Pose correction method, automatic walking equipment and computer storage medium

Abstract

The invention relates to the field of automatic walking equipment, and provides a driving control method, automatic walking equipment and a computer storage medium, the method comprises the following steps: in response to a control instruction of a user, starting a working mode of the automatic walking equipment; acquiring a boundary image corresponding to the boundary position of the automatic walking equipment in the target working area; constructing boundary sub-maps corresponding to the at least two different boundary positions according to the boundary image; according to the boundary sub-map, checking image data acquired by the automatic walking equipment to obtain a first image pose; and according to the first image pose, performing error correction on a historical pose of the automatic walking equipment to obtain a target pose.

Description

Technical Field

[0001] The present application relates to the field of automatic walking devices, and particularly to a pose correction method for an automatic walking device, an automatic walking device, and a computer storage medium. Background Art

[0002] In order to reduce the labor intensity and cost of lawn maintenance, automatic walking devices for cutting and trimming lawns, such as lawn mowers, have emerged. Currently, when some lawn mowers are driving and operating on the lawn, the vehicle position is mainly estimated in real time through a VIO (visual-inertial odometry) system, and at the same time, a surrounding environment map of the automatic walking device is constructed to achieve simultaneous localization and mapping. However, the existing VIO scheme requires a large amount of image data for map construction, and the automatic walking device needs to search through a large amount of image data when performing pose correction, resulting in low pose correction efficiency and accuracy. Summary of the Invention

[0003] The main purpose of the present application is to provide a pose correction method for an automatic walking device, an automatic walking device, and a computer storage medium, aiming to improve the accuracy of pose correction of the automatic walking device.

[0004] In a first aspect, the present application provides a pose correction method for an automatic walking device, and the pose correction method for the automatic walking device includes the following steps: Respond to a user's control instruction to start the working mode of the automatic walking device; Obtain a boundary image corresponding to a boundary position of the automatic walking device in a target working area; Construct boundary submaps corresponding to at least two different boundary positions according to the boundary image; Verify the image data obtained by the automatic walking device according to the boundary submaps to obtain a first image pose; Correct the error of the historical pose of the automatic walking device according to the first image pose to obtain a target pose.

[0005] In a second aspect, the present application further provides an automatic walking device, and the automatic walking device includes: A main body; A driving module for driving the main body to travel; A working module disposed on the main body and used for performing lawn mowing operations on the position where the automatic walking device is located; An acquisition module at least used for collecting images of the environment where the automatic walking device is located and collecting the pose of the automatic walking device; A controller, connected to the driving module and the acquisition module, is configured to execute the pose correction method of the automatic walking device according to any one of the embodiments of the present application.

[0006] In a third aspect, the present application further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the pose correction method of the automatic walking device as described above is implemented.

[0007] The present application provides a pose correction method for an automatic walking device, an automatic walking device, and a computer storage medium. The present application starts the working mode of the automatic walking device in response to a control instruction of a user; acquires a boundary image corresponding to a boundary position of the automatic walking device in a target working area; constructs boundary submaps corresponding to at least two different boundary positions according to the boundary image; validates image data acquired by the automatic walking device according to the boundary submaps to obtain a first image pose; and corrects an error of a historical pose of the automatic walking device according to the first image pose to obtain a target pose. Since the historical pose is corrected according to the first image pose verified by the boundary submap, the accuracy of pose correction of the automatic walking device is improved, thereby improving the accuracy of movement of the automatic walking device. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] To more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0009] Figure 1 It is a schematic flowchart of a pose correction method for an automatic walking device provided by an embodiment of the present application; Figure 2 It is a schematic diagram of a boundary image in a boundary submap provided by an embodiment of the present application; Figure 3 It is a schematic diagram of a first image pose provided by an embodiment of the present application; Figure 4 It is a schematic diagram of error correction of a historical pose according to a first image pose provided by an embodiment of the present application; Figure 5 It is a schematic diagram of a movement trajectory of a first driving operation provided by an embodiment of the present application; Figure 6 It is a schematic diagram of a movement trajectory of an automatic walking device provided in the related art of the present application; Figure 7Schematic diagram of the structure of an automatic walking device provided by an embodiment of the present application. Detailed implementation manners

[0010] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0011] The flowcharts shown in the accompanying drawings are only illustrative examples, and do not necessarily include all the contents and operations / steps, nor do they necessarily need to be executed in the described order. For example, some operations / steps can also be decomposed, combined, or partially merged. Therefore, the actual execution order may change according to the actual situation.

[0012] The embodiments of the present application provide a pose correction method for an automatic walking device, an automatic walking device, and a computer storage medium.

[0013] Next, some implementation manners of the present application will be described in detail in conjunction with the accompanying drawings. Without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0014] Please refer to Figure 1 , Figure 1 which is a schematic flowchart of a pose correction method for an automatic walking device provided by an embodiment of the present application. The pose correction method for the automatic walking device can be used in a terminal or a server to correct the pose of the automatic walking device. Among them, the terminal can be an automatic walking device, such as an electronic device like a lawn mowing robot or a floor sweeping robot; the server can be an independent server, a server cluster, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, Content Delivery Network (CDN), and big data and artificial intelligence platforms.

[0015] As Figure 1 shown, the pose correction method for the automatic walking device includes steps S101 to S105.

[0016] Step S101, in response to a control instruction from the user, start the working mode of the automatic walking device.

[0017] Exemplarily, the control instruction of the user can be remotely sent to the automatic walking device through wireless communication technology, or can be triggered by a button provided on the automatic walking device. Specifically, the user can operate an application for controlling the automatic walking device through a smart phone, so that the server sends the control instruction to the automatic walking device. Of course, it is not limited to this, and no limitation is made here.

[0018] Step S102: Obtain a boundary image corresponding to the boundary position of the automatic walking device in the target working area.

[0019] Exemplarily, different from the related art in which the automatic walking device acquires images without difference at all locations in the entire target working area and uses all the acquired images to construct a map, in the map construction method provided by the technical solution of the present application, only the boundary images acquired by the automatic walking device at the boundary positions in the target working area are used to construct the boundary sub-map. Specifically, the boundary position is the position where the automatic walking device needs to turn around, and the pose of the automatic walking device is likely to deviate at the boundary position. Since the boundary position is the position where the automatic walking device is likely to deviate during the movement, the boundary sub-map constructed from the boundary images can effectively correct the deviation of the automatic walking device while having a relatively small data volume. Among them, the boundary position is not only the boundary of the working area, but also the image area range when the image acquisition device can acquire the boundary features, all of which can be the boundary positions. It can also be the area segment of the position point where the automatic walking device needs to turn around or the time period corresponding to the position point when turning around. The boundary image is not only the image when the automatic walking device is exactly at the edge, but also several consecutive frames of images collected when approaching the boundary (such as when the distance from the boundary is less than a certain threshold).

[0020] Step S103: Construct boundary sub-maps corresponding to at least two different boundary positions according to the boundary images.

[0021] Exemplarily, boundary sub - maps are set for each boundary position respectively to reduce the computational complexity of subsequent rectification of the automatic walking device. Taking the solution of the bundle adjustment (BA) problem based on the Levenberg–Marquardt (LM) algorithm as an example, the computational complexity of the LM algorithm is O(N^3), where N represents the number of images in the map. Suppose there are 100 images in the map of the target working area, then the computational complexity is 100^3 = 1000000; if the 100 images are divided into two sub - maps, then the computational complexity is (100 / 2)^3*2 = 250000, and the calculation speed is increased by 4 times; if the 100 images are divided into four sub - maps, then the computational complexity is (100 / 4)^3*4 = 62500, and the calculation speed is increased by 16 times. Thus, compared with maintaining a single map, dividing a map into multiple boundary sub - maps can significantly reduce the computational complexity and improve the computational efficiency.

[0022] In some embodiments, constructing boundary sub - maps corresponding to at least two different boundary positions according to the boundary images includes: Obtaining the acquisition pose of the automatic walking device when each of the boundary images is acquired; specifically, it can obtain the device pose when each boundary image is acquired through an odometer, an IMU, a GPS / RTK module or a visual SLAM system, and the device pose includes at least the position information and the orientation information of the automatic walking device.

[0023] Classifying the boundary images according to the acquisition pose to obtain boundary sub - maps corresponding to at least two different boundary positions.

[0024] Exemplarily, boundary images acquired near different boundary positions respectively correspond to different boundary sub - maps, so as to divide the boundary images into multiple boundary sub - maps. Clustering all boundary images according to their acquisition poses and dividing them into multiple "boundary image sets", and a unique ID can be assigned to each boundary sub - map, such as Map_A, Map_B, etc. It can be understood that the automatic walking device has different poses when approaching different boundary positions. For example, the positions where the automatic walking device is located are different when approaching different boundary positions, and the directions the automatic walking device faces are also different when moving towards different boundary positions. Therefore, the boundary sub - map to which a boundary image belongs can be determined according to the pose of the automatic walking device when the boundary image is acquired, and the boundary image can be updated to the corresponding boundary sub - map.

[0025] As Figure 2 shown, Figure 2 is a schematic diagram of boundary images in a boundary sub - map provided by an embodiment of the present application.

[0026] AsFigure 2 As shown, assuming that the target working area is a 100-meter * 100-meter working area, the automatic walking device collects images during the traveling process. Figure 2 In it, the vertex of the triangle represents the position where the image is collected, and the vertex orientation represents the angle of the collected image. If Map_A in the dashed box is the north boundary map, when the automatic walking device collects the boundary image in Map_A, the position is close to the north boundary of the working area, that is, the Y coordinate is close to the maximum value. For example, the boundary image in Map_A is collected near the position of Y≈99 meters, and the orientation angle is θ90°; it can be understood that since there may be a certain angular deviation in the orientation of the automatic walking device during the walking process, as long as the orientation angle is within the preset angular deviation range, the collected images can all be used as the boundary images in the boundary sub-map corresponding to this boundary. Assuming that the collection pose of the boundary image is represented by (X, Y, θ), the collection pose of the boundary image in Map_A can be, for example: (5,99, 90°), (25,99, 90°), (50,99,92°), (75,99,88°), (95,99,90°), etc. Of course, it is not limited to this. There can also be boundary images corresponding to the positions of Y≈98 meters and Y≈97 meters in Map_A, which are not limited here. In some embodiments, the collection pose includes the collection position and the collection direction. Classifying the boundary images according to the collection pose to obtain boundary sub-maps corresponding to at least two different boundary positions includes: Obtaining the characteristic direction and characteristic position of each of the boundary sub-maps; When the collection direction of the boundary image matches the characteristic direction and the collection position matches the characteristic position, it is determined that the boundary image corresponds to the boundary sub-map.

[0027] Exemplarily, matching the collection direction of the boundary image with the characteristic direction of the boundary sub-map, that is, comparing the collection direction with the characteristic direction. When the included angle formed by the two is less than the preset included angle threshold, it is determined that the collection direction matches the characteristic direction.

[0028] Exemplarily, matching the collection position of the boundary image with the characteristic position of the boundary sub-map, that is, comparing the collection position with the characteristic position. When the distance between the two is less than the preset distance threshold, it is determined that the collection position matches the characteristic position.

[0029] Among them, the characteristic direction of the boundary sub-map can be the collection direction of the existing boundary images in the boundary sub-map, and the characteristic position of the boundary sub-map can be the collection position of the existing boundary images in the boundary sub-map.

[0030] Step S104: Verify the image data obtained by the automatic walking device according to the boundary sub-map to obtain the first image pose.

[0031] Exemplarily, the boundary image in the boundary sub-map is an image collected by the automatic walking device at a position near the boundary during the process of constructing the boundary sub-map. The automatic walking device moves based on the indication of Visual-Inertial Odometry (VIO for short) during the operation process, and collects image data during the movement. It can determine the boundary position where the current image is located according to the VIO output result, and in the corresponding boundary sub-map, select a set of boundary images with the closest spatial distance and direction distance to the current pose as the comparison image set.

[0032] It can be understood that, in an ideal situation, when the automatic walking device moves to a position near the boundary during the operation process, the collected image data should be consistent with the boundary image in the boundary sub-map. However, due to the error generated by VIO, there will be a certain degree of deviation between the pose when the automatic walking device actually captures the image data and the pose when the boundary image is captured. Therefore, there will also be a certain difference between the image data obtained by the automatic walking device and the boundary image in the boundary sub-map. At this time, the image data can be verified by comparing the gap between the boundary image and the image data, and the actual first image pose when the automatic walking device collects the image data during the movement can be obtained.

[0033] In some embodiments, the step of verifying the image data obtained by the automatic walking device according to the boundary sub-map to obtain the first image pose includes: The image data includes sub-images collected by the automatic walking device within a preset time window; Extract the feature points of the boundary sub-map, perform position verification with the feature points in the sub-images, and obtain the first image pose.

[0034] Please refer to Figure 3 , Figure 3 which is a schematic diagram of the first image pose provided by an embodiment of the present application.

[0035] As Figure 3 shown, Figure 3 the circles in it represent 3D points, and the triangles represent the poses when taking images. Among them, the position of the triangle vertex represents the position where the image is taken, and the orientation of the apex angle represents the direction of image taking. Figure 3The triangle formed by the solid lines is the boundary image pose, i.e., the pose when collecting the boundary image in the boundary sub-map; the triangle formed by the dashed lines is the first image pose, i.e., the actual pose when collecting the image data. In an ideal situation, the first image pose should be the same as the boundary image pose. However, there will actually be a certain deviation between the first image pose and the boundary image pose, that is Figure 3 the gap between the first image pose and the boundary image pose shown.

[0036] Exemplarily, the actual pose of the image data, i.e., the first image pose, can be determined by comparing the feature points in the sub-image obtained by the autonomous mobile device and the boundary image.

[0037] Specifically, the first image pose can be determined by optimizing the following objective function:

[0038] where MapPose represents the first image pose, and VioPose represents the boundary image pose. represents the function of projecting 3D points into pixels using the camera model given the internal and external parameters of the camera of the autonomous mobile device and the given 3D points. represents the covariance matrix corresponding to the residual term. The form of represents the Mahalanobis distance. The first term is the position constraint, and the true scale is ensured during the optimization process through the position-prior; the second term is in the form of the common bundle adjustment (BA).

[0039] Figure 3 The bold line connection between the first image pose and the boundary image pose in represents the position constraint between the two. The connection between the first image pose and multiple 3D points represents the constraint of the reprojection error, i.e., the bundle adjustment. By comparing the feature points of the actually collected image data with the image in the boundary sub-map (image pose verification) and optimizing the objective function to obtain a more realistic image pose (the first image pose), the rough historical pose information obtained based on VIO is effectively corrected.

[0040] Step S105: According to the first image pose, correct the error of the historical pose of the autonomous mobile device to obtain the target pose.

[0041] Exemplarily, after verifying the image data to obtain the first image data, since the first image pose can more accurately represent the actual pose of the autonomous mobile device during movement than the pose of the image data before verification, the historical pose recorded based on VIO is corrected by the first image pose to obtain a more accurate target pose, so as to use the more accurate target pose as the basis for the next movement of the autonomous mobile device.

[0042] In some embodiments, correcting the error of the historical pose of the automatic walking device according to the first image pose to obtain a target pose includes: Obtaining a first time point matching the first image pose, and obtaining first pose information at the first time point; Based on the first image pose, correcting the error of the first pose information at the first time point to obtain a first target pose.

[0043] It can be understood that the first pose information is the historical pose recorded by VIO at the first time point, and the first image pose is the pose obtained by verifying the image data at the first time point. Therefore, the accuracy of the first image pose is higher than that of the first pose information. Correcting the error of the first pose information based on the first image pose to obtain a more accurate first target pose.

[0044] Please refer to Figure 4 , Figure 4 , which is a schematic diagram of correcting the error of the historical pose according to the first image pose provided by an embodiment of the present application.

[0045] Figure 4 In, the solid circle represents the historical pose, and the dashed circle represents the first image pose obtained in step S103.

[0046] In some embodiments, the historical pose is based on a set of pose information within the preset time window. Obtaining second pose information at a second time point adjacent to the first time point, correcting the second pose information based on the first target pose to obtain a second target pose, and gradually correcting the historical pose through this process to obtain the target pose.

[0047] As Figure 4 shown, the historical pose includes a series of pose information arranged in chronological order. For example, the first pose information is adjacent to the second pose information. It can be understood that due to the continuity of motion, each historical pose except the first historical pose is related to the historical pose at the previous time point. Therefore, each pose information in the historical pose is corrected one by one in chronological order to obtain the target pose.

[0048] In some embodiments, the gradually correcting the historical pose to obtain the target pose includes: Establishing a first constraint condition between the first pose information matching the first time point and the first image pose; Establishing a second constraint condition between the first pose information adjacent to the first time point and the second pose information; Correcting the error of the historical pose according to the first constraint condition and the second constraint condition to obtain the target pose.

[0049] As shown Figure 4 in the figure, the constraints on the first pose information include the constraints on the second pose information adjacent to the first pose information, that is, the first constraint condition; and the constraints on the first image information obtained by verification in step S103, that is, the second constraint condition. Among them, Figure 4 the connection lines between the respective historical poses represent the first constraint condition, and the connection lines between each historical pose and the first image pose represent the second constraint condition.

[0050] Exemplarily, the error correction of the historical pose can be achieved by optimizing the following objective function:

[0051] Among them, the meaning of is the inverse of, and its meaning is expressed as the inverse of multiplied by , and the subsequent formulas have the same similar meaning; represents the first pose information, represents the second pose information adjacent to the first pose information, represents the first image pose. The first term in the objective function represents the constraint of the second pose information on the first pose information, that is, the second constraint condition; the second term represents the constraint of the first image pose on the first pose information, that is, the first constraint condition.

[0052] In some embodiments, during the process of the first driving operation of the automatic walking device in the working mode, it enters the same boundary area at least twice, and the positions where it enters the same boundary area at least twice do not coincide in the extending direction of the boundary area.

[0053] Please refer to Figure 5 , Figure 5 which is a schematic diagram of the motion trajectory of the first driving operation provided by an embodiment of the present application.

[0054] As Figure 5 shown, the pose deviation correction provided by the embodiment of the present application is based on the boundary sub-map corresponding to the boundary position, and the boundary area corresponding to the boundary sub-map should be the area that the automatic walking device needs to repeatedly enter during the first driving operation. For example Figure 5 the position corresponding to the dashed box in. Since the automatic walking device enters the same boundary area at least twice, one boundary sub-map can be used for pose deviation correction at least twice during the movement of the automatic walking device, improving the utilization rate of the boundary sub-map.

[0055] Moreover, it can be understood that the boundary position is the position where the automatic walking device needs to turn around during movement, and it is the position where the pose is prone to errors. Correcting the deviation in the boundary area near the boundary position can effectively reduce the pose error.

[0056] Among them, the paths of the automatic walking device entering the same boundary area at least twice are different, so the positions where the automatic walking device enters the same boundary area at least twice do not coincide in the extension direction of the boundary area.

[0057] It can be understood that the movement trajectory of the automatic walking device during the first driving operation in the working mode is not limited to Figure 5 the movement path shown in Figure 5 The movement trajectory shown in is only for illustration. The actual operation target working area can be of other shapes, and the movement trajectory of the automatic walking device is adapted to the shape of the target working area. For example, when the shape of the target working area is circular or elliptical, the path in the movement trajectory of the automatic walking device can be arc-shaped.

[0058] In some embodiments, the driving path of the first driving operation is bow-shaped.

[0059] As Figure 5 shown, the automatic walking device reciprocates along a bow-shaped path in the target working area, so that the movement trajectory can cover the target working area as much as possible. Specifically, the automatic walking device moves north to one of the boundary positions of the target working area and then turns around to change direction, and then moves south to the other boundary position of the target working area, and repeats this process until the movement trajectory covers the entire target working area.

[0060] Please refer to Figure 6 , Figure 6 which is a schematic diagram of the movement trajectory of the automatic walking device provided in the related art of the present application.

[0061] The actual movement trajectory of the automatic walking device may have an offset as Figure 4 shown. For example, due to the cumulative error generated during the movement of the automatic walking device, it is usually difficult to ensure that the movement trajectories are completely parallel, which is likely to cause different angles of different paths, resulting in some positions in the target working area not being covered. Taking a lawn mowing robot as an example, Figure 6 the black area in is the missed area of the lawn mowing operation.

[0062] To avoid the above situation, in the embodiments of the present application, a boundary sub-map is constructed for the boundary position before the automatic walking device turns around in the target working area, and the movement trajectory of the automatic walking device is corrected according to the boundary sub-map before each turn, so as to avoid as Figure 6The automatic walking device continuously generates and accumulates errors during the movement process, resulting in an increasing error. Through the pose correction method provided by the embodiments of the present application, the actual movement trajectory of the automatic walking device is made to correspond as much as possible to the Figure 5 in the movement trajectory, and the situation of Figure 6 in the movement trajectory is avoided, improving the aesthetics of the movement trajectory and reducing the omitted areas in the target working area. Taking the automatic walking device as a lawn mowing robot as an example, the omitted areas of the lawn mowing operation performed by the lawn mowing robot are reduced.

[0063] Please refer to Figure 7 , Figure 7 which is a schematic structural diagram of the automatic walking device provided by an embodiment of the present application.

[0064] As Figure 7 shown, the embodiments of the present application further provide an automatic walking device, and the automatic walking device includes: a main body 10; a driving module 20 for driving the main body to travel; a working module 30 provided on the main body and used for performing a lawn mowing operation on the position where the automatic walking device is located; a collection module (not shown in the figure), at least used for collecting image data of the environment where the automatic walking device is located and collecting the pose of the automatic walking device; a controller (not shown in the figure), connected to the driving module and the collection module, and used for executing the pose correction method of the automatic walking device according to any one of the embodiments of the present application.

[0065] Exemplarily, the automatic walking device provided by the embodiments of the present application can be used to perform a lawn mowing operation during movement. Of course, it is not limited thereto. The automatic walking device provided by the embodiments of the present application can also perform operations such as cleaning, snow sweeping, and leaf blowing, which are not limited herein.

[0066] The collection module of the automatic walking device is used to collect image data of the environment where the robot is located and collect the pose. For example, it may include: a visible light camera or an infrared light camera for collecting image data, an inertial measurement module for obtaining the collection direction, and a real-time kinematic positioning module for obtaining the collection position. Of course, it is not limited thereto. The collection module can also be replaced by other devices or include more components, which are not limited herein.

[0067] Exemplarily, the controller is used to implement the map construction method provided by the embodiments of the present application to obtain a boundary sub-map, and correct the movement of the automatic walking device through the boundary sub-map.

[0068] An embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored. The computer program includes program instructions, and the method implemented when the program instructions are executed can refer to the various embodiments of the pose correction method of the automatic walking device of the present application.

[0069] Among them, the computer-readable storage medium may be an internal storage unit of the computer device described in the foregoing embodiments, such as the hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc. equipped on the computer device.

[0070] It should be understood that the terms used in the specification of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in the specification of the present application and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include the plural forms.

[0071] It should also be understood that the term "and / or" used in the specification of the present application and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations. It should be noted that in this article, the term "comprises", "comprising" or any other variation thereof is intended to cover a non-exclusive inclusion, so that a process, method, article or system including a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or system. Without further limitation, an element defined by the statement "comprising a..." does not exclude the existence of additional identical elements in the process, method, article or system including the element.

[0072] The serial numbers of the embodiments of the present application above are only for description and do not represent the advantages and disadvantages of the embodiments. The above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of various equivalent modifications or substitutions, and these modifications or substitutions should be covered by the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for correcting the pose of an automatic walking device, characterized in that, The method includes: In response to a control instruction of a user, starting a working mode of the automatic walking device; Obtaining a boundary image corresponding to a boundary position of the automatic walking device in a target working area; Constructing boundary sub-maps corresponding to at least two different boundary positions according to the boundary image; According to the boundary sub-maps, verifying image data obtained by the automatic walking device to obtain a first image pose; According to the first image pose, correcting an error of a historical pose of the automatic walking device to obtain a target pose.

2. The method according to claim 1, characterized in that The step of verifying the image data obtained by the automatic walking device according to the boundary sub-maps to obtain a first image pose includes: the image data includes sub-images collected by the automatic walking device within a preset time window; Extracting feature points of the boundary sub-maps and performing position verification with feature points in the sub-images to obtain a first image pose.

3. The method according to claim 1, characterized in that, The step of correcting an error of a historical pose of the automatic walking device according to the first image pose to obtain a target pose includes: Obtaining a first time point matching the first image pose and obtaining first pose information at the first time point; Based on the first image pose, correcting an error of the first pose information at the first time point to obtain a first target pose.

4. The method according to claim 3, characterized in that, The historical pose is based on a set of pose information within a preset time window. Obtaining second pose information of a second time point adjacent to the first time point, correcting the second pose information based on the first target pose to obtain a second target pose, and gradually correcting the historical pose through this process to obtain the target pose.

5. The method according to claim 4, characterized in that The step of gradually correcting the historical pose to obtain the target pose includes: Establishing a first constraint condition between the first pose information matching the first time point and the first image pose; Establishing a second constraint condition between the first pose information adjacent to the first time point and the second pose information; According to the first constraint condition and the second constraint condition, correcting an error of the historical pose to obtain a target pose.

6. The method according to claim 1, wherein During at least two times of entering the same boundary area in the process of a first driving operation of the automatic walking device in the working mode, and positions of at least two times of entering the same boundary area do not coincide in the extending direction of the boundary area.

7. The method according to claim 6, characterized in that, The driving path of the first driving operation is in a bow shape.

8. The method according to claim 1, wherein The step of constructing boundary sub-maps corresponding to at least two different boundary positions according to the boundary image includes: Obtaining a collection pose of the automatic walking device when each boundary image is collected; Classifying the boundary images according to the collection pose to obtain boundary sub-maps corresponding to at least two different boundary positions.

9. The method according to claim 8, characterized in that, The collection pose includes a collection position and a collection direction. The step of classifying the boundary images according to the collection pose to obtain boundary sub-maps corresponding to at least two different boundary positions includes: Obtaining a feature direction and a feature position of each boundary sub-map; When the acquisition direction of the boundary image matches the feature direction and the acquisition position matches the feature position, it is determined that the boundary image corresponds to the boundary sub-map.

10. An automatic walking device, characterized in that, The device includes: A body; A driving module for driving the body to travel; A working module provided on the body and used for mowing the position where the automatic walking device is located; An acquisition module at least used for acquiring images of the environment where the automatic walking device is located and acquiring the pose of the automatic walking device; A controller connected to the driving module and the acquisition module, and used for executing the pose correction method of the automatic walking device according to any one of claims 1-9.

11. A computer storage medium, characterized in that, The computer storage medium stores a computer program, and when the computer program is executed by a processor, it implements the pose correction method of the automatic walking device according to any one of claims 1 to 9.

Images