapping relationship between multiple two-dimensional grayscale datasets and multiple three-dimensional planar objects; based on the mapping relationship, determining the grayscale regions mapped from the two-dimensional grayscale datasets to the three-dimensional planar objects; and determining the three-dimensional planar objects corresponding to grayscale regions with an area smaller than a preset threshold as non-ground objects.

[0007] Optionally, plane detection is performed on the 3D point cloud image to obtain multiple 3D planar objects, including: performing plane detection on the 3D point cloud image to obtain multiple initial 3D planar objects; performing coordinate system transformation based on pose data and the initial 3D planar objects to obtain a first 3D planar object; determining the first 3D planar object whose angle is less than a preset angle threshold as a second 3D planar object based on the angle between the plane normal vector of the first 3D planar object and the horizontal plane normal vector; and determining the second 3D planar object as a 3D planar object.

[0008] Optionally, determining the second three-dimensional plane object as a three-dimensional plane object includes: determining the second three-dimensional plane object with a height less than the preset height threshold as a three-dimensional plane object based on the second three-dimensional plane object and the preset height threshold.

[0009] Optionally, the method further includes: acquiring velocity data; if the velocity data is less than a preset velocity threshold, then acquiring a three-dimensional point cloud image.

[0010] Optionally, the method further includes: acquiring angle data; if the angle data is greater than a preset angle threshold, then acquiring a three-dimensional point cloud image.

[0011] Optionally, the method further includes: identifying non-ground objects according to preset identification rules.

[0012] According to another aspect of this application, a ground object recognition device is provided, including a detection unit, a calculation unit, a filtering unit, and an output unit. The detection unit is used to perform planar detection on a three-dimensional point cloud image to obtain multiple three-dimensional planar objects. The calculation unit is used to cluster the pixel grayscale values ​​of the projection image of the three-dimensional point cloud image to obtain multiple two-dimensional grayscale datasets in the projection image. The filtering unit is used to determine at least one non-ground object among the multiple three-dimensional planar objects based on the multiple two-dimensional grayscale datasets. The output unit is used to filter out non-ground planar objects from the multiple three-dimensional planar objects to obtain ground objects.

[0013] According to another aspect of this application, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause a computer to perform the methods described above.

[0014] According to another aspect of this application, an electronic device is provided, comprising: one or more processors; and a memory storing a program; wherein the program includes instructions that, when executed by the processor, cause the processor to perform the methods described above.

[0015] This application provides a ground object recognition method that obtains multiple 3D planar objects by performing planar detection on a 3D point cloud image; clustering the pixel grayscale values ​​of the projected image of the 3D point cloud image to obtain multiple 2D grayscale datasets in the projected image; identifying at least one non-ground object among the multiple 3D planar objects based on the multiple 2D grayscale datasets; and filtering out non-ground planar objects from the multiple 3D planar objects to obtain ground objects. This method combines 2D grayscale datasets with 3D planar objects, further subdividing the 3D planar objects based on the 2D grayscale datasets to exclude non-ground planes, thereby achieving ground recognition and improving the accuracy of ground recognition. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A flowchart illustrating a ground object recognition method as an exemplary embodiment of this application;

[0018] Figure 2 A flowchart illustrating a ground object recognition method as another exemplary embodiment of this application;

[0019] Figure 3 A flowchart illustrating a ground object recognition method as another exemplary embodiment of this application;

[0020] Figure 4 A structural block diagram of a ground object recognition device as an exemplary embodiment of this application;

[0021] Figure 5 A structural block diagram of an electronic device that is an exemplary embodiment of this application;

[0022] Explanation of reference numerals in the attached figures:

[0023] 400. Ground object recognition device; 401. Detection unit; 402. Calculation unit; 403. Filtering unit; 404. Output unit; 500. Electronic device; 501. Calculation unit; 502. ROM; 503. RAM; 504. Bus; 505. Input / output interface; 506. Input unit; 507. Output unit; 508. Storage unit; 509. Communication unit. Detailed Implementation

[0024] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present application. All other embodiments obtained by those skilled in the art based on the embodiments of the present application without creative effort are within the scope of protection of the present application.

[0025] To facilitate understanding, before describing the specific embodiments of this application in detail, the application scenarios of the ground object recognition method, device, storage medium and electronic device of this application will be illustrated by example.

[0026] The ground object recognition method of this application can be applied to virtual reality head-mounted display devices that employ virtual reality technology. Virtual reality (VR) technology is a combination of simulation technology and various other technologies such as computer graphics, human-computer interface technology, multimedia technology, sensing technology, and network technology. VR technology simulates a virtual environment through computers, thereby providing a sense of immersion. This technology is widely used in various industries.

[0027] In the initial stage of VR tracking and positioning, ground recognition is required to determine ground height. Existing technologies typically rely on 3D planar recognition to detect planes within the field of view. However, since multiple planes may exist within the field of view, determining the ground from these multiple planes may result in misidentifying non-ground surfaces such as tabletops, chair surfaces, and carpets as the ground, leading to low accuracy in ground recognition. Therefore, this application proposes a ground object recognition method, apparatus, storage medium, and electronic device to solve the various problems existing in the prior art.

[0028] The specific embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0029] Figure 1 The flowchart of a ground object recognition method, which is an exemplary embodiment of this application, is shown in the figure. This embodiment mainly includes the following steps:

[0030] S101. Perform plane detection on the 3D point cloud image to obtain multiple 3D plane objects.

[0031] For example, a 3D point cloud image can be obtained by performing coordinate transformation calculations on a depth image captured by a virtual reality head-mounted display device. The point cloud size can be defined by its height and width. When performing planar detection on a 3D point cloud image, the points in the image can be uniformly divided into A×B nodes. For example, if the 3D point cloud image has a height of 10×width of 10 and A=B=2, it can be divided into (10 / 2)×(10 / 2)=25 nodes. Nodes lacking data, nodes with discontinuous depth, and edges corresponding to these nodes can be removed. A hierarchical clustering algorithm can be used to cluster the remaining nodes to obtain multiple segmented nodes. Further optimization can be performed based on pixels. Multiple 3D planar objects can be extracted from the optimized segmented nodes. A 3D planar object can include the number of points, point coordinates, and plane normal vectors.

[0032] S102. Cluster the pixel grayscale values ​​of the projected image of the 3D point cloud image to obtain multiple 2D grayscale datasets in the projected image.

[0033] For example, the projected image of a 3D point cloud image can be a grayscale image or a color image. Grayscale refers to the color depth of a point in a black and white image, typically ranging from 0 to 255, where white is 255 and black is 0. The grayscale of a color image can be obtained by converting it to the pixel values ​​of a black and white image. Clustering methods can use K-means clustering or hierarchical clustering algorithms. Multiple two-dimensional grayscale datasets can be obtained through clustering algorithms. These datasets can include the number of pixels, pixel coordinates, grayscale values, etc.

[0034] S103. Based on multiple two-dimensional grayscale datasets, identify at least one non-ground object among multiple three-dimensional planar objects.

[0035] For example, by mapping multiple two-dimensional grayscale datasets to multiple three-dimensional planar objects, the grayscale values ​​of multiple three-dimensional planar objects can be obtained. Based on the grayscale value of each three-dimensional planar object and the number of points it contains, a point count threshold can be set to filter the multiple three-dimensional planar objects and determine non-ground objects. For example, the filtering condition can be that three-dimensional planar objects with fewer points than the point count threshold are considered non-ground objects, such as desktops and chair surfaces; or, if the number of points corresponding to a certain grayscale value in a three-dimensional planar object is less than the point count threshold, then a portion of the three-dimensional planar object composed of those points can be determined as a non-ground object, such as carpets or books close to the ground.

[0036] S104. From multiple three-dimensional planar objects, filter out non-ground planar objects to obtain ground objects.

[0037] For example, from multiple three-dimensional planar objects, non-ground planar objects can be removed, such as desktops, chair surfaces, carpets, books, etc. that are close to the ground. By removing non-ground planar objects, the remaining three-dimensional planar objects can be obtained, and the remaining three-dimensional planar objects can be determined as ground objects.

[0038] This embodiment combines a two-dimensional grayscale dataset with three-dimensional planar objects, further subdivides the three-dimensional planar objects based on the two-dimensional grayscale dataset, excludes non-ground planes, and achieves ground object recognition, thereby improving the accuracy of ground object recognition.

[0039] Figure 2 This is a flowchart illustrating a ground object recognition method as another exemplary embodiment of this application. This embodiment mainly shows a specific implementation of step S103 described above. As shown in the figure, this embodiment mainly includes the following steps:

[0040] S201. Establish the mapping relationship between multiple two-dimensional grayscale datasets and multiple three-dimensional planar objects.

[0041] S202. Based on the mapping relationship, determine the grayscale region in the three-dimensional planar object that the two-dimensional grayscale dataset is mapped to.

[0042] S203. Determine that the three-dimensional planar object corresponding to the grayscale area with an area smaller than the preset threshold is a non-ground object.

[0043] For example, the mapping relationship between the two-dimensional grayscale dataset and the three-dimensional planar object can be obtained by mapping the coordinates of pixels in the two-dimensional grayscale dataset to the coordinates of points in the three-dimensional planar object. Based on this mapping relationship, the grayscale region corresponding to the two-dimensional grayscale dataset in the three-dimensional planar object can be obtained. The grayscale region may include the number of points and grayscale values. The area of ​​the grayscale region can be determined by the number of points in the grayscale region. If the number of points in the grayscale region is less than a preset threshold, the three-dimensional planar object corresponding to that grayscale region is determined to be a non-ground object.

[0044] This embodiment establishes a mapping relationship between a two-dimensional grayscale dataset and a three-dimensional planar object, thereby mapping the two-dimensional grayscale dataset to the three-dimensional planar object and obtaining the grayscale region corresponding to the two-dimensional grayscale dataset in the three-dimensional planar object. By subdividing the three-dimensional planar object through the grayscale region, non-ground objects are filtered out, thereby improving the accuracy of ground object recognition.

[0045] Figure 3 This is a flowchart illustrating a ground object recognition method as another exemplary embodiment of this application. This embodiment mainly shows a specific implementation of step S101 described above. As shown in the figure, this embodiment mainly includes the following steps:

[0046] S301. Perform plane detection on the 3D point cloud image to obtain multiple initial 3D plane objects.

[0047] S302. Based on the posture data and the initial three-dimensional plane object, perform coordinate system transformation to obtain the first three-dimensional plane object.

[0048] S303. Based on the angle between the plane normal vector of the first three-dimensional plane object and the horizontal plane normal vector, determine the first three-dimensional plane object whose angle is less than a preset angle threshold as the second three-dimensional plane object.

[0049] S304. Determine the second three-dimensional planar object as a three-dimensional planar object.

[0050] For example, a virtual reality head-mounted display device may be equipped with an IMU sensor (Inertial Measurement Unit). The main components of the IMU sensor include a gyroscope, an accelerometer, and a magnetometer. The gyroscope obtains the angular velocity of each axis, the accelerometer obtains the acceleration in the x, y, and z directions, and the magnetometer obtains information about the surrounding magnetic field. Based on the information obtained from the IMU sensor, attitude data is obtained through an attitude fusion algorithm. The attitude data includes the three-dimensional coordinate data of the camera in the virtual reality head-mounted display device. A coordinate system transformation is performed based on the attitude data and an initial three-dimensional planar object, converting the coordinate system of the initial three-dimensional planar object to the world coordinate system to obtain the first three-dimensional planar object.

[0051] Specifically, the angle between the plane normal vector of the first 3D planar object and the horizontal plane normal vector is calculated. The angle is then compared to a preset angle threshold, which can be 10 degrees. If the angle is less than the preset angle threshold, the corresponding first 3D planar object can be used as the second 3D planar object. If the angle is greater than or equal to the preset angle threshold, the corresponding first 3D planar object is discarded. The second 3D planar object is then identified as the 3D planar object in S101 for subsequent processing.

[0052] This embodiment determines the angle between the plane normal vector of the first three-dimensional plane object and the horizontal plane normal vector, and compares the angle with a preset angle threshold. The first three-dimensional plane object with an angle smaller than the preset angle threshold is retained as the second three-dimensional plane object, thereby eliminating the first three-dimensional plane with a larger angle, that is, eliminating some tilted planes. This reduces the amount of data processing for subsequent combination of two-dimensional grayscale datasets and three-dimensional plane objects, thereby improving the efficiency and accuracy of ground object recognition.

[0053] In one specific implementation, a second three-dimensional plane object with a height less than the preset height threshold is determined as a three-dimensional plane object based on the second three-dimensional plane object and the preset height threshold.

[0054] For example, the second three-dimensional plane object can be further filtered by setting a preset height threshold and comparing the height of the second three-dimensional plane object with the preset height threshold. The preset height threshold can be set to 10 centimeters. Second three-dimensional plane objects with a height less than the preset height threshold are retained as three-dimensional plane objects, while second three-dimensional plane objects with a height greater than or equal to the preset height threshold are discarded.

[0055] This embodiment removes second-dimensional planar objects with a height greater than or equal to a preset height threshold by setting a height threshold. This eliminates high-height planes such as desktops and chair surfaces. This reduces data processing volume for subsequent combination of 2D grayscale datasets and 3D planar objects, further improving the efficiency and accuracy of ground object recognition.

[0056] In one specific implementation, the method further includes a specific implementation scheme for acquiring a three-dimensional point cloud image: acquiring velocity data; if the velocity data is less than a preset velocity threshold, then acquiring a three-dimensional point cloud image.

[0057] For example, angular velocity information can be obtained from the gyroscope in the IMU sensor as velocity data. The velocity data is compared with a preset velocity threshold. If the velocity data is less than the preset velocity threshold, a depth image is captured by a virtual reality head-mounted display device to obtain a three-dimensional point cloud image.

[0058] This embodiment compares the motion speed data of the virtual reality head-mounted display device with a preset speed threshold, enabling the virtual reality head-mounted display device to capture images at lower motion speeds, thereby making the acquired 3D point cloud image data more reliable and accurate.

[0059] In one specific implementation, the method further includes an exemplary implementation for acquiring a three-dimensional point cloud image: acquiring angle data, and if the angle data is greater than a preset angle threshold, acquiring a three-dimensional point cloud image.

[0060] For example, the posture data of the virtual reality head-mounted display device can be obtained through the IMU sensor. Based on the posture data and the position data of the camera on the virtual reality head-mounted display device, the angle between the camera and the horizontal plane can be calculated as angle data. A preset angle threshold of 30 degrees can be set. If the angle data is greater than the preset angle threshold, it means that the downward rotation angle of the camera is greater than 30 degrees, and a depth image of the target field of view can be captured to obtain a three-dimensional point cloud image.

[0061] This embodiment compares the angle of the camera in the virtual reality head-mounted display device with a preset angle threshold, enabling the virtual reality head-mounted display device with the camera in front to accurately obtain the three-dimensional point cloud image of the target field of view, thereby avoiding the situation where the acquired three-dimensional point cloud image data is insufficient and ground object recognition cannot be performed.

[0062] In one specific implementation, the ground object recognition method of this application may further include a subsequent implementation scheme after S104: identifying non-ground objects according to a preset identification rule.

[0063] For example, non-ground objects may be desktops, chair surfaces, and carpets, books, etc. that are close to the ground. The preset marking rules may be to use different colors to mark non-ground objects according to their height. For example, non-ground objects with a height of 10 centimeters or more can be marked orange, while non-ground objects with a height of less than 10 centimeters can be marked red.

[0064] This embodiment identifies non-ground objects, enabling users to clearly distinguish them while using virtual reality headsets. This is especially true for non-ground objects that may affect user safety, such as carpets and books that are close to the ground, thereby reducing safety risks during use.

[0065] Figure 4 This is a structural block diagram of a ground object recognition device as an exemplary embodiment of this application.

[0066] The ground object recognition device 400 of this embodiment can be installed in a virtual reality head-mounted display device, wherein the virtual reality head-mounted display device can be used to perform ground object recognition tasks.

[0067] As shown in the figure, the ground object recognition device 400 of this embodiment mainly includes: a detection unit 401, a calculation unit 402, a filtering unit 403 and an output unit 404.

[0068] The detection unit 401 is used to perform plane detection on the 3D point cloud image to obtain multiple 3D plane objects; the calculation unit 402 is used to cluster the pixel grayscale of the projection image of the 3D point cloud image to obtain multiple 2D grayscale datasets in the projection image; the filtering unit 403 is used to determine at least one non-ground object among the multiple 3D plane objects based on the multiple 2D grayscale datasets; and the output unit 404 is used to filter out non-ground plane objects from the multiple 3D plane objects to obtain ground objects.

[0069] This embodiment combines a two-dimensional grayscale dataset with three-dimensional planar objects, further subdivides the three-dimensional planar objects based on the two-dimensional grayscale dataset, excludes non-ground planes, and achieves ground object recognition, thereby improving the accuracy of ground object recognition.

[0070] Furthermore, the ground object recognition device 400 of this application embodiment can also be used to implement other steps in the aforementioned ground object recognition method embodiments, and has the beneficial effects of the corresponding method step embodiments, which will not be repeated here.

[0071] An exemplary embodiment of this application also provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to perform the methods of various embodiments of this application.

[0072] An exemplary embodiment of this application also provides an electronic device, including: one or more processors; and a memory storing a program; wherein the program includes instructions that, when executed by the processor, cause the processor to perform the methods of various embodiments of this application.

[0073] refer to Figure 5 The present invention describes a structural block diagram of an electronic device 500 that can serve as a server or client of this application, which is an example of a hardware device that can be applied to various aspects of this application. The electronic device is intended to represent various forms of digital electronic computer devices, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the application described and / or claimed herein.

[0074] like Figure 5 As shown, the electronic device 500 includes a computing unit 501, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 502 or a computer program loaded from a storage unit 508 into a random access memory (RAM) 503. The RAM 503 may also store various programs and data required for the operation of the device 500. The computing unit 501, ROM 502, and RAM 503 are interconnected via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0075] Multiple components in electronic device 500 are connected to I / O interface 505, including: input unit 506, output unit 507, storage unit 508, and communication unit 509. Input unit 506 can be any type of device capable of inputting information to electronic device 500. Input unit 506 can receive input digital or character information and generate key signal inputs related to user settings and / or function control of electronic device. Output unit 507 can be any type of device capable of presenting information and may include, but is not limited to, a display, speaker, video / audio output terminal, vibrator, and / or printer. Storage unit 504 may include, but is not limited to, disks and optical discs. Communication unit 509 allows electronic device 500 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks, and may include, but is not limited to, modems, network cards, infrared communication devices, wireless communication transceivers, and / or chipsets, such as Bluetooth™ devices, WiFi devices, WiMax devices, cellular communication devices, and / or the like.

[0076] The computing unit 501 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 501 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 501 performs the various methods and processes described above. For example, in some embodiments, the methods of the foregoing embodiments can be implemented as computer software programs tangibly contained in a machine-readable medium, such as storage unit 508. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 500 via ROM 502 and / or communication unit 509. In some embodiments, the computing unit 501 can be configured to perform the foregoing methods by any other suitable means (e.g., by means of firmware).

[0077] The program code used to implement the methods of this application may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0078] It should be noted that in the description of this application, the terms "first" and "second" are used only for convenience in describing different components or names, and should not be construed as indicating or implying a sequential relationship, relative importance, or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first" or "second" may explicitly or implicitly include at least one of those features.

[0079] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application.

[0080] It should be noted that although specific embodiments of this application have been described in detail with reference to the accompanying drawings, this should not be construed as limiting the scope of protection of this application. Various modifications and variations that can be made by those skilled in the art without inventive effort within the scope described in the claims still fall within the scope of protection of this application.

[0081] The examples of the embodiments in this application are intended to concisely illustrate the technical features of the embodiments in this application, so that those skilled in the art can intuitively understand the technical features of the embodiments in this application, and are not intended to be improper limitations on the embodiments in this application.

[0082] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for recognizing ground objects, characterized in that, include: Perform plane detection on a 3D point cloud image to obtain multiple 3D planar objects; Clustering the pixel grayscale values ​​of the projected image of the 3D point cloud image yields multiple 2D grayscale datasets in the projected image. Based on multiple two-dimensional grayscale datasets, at least one non-ground object is identified from among multiple three-dimensional planar objects; From the plurality of three-dimensional planar objects, the non-ground objects are filtered out to obtain ground objects; The step of determining at least one non-ground object among the plurality of three-dimensional planar objects based on the two-dimensional grayscale data includes: Establish a mapping relationship between the multiple two-dimensional grayscale datasets and the multiple three-dimensional planar objects; Based on the mapping relationship, the grayscale region in the three-dimensional planar object is determined to be mapped from the two-dimensional grayscale data. The three-dimensional planar object corresponding to the grayscale region with an area smaller than a preset threshold is determined to be a non-ground object.

2. The method according to claim 1, characterized in that, The process of performing planar detection on a 3D point cloud image yields multiple 3D planar objects, including: Plane detection is performed on the 3D point cloud image to obtain multiple initial 3D planar objects; Based on the posture data and the initial three-dimensional planar object, a coordinate system transformation is performed to obtain the first three-dimensional planar object; Based on the angle between the plane normal vector and the horizontal plane normal vector of the first three-dimensional plane object, the first three-dimensional plane object with an angle smaller than a preset angle threshold is determined as the second three-dimensional plane object; The second three-dimensional planar object is identified as the three-dimensional planar object.

3. The method according to claim 2, characterized in that, The step of determining the second three-dimensional planar object as the three-dimensional planar object includes: Based on the second three-dimensional planar object and the preset height threshold, the second three-dimensional planar object with a height less than the preset height threshold is determined as the three-dimensional planar object.

4. The method according to claim 1, characterized in that, The method further includes: Acquire velocity data; if the velocity data is less than a preset velocity threshold, then acquire the three-dimensional point cloud image.

5. The method according to claim 4, characterized in that, The method further includes: Obtain angle data; if the angle data is greater than a preset angle threshold, then obtain the three-dimensional point cloud image.

6. The method according to claim 1, characterized in that, The method further includes: The non-ground objects are identified according to preset identification rules.

7. A ground object recognition device, characterized in that, include: The detection unit is used to perform planar detection on the 3D point cloud image to obtain multiple 3D planar objects; The computing unit is used to cluster the pixel grayscale values ​​of the projected image of the three-dimensional point cloud image to obtain multiple two-dimensional grayscale datasets in the projected image. A filtering unit is used to determine at least one non-ground object among a plurality of three-dimensional planar objects based on a plurality of the two-dimensional grayscale datasets; The output unit is used to filter out the non-ground objects from the plurality of three-dimensional planar objects to obtain ground objects; Specifically, the filtering unit is used for: Establish a mapping relationship between the multiple two-dimensional grayscale datasets and the multiple three-dimensional planar objects; Based on the mapping relationship, the grayscale region in the three-dimensional planar object is determined to be mapped from the two-dimensional grayscale data. The three-dimensional planar object corresponding to the grayscale region with an area smaller than a preset threshold is determined to be a non-ground object.

8. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-6.

9. An electronic device, characterized in that, include: One or more processors; as well as Memory for stored programs; The program includes instructions that, when executed by the processor, cause the processor to perform the method of any one of claims 1-6.

Images