Localization and mapping method and system for track robot in power distribution room

By employing image acquisition by a tracked robot and the Manhattan world hypothesis, the problems of large positioning errors and low real-time information in the power distribution room environment were solved, achieving high-precision and real-time positioning and mapping of the tracked robot.

CN118470288BActive Publication Date: 2026-11-17GUANGDONG KEYSTAR INTELLIGENCE ROBOT CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202410493177.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-04-23
Publication Date
2026-11-17
Estimated Expiration
2044-04-23

AI Technical Summary

Technical Problem

Existing track robots suffer from large positioning errors and low real-time information in power distribution room environments, making them unusable immediately upon activation. Furthermore, their positioning accuracy is affected by pulley slippage when relying on encoders, and they cannot provide extensive environmental data.

Method used

By using an orbital robot to collect images of a power distribution room for OCR text detection and feature point matching, and combining the Manhattan world hypothesis and a preset pose tracking method, pose constraints for the orbital robot are constructed to achieve 3D map construction.

Benefits of technology

It improves the accuracy and real-time performance of positioning, maintains positioning precision even when the wheels slip, makes full use of power distribution room information for real-time mapping, and supports complex navigation and task execution.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118470288B_ABST
    Figure CN118470288B_ABST
Patent Text Reader

Abstract

This invention relates to the field of automation technology, and more particularly to a method and system for localization and mapping of a tracked robot in a power distribution room. The invention proposes a localization and mapping method for a tracked robot in a power distribution room based on pure visual map modeling and Manhattan world plane matching. This method can operate directly using real-time images captured by the robot's camera and estimates the robot's pose based on real-time image information, maintaining positioning accuracy even when the robot's wheels slip. Simultaneously, by using text labels and feature points of the power distribution room, and the Manhattan world assumption, the matching constraints between the plane and the localization are improved, enhancing the real-time performance and accuracy of localization.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of automation technology, and in particular to a method and system for positioning and mapping a track robot in a power distribution room. Background Technology

[0002] A track-mounted robot is an automated robotic device commonly used for inspection tasks in environments with a large number of devices, such as power distribution rooms. Track-mounted robots move along tracks using pulleys and other sliding mechanisms to perform inspection or monitoring tasks.

[0003] Generally, tracked robots use a unified control system to aggregate data, allowing operators to locate the constantly moving robot's position. In related technologies, most tracked robot positioning methods are based on encoders and manually deployed tags (NFC, QR codes, or RFID modules, etc.). The basic idea is to use encoders for continuous positioning and then use the truth information contained in the manually deployed tags to eliminate the cumulative errors of the encoder.

[0004] However, the above positioning method has the following problems:

[0005] Relying on manual placement of tags to correct positioning errors makes it impossible to fully automate the entire positioning process, lacks the flexibility to be used immediately, and reduces the level of intelligence.

[0006] When relying on encoders for positioning, if the pulleys slip or spin freely, the positioning accuracy will be severely affected and cannot be guaranteed.

[0007] Limited by the robot's pose information, this method cannot provide broader environmental data, such as environmental features, which limits its widespread use in business applications.

[0008] On the other hand, the position of a tracked robot needs to be described based on its operating environment. A precise and easily descriptive 3D spatial map is beneficial for accurately representing the position of the tracked robot. Related technologies utilize SLAM (Simultaneous Localization and Mapping) to reconstruct a 3D spatial map of the power distribution room. After 3D reconstruction, convolutional neural networks are used to identify the quantity and type of equipment.

[0009] However, the above-mentioned three-dimensional reconstruction method has the following problems:

[0010] Relying on existing 3D reconstruction data of the power distribution room, and then using a convolutional neural network to identify the number and type of equipment, the real-time performance is low and it cannot provide the robot's status and the real-time situation of the power distribution room in real time.

[0011] The information in the power distribution room was not fully utilized for positioning and mapping, resulting in a waste of information. Summary of the Invention

[0012] The present invention aims to solve the problems of large position information errors and low real-time information performance that are easily generated when using track robots in power distribution room environments in the existing technology.

[0013] To solve the above-mentioned technical problems, in a first aspect, the present invention provides a method for positioning and mapping a power distribution room track robot, comprising the following steps:

[0014] S1. Collect images of the power distribution room using a track robot, and perform OCR text detection based on the images of the power distribution room to obtain planar text data of the power distribution room;

[0015] S2. Perform feature detection on the power distribution room image to obtain power distribution room feature point data;

[0016] S3. Match the power distribution room text plane data and the power distribution room feature point data corresponding to the power distribution room images in different frames;

[0017] S4. Recognize the power distribution room image according to the Manhattan world hypothesis and extract Manhattan plane data in the power distribution room environment;

[0018] S5. Match the power distribution room text plane data in the current frame with the Manhattan plane data;

[0019] S6. Perform pose tracking on the track robot based on a preset pose tracking method to obtain pose data, and determine whether the power distribution room image collected by the track robot is a key frame based on the pose data. If yes, proceed to step S7; otherwise, return to step S1.

[0020] S7. Based on the pose data, perform coordinate transformation on the power distribution room text plane data and the power distribution room feature point data to obtain the world map text plane and world map feature points respectively;

[0021] S8. Based on the distance error between the feature point data of the power distribution room and the text plane data of the power distribution room, and the angle error between the text plane data of the power distribution room and the Manhattan plane data, construct the pose constraint of the track robot, and construct a three-dimensional map of the power distribution room according to the pose constraint of the track robot.

[0022] Furthermore, the power distribution room image is an RGB-D depth image or a binocular stereo image.

[0023] Furthermore, in step S3, the method for matching the power distribution room text plane data and the power distribution room feature point data corresponding to the power distribution room images of different frames is specifically as follows:

[0024] A preset first threshold is set. Based on the centroid and covariance distribution of the feature points in the power distribution room feature point data, the Mahalanobis distance between the power distribution room text plane data and the power distribution room feature point data is calculated. If the calculated Mahalanobis distance is less than the preset first threshold, the power distribution room text plane data and the power distribution room feature point data are matched and bound.

[0025] Furthermore, the centroid and covariance distributions of the feature points in the power distribution room feature point data are defined as C and S, respectively, where N is the current number of feature points, and x is the number of feature points. i y i z i Let x, y, and z be the x, y, and z coordinates of feature point i in the three-dimensional coordinate system, respectively. The Mahalanobis distance between the power distribution room text plane data and the power distribution room feature point data is D, which satisfies the following relationship:

[0026] C = (C x C y C z (1);

[0027]

[0028]

[0029] Furthermore, in step S5, the method for matching the power distribution room text plane data with the Manhattan plane data in the current frame is specifically as follows:

[0030] A second preset threshold is set, and the respective planes of the power distribution room text plane data and the Manhattan plane data are compared. If the angle between the planes is less than the second preset threshold, the power distribution room text plane data and the Manhattan plane data are matched and bound.

[0031] Furthermore, in step S8, the orbital robot pose constraint constructed based on the distance error from the power distribution room feature point data to the power distribution room text plane data, and the angle error between the power distribution room text plane data and the Manhattan plane data, is a nonlinear optimization residual that satisfies the following relationship:

[0032] E textplane =nT wc P c -d (4);

[0033]

[0034]

[0035] Among them, E reproj E represents the reprojection error. textplaneE represents the distance error from the point to the text plane. plane The error is the angle between the text plane of the power distribution room and the Manhattan plane.

[0036] In relation (4), n is the normal vector of the power distribution room text plane, and T wc For the pose transformation of the camera to the world coordinate system, P c Let d be a feature point of the power distribution room, and d be the text plane parameter of the power distribution room;

[0037] In relation (5), n mw Let R be the normal vector of the Manhattan world plane, R be the rotation matrix from the camera to the world coordinate system, and n be the normal vector. text This is the normal vector of the text plane of the power distribution room.

[0038] Furthermore, the preset pose tracking method is either the PnP method or the constant velocity model method.

[0039] Secondly, the present invention also provides a positioning and mapping system for a power distribution room track robot, comprising:

[0040] The data acquisition module is used to acquire images of the power distribution room through the track robot, and to perform OCR text detection based on the images of the power distribution room to obtain planar text data of the power distribution room.

[0041] The feature detection module is used to perform feature detection on the power distribution room image to obtain power distribution room feature point data;

[0042] The feature matching module is used to match the power distribution room text plane data and the power distribution room feature point data corresponding to the power distribution room images in different frames;

[0043] The Manhattan detection module is used to identify the substation image according to the Manhattan world hypothesis and extract Manhattan plane data in the substation environment;

[0044] The Manhattan matching module is used to match the power distribution room text plane data in the current frame with the Manhattan plane data;

[0045] The pose tracking module is used to perform pose tracking on the track robot based on a preset pose tracking method to obtain pose data, and to determine whether the power distribution room image collected by the track robot is a key frame based on the pose data. If so, the steps are executed.

[0046] The data conversion module is used to perform coordinate transformation on the power distribution room text plane data and the power distribution room feature point data according to the pose data, so as to obtain the world map text plane and the world map feature points respectively;

[0047] The mapping module is used to construct the pose constraints of the track robot based on the distance error between the feature point data of the power distribution room and the text plane data of the power distribution room, as well as the angle error between the text plane data of the power distribution room and the Manhattan plane data, and to construct a three-dimensional map of the power distribution room based on the pose constraints of the track robot.

[0048] The beneficial effects achieved by this invention are that it proposes a localization and mapping method for a power distribution room track robot based on pure visual map modeling and Manhattan world plane matching. This method can work directly using real-time images captured by the track robot's camera and estimate the robot's pose based on real-time image information, maintaining positioning accuracy even when the robot's wheels slip. At the same time, it improves the matching constraints of the plane and positioning based on the text labels and feature points of the power distribution room and the Manhattan world hypothesis, thereby improving the real-time performance and accuracy of positioning. Attached Figure Description

[0049] Figure 1 This is a flowchart illustrating the steps of the positioning and mapping method for a power distribution room track robot provided in an embodiment of the present invention;

[0050] Figure 2 This is a schematic diagram of OCR text detection in a power distribution room provided in an embodiment of the present invention;

[0051] Figure 3 This is a diagram illustrating the effect of inter-frame feature point and planar matching provided in an embodiment of the present invention.

[0052] Figure 4 This is a schematic diagram of nonlinear optimization residuals provided in an embodiment of the present invention;

[0053] Figure 5 This is a schematic diagram of a three-dimensional map provided in an embodiment of the present invention;

[0054] Figure 6 This is another schematic diagram of a three-dimensional map provided in an embodiment of the present invention. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0056] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating the steps of the positioning and mapping method for a power distribution room track robot provided in an embodiment of the present invention. The positioning and mapping method for the power distribution room track robot includes the following steps:

[0057] S1. Collect images of the power distribution room using a tracked robot, and perform OCR text detection based on the images to obtain planar text data of the power distribution room.

[0058] The image of the power distribution room is an RGB-D depth image or a binocular stereo image. The power distribution room is filled with a large amount of signage and text information; this macroscopic and semantic information can serve as an important constraint for localization. OCR (Optical Character Recognition) text detection technology is a technique for accurately locating text regions from an image, as illustrated in the diagram below. Figure 2 As shown.

[0059] S2. Perform feature detection on the power distribution room image to obtain power distribution room feature point data.

[0060] S3. Match the power distribution room text plane data and the power distribution room feature point data corresponding to the power distribution room images in different frames.

[0061] Feature point matching can be described using the ORB operator. In step S3, the method for matching the power distribution room text plane data and the power distribution room feature point data corresponding to the power distribution room images in different frames is as follows:

[0062] A preset first threshold is set. Based on the centroid and covariance distribution of the feature points in the power distribution room feature point data, the Mahalanobis distance between the power distribution room text plane data and the power distribution room feature point data is calculated. If the calculated Mahalanobis distance is less than the preset first threshold, the power distribution room text plane data and the power distribution room feature point data are matched and bound.

[0063] In this embodiment of the invention, the preset first threshold is set to 10 degrees.

[0064] Define the centroid and covariance distributions of the feature points in the power distribution room feature point data as C and S, respectively, where N is the current number of feature points, and x is the number of feature points. i y i z i Let x, y, and z be the x, y, and z coordinates of feature point i in the three-dimensional coordinate system, respectively. The Mahalanobis distance between the power distribution room text plane data and the power distribution room feature point data is D, which satisfies the following relationship:

[0065] C = (C x C y C z (1);

[0066]

[0067]

[0068] like Figure 3As shown, Figure 3 This is a diagram showing the effect of inter-frame feature point and planar matching.

[0069] S4. Recognize the power distribution room image according to the Manhattan world hypothesis and extract Manhattan plane data in the power distribution room environment.

[0070] The Manhattan world hypothesis posits that the main structures in an environment are arranged along orthogonal directions. Planes that conform to the Manhattan world hypothesis are either perpendicular or parallel to each other. In this embodiment of the invention, identifying Manhattan plane data in the environment helps to improve the constraint strength of positioning.

[0071] S5. Match the power distribution room text plane data in the current frame with the Manhattan plane data.

[0072] In step S5, the method for matching the power distribution room text plane data with the Manhattan plane data in the current frame is as follows:

[0073] A second preset threshold is set, and the respective planes of the power distribution room text plane data and the Manhattan plane data are compared. If the angle between the planes is less than the second preset threshold, the power distribution room text plane data and the Manhattan plane data are matched and bound.

[0074] S6. Perform pose tracking on the track robot based on a preset pose tracking method to obtain pose data, and determine whether the power distribution room image collected by the track robot is a keyframe based on the pose data. If yes, proceed to step S7; otherwise, return to step S1.

[0075] In traditional encoder-based positioning systems, wheel slippage can lead to inaccurate positioning. In this embodiment of the invention, image information captured by a vision sensor can accurately estimate the robot's pose independently of the wheel's motion, maintaining positioning accuracy even when the wheels slip.

[0076] The preset pose tracking method is either the PnP method or the constant velocity model method.

[0077] S7. Based on the pose data, perform coordinate transformation on the power distribution room text plane data and the power distribution room feature point data to obtain the world map text plane and world map feature points, respectively.

[0078] S8. Based on the distance error between the feature point data of the power distribution room and the text plane data of the power distribution room, and the angle error between the text plane data of the power distribution room and the Manhattan plane data, construct the pose constraint of the track robot, and construct a three-dimensional map of the power distribution room according to the pose constraint of the track robot.

[0079] In step S8, the orbital robot pose constraint constructed based on the distance error from the power distribution room feature point data to the power distribution room text plane data, and the angle error between the power distribution room text plane data and the Manhattan plane data, is a nonlinear optimization residual that satisfies the following relationship:

[0080] E textplane =nT wc P c -d (4);

[0081]

[0082]

[0083] Among them, E reproj E represents the reprojection error. textplane E represents the distance error from the point to the text plane. plane The error is the angle between the text plane of the power distribution room and the Manhattan plane.

[0084] In relation (4), n is the normal vector of the power distribution room text plane, and T wc For the pose transformation of the camera to the world coordinate system, P c Let d be a feature point of the power distribution room, and d be the text plane parameter of the power distribution room;

[0085] In relation (5), n mw Let R be the normal vector of the Manhattan world plane, R be the rotation matrix from the camera to the world coordinate system, and n be the normal vector. text This is the normal vector of the text plane of the power distribution room.

[0086] The execution process of the above nonlinear optimization residual is as follows: Figure 4 As shown, Figure 4 In this context, π represents a plane vertex, P represents a map vertex, and T represents a pose vertex.

[0087] In this invention, the final output 3D map can be as follows: Figure 5 The image shown includes a 3D map with sparse feature points and text boxes, and as shown in the image. Figure 6 The map shown is a dense 3D map containing only text.

[0088] In this embodiment of the invention, all processes are performed in real time, eliminating the need for prior 3D mapping and equipment identification. Instead, mapping and identification occur simultaneously with the robot's movement, meaning that the robot's pose and information about the power distribution room equipment can be acquired in real time. Furthermore, utilizing the power distribution room information allows for the simple acquisition of information about meters and text tags, which is more comprehensive than relying solely on encoders or other traditional methods for location information. This supports more complex navigation and task execution, improving data utilization.

[0089] The beneficial effects achieved by this invention are that it proposes a localization and mapping method for a power distribution room track robot based on pure visual map modeling and Manhattan world plane matching. This method can work directly using real-time images captured by the track robot's camera and estimate the robot's pose based on real-time image information, maintaining positioning accuracy even when the robot's wheels slip. At the same time, it improves the matching constraints of the plane and positioning based on the text labels and feature points of the power distribution room and the Manhattan world hypothesis, thereby improving the real-time performance and accuracy of positioning.

[0090] This invention also provides a positioning and mapping system for a power distribution room track robot, comprising:

[0091] The data acquisition module is used to acquire images of the power distribution room through the track robot, and to perform OCR text detection based on the images of the power distribution room to obtain planar text data of the power distribution room.

[0092] The feature detection module is used to perform feature detection on the power distribution room image to obtain power distribution room feature point data;

[0093] The feature matching module is used to match the power distribution room text plane data and the power distribution room feature point data corresponding to the power distribution room images in different frames;

[0094] The Manhattan detection module is used to identify the substation image according to the Manhattan world hypothesis and extract Manhattan plane data in the substation environment;

[0095] The Manhattan matching module is used to match the power distribution room text plane data in the current frame with the Manhattan plane data;

[0096] The pose tracking module is used to perform pose tracking on the track robot based on a preset pose tracking method to obtain pose data, and to determine whether the power distribution room image collected by the track robot is a key frame based on the pose data. If so, the steps are executed.

[0097] The data conversion module is used to perform coordinate transformation on the power distribution room text plane data and the power distribution room feature point data according to the pose data, so as to obtain the world map text plane and the world map feature points respectively;

[0098] The mapping module is used to construct the pose constraints of the track robot based on the distance error between the feature point data of the power distribution room and the text plane data of the power distribution room, as well as the angle error between the text plane data of the power distribution room and the Manhattan plane data, and to construct a three-dimensional map of the power distribution room based on the pose constraints of the track robot.

[0099] The positioning and mapping system for the power distribution room track robot can implement the steps in the positioning and mapping method for the power distribution room track robot in the above embodiments, and can achieve the same technical effect. Refer to the description in the above embodiments, which will not be repeated here.

[0100] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0101] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0102] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0103] The embodiments of the present invention have been described above with reference to the accompanying drawings. The disclosed embodiments are merely preferred embodiments of the present invention. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many equivalent changes in form without departing from the spirit and scope of the claims of the present invention, and all such changes are within the protection scope of the present invention.

Claims

1. A method for positioning and mapping of a substation track robot, characterized in that, Includes the following steps: S1. Collect images of the power distribution room using a track robot, and perform OCR text detection based on the images of the power distribution room to obtain planar text data of the power distribution room; S2. Perform feature detection on the power distribution room image to obtain power distribution room feature point data; S3. Match the power distribution room text plane data and the power distribution room feature point data corresponding to the power distribution room images in different frames; S4. Recognize the power distribution room image according to the Manhattan world hypothesis and extract Manhattan plane data in the power distribution room environment; S5. Match the power distribution room text plane data with the Manhattan plane data in the current frame; S6. Perform pose tracking on the track robot based on a preset pose tracking method to obtain pose data, and determine whether the power distribution room image collected by the track robot is a key frame based on the pose data. If yes, proceed to step S7; otherwise, return to step S1. S7. Based on the pose data, perform coordinate transformation on the power distribution room text plane data and the power distribution room feature point data to obtain the world map text plane and world map feature points respectively; S8. Based on the distance error between the feature point data of the power distribution room and the text plane data of the power distribution room, and the angle error between the text plane data of the power distribution room and the Manhattan plane data, construct the pose constraint of the track robot, and construct a three-dimensional map of the power distribution room according to the pose constraint of the track robot.

2. The method of Claim 1, wherein, The images of the power distribution room are RGB-D depth images or binocular stereo images.

3. The method of Claim 1, wherein, In step S3, the method for matching the power distribution room text plane data and the power distribution room feature point data corresponding to the power distribution room images in different frames is as follows: A preset first threshold is set. Based on the centroid and covariance distribution of the feature points in the power distribution room feature point data, the Mahalanobis distance between the power distribution room text plane data and the power distribution room feature point data is calculated. If the calculated Mahalanobis distance is less than the preset first threshold, the power distribution room text plane data and the power distribution room feature point data are matched and bound.

4. The method of Claim 3, wherein, Define the centroid and covariance distribution of the feature points in the power distribution room feature point data C, S, N is the current number of feature points, x i , y i , z i x-axis, y-axis, z-axis coordinates of feature point i in the three-dimensional coordinate system, the Mahalanobis distance between the power distribution room text plane data and the power distribution room feature point data D, which satisfies the following relationship: C = (C x , C y , C z ) (1); 5. The method of Claim 1, wherein, In step S5, the method for matching the power distribution room text plane data with the Manhattan plane data in the current frame is as follows: A second preset threshold is set, and the respective planes of the power distribution room text plane data and the Manhattan plane data are compared. If the angle between the planes is less than the second preset threshold, the power distribution room text plane data and the Manhattan plane data are matched and bound.

6. The method of Claim 1, wherein, In step S8, the orbital robot pose constraint constructed based on the distance error from the power distribution room feature point data to the power distribution room text plane data, and the angle error between the power distribution room text plane data and the Manhattan plane data, is a nonlinear optimization residual that satisfies the following relationship: E textplane = nT wc P c - d (4); wherein E reproj is the reprojection error, E textplane is the point-to-texture plane distance error, E plane is the electrical room texture plane and Manhattan plane angle error; In relation (4), n is the normal vector of the power distribution room text plane, and T wc For the pose transformation of the camera to the world coordinate system, P c Let d be a feature point of the power distribution room, and d be the text plane parameter of the power distribution room; In relation (5), n mw Let R be the normal vector of the Manhattan world plane, R be the rotation matrix from the camera to the world coordinate system, and n be the normal vector. text This is the normal vector of the text plane of the power distribution room.

7. The method of Claim 1, wherein, The preset pose tracking method is either the PnP method or the constant velocity model method.

8. A positioning and mapping system for a power distribution room track robot, characterized in that, include: The data acquisition module is used to acquire images of the power distribution room through the track robot, and to perform OCR text detection based on the images of the power distribution room to obtain planar text data of the power distribution room. The feature detection module is used to perform feature detection on the power distribution room image to obtain power distribution room feature point data; The feature matching module is used to match the power distribution room text plane data and the power distribution room feature point data corresponding to the power distribution room images in different frames; The Manhattan detection module is used to identify the substation image according to the Manhattan world hypothesis and extract Manhattan plane data in the substation environment; The Manhattan matching module is used to match the power distribution room text plane data with the Manhattan plane data in the current frame; The pose tracking module is used to track the pose of the track robot based on a preset pose tracking method, obtain pose data, and determine whether the power distribution room image collected by the track robot is a keyframe based on the pose data. If yes, the data conversion module is executed; otherwise, the data acquisition module is returned. The data conversion module is used to perform coordinate transformation on the power distribution room text plane data and the power distribution room feature point data according to the pose data, so as to obtain the world map text plane and the world map feature points respectively; The mapping module is used to construct the pose constraints of the track robot based on the distance error between the feature point data of the power distribution room and the text plane data of the power distribution room, as well as the angle error between the text plane data of the power distribution room and the Manhattan plane data, and to construct a three-dimensional map of the power distribution room based on the pose constraints of the track robot.

Citation Information

Patent Citations

  • Robot pose estimation method and robot point cloud map construction method and device

    CN116973883A

  • Visual SLAM method and system based on Manhattan structural surface feature constraint

    CN117490679A