Robot and navigation method thereof

Abstract

The present invention discloses a robot and a navigation method thereof, the navigation method comprising: collecting an image of a working scene through the camera; calculating a reference direction of the working scene according to the collected image of the working scene; correcting the direction of the coordinate axis in the robot's cruise coordinate system with the reference direction; and navigating the robot using the corrected cruise coordinate system. When the robot is turned on or in the working process, the present invention searches for and selects a group of mutually parallel lines with the largest number of lines parallel or perpendicular to them as reference lines in the collected image of the working scene, uses the direction of the reference line as the reference direction, and corrects the direction of the coordinate axis in the robot's cruise coordinate system with the reference direction, so that the robot works in a scientific and reasonable navigation path, which is conducive to improving the robot's work efficiency.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a robot and a navigation method thereof. Background Art

[0002] Artificial Intelligence (AI) has been a very hot field in recent years. It is a technical science that studies and develops theories, methods, technologies and application systems for simulating, extending and expanding human intelligence.

[0003] A robot is a machine device that performs tasks automatically. It can act according to the action patterns set by artificial intelligence, assisting or replacing humans in doing certain jobs, such as manufacturing, construction, etc. Summary of the invention

[0004] The technical problem solved by the present invention is how a robot can work in a scientific and reasonable navigation path.

[0005] In order to solve the above technical problems, an embodiment of the present invention provides a navigation method of a robot, wherein a camera is installed on the robot, and the navigation method comprises:

[0006] Collecting images of the working scene through the camera;

[0007] Calculating a reference direction of the working scene according to the collected image of the working scene;

[0008] Correcting the direction of the coordinate axis in the robot's cruising coordinate system using the reference direction;

[0009] The corrected cruise coordinate system is used to navigate the robot.

[0010] Optionally, calculating a reference direction of the work scene according to the collected image of the work scene includes:

[0011] Performing homography transformation on the collected image of the working scene to obtain a transformed image, wherein an image plane of the transformed image is parallel to the ceiling of the working scene;

[0012] A search is performed in the transformed image to identify line segments on the ceiling of the working scene.

[0013] Optionally, calculating a reference direction of the work scene according to the collected image of the work scene includes:

[0014] A search is performed in the collected image of the work scene to identify line segments on the ceiling of the work scene.

[0015] Optionally, calculating a reference direction of the work scene according to the collected image of the work scene includes:

[0016] Among the searched line segments, a line segment that meets a preset condition is selected as a reference line, and a reference direction is determined according to the direction of the reference line.

[0017] Optionally, the preset condition is: being parallel to other searched line segments.

[0018] Optionally, the preset condition is: being perpendicular to other searched line segments.

[0019] Optionally, the method of correcting the direction of the coordinate axis in the cruising coordinate system of the robot with the reference direction includes:

[0020] Calculating a first angle, where the first angle is the angle between the main axis direction of the image and the direction of the coordinate axis in the cruising coordinate system of the robot before correction;

[0021] Calculating a second angle, where the second angle is the angle between the reference direction and the main axis direction of the image;

[0022] Calculate a third angle according to the first angle and the second angle, wherein the third angle is the angle between the reference direction and the direction of the coordinate axis in the cruising coordinate system of the robot before correction;

[0023] The directions of the coordinate axes in the robot's cruising coordinate system are corrected according to the third angle.

[0024] Optionally, the camera captures the image of the working scene when the robot is turned on or when the robot is working.

[0025] Optionally, the robot is controlled to rotate and / or move so that the shooting area of ​​the robot's camera rotates once, and the camera captures images of the working scene at predetermined time intervals.

[0026] In order to solve the above technical problems, an embodiment of the present invention further provides a robot, comprising: an image acquisition unit, a calculation unit, a coordinate correction unit and a navigation unit; wherein:

[0027] An image acquisition unit, which acquires images of the working scene through the camera;

[0028] A calculation unit, which calculates a reference direction of the working scene according to the collected image of the working scene;

[0029] A coordinate correction unit, which corrects the direction of the coordinate axis in the robot's cruising coordinate system with the reference direction;

[0030] The navigation unit uses the corrected cruise coordinate system to navigate the robot.

[0031] Optionally, the calculating unit calculates the reference direction of the working scene according to the collected image of the working scene, including:

[0032] Performing homography transformation on the collected image of the working scene to obtain a transformed image, wherein an image plane of the transformed image is parallel to the ceiling of the working scene;

[0033] A search is performed in the transformed image to identify line segments on the ceiling of the working scene.

[0034] Optionally, the calculating unit calculates the reference direction of the working scene according to the collected image of the working scene, including:

[0035] A search is performed in the collected image of the work scene to identify line segments on the ceiling of the work scene.

[0036] Optionally, the calculating unit calculates a reference direction of the working scene according to the collected image of the working scene, including:

[0037] Among the searched line segments, a line segment that meets a preset condition is selected as a reference line, and a reference direction is determined according to the direction of the reference line.

[0038] Optionally, the preset condition is: being parallel to other searched line segments.

[0039] Optionally, the preset condition is: being perpendicular to other searched line segments.

[0040] Optionally, the coordinate correction unit corrects the direction of the coordinate axis in the cruising coordinate system of the robot with the reference direction, including:

[0041] Calculating a first angle, where the first angle is the angle between the main axis direction of the image and the direction of the coordinate axis in the cruising coordinate system of the robot before correction;

[0042] Calculating a second angle, where the second angle is the angle between the reference direction and the main axis direction of the image;

[0043] Calculate a third angle according to the first angle and the second angle, wherein the third angle is the angle between the reference direction and the direction of the coordinate axis in the cruising coordinate system of the robot before correction;

[0044] The directions of the coordinate axes in the robot's cruising coordinate system are corrected according to the third angle.

[0045] Optionally, the camera of the image acquisition unit acquires the image of the working scene when the robot is turned on or when the robot is working.

[0046] Optionally, the robot is controlled to rotate and / or move so that the shooting area of ​​the camera of the robot rotates one circle, and the image acquisition unit acquires the image of the working scene through the camera at predetermined time intervals.

[0047] Compared with the prior art, the technical solution of the present invention has the following beneficial effects:

[0048] When the robot is turned on or in the working process, an image of the working scene is collected. In the collected image of the working scene, a group of mutually parallel lines with the largest number of lines parallel or perpendicular to them is searched and selected as reference lines. The direction of the reference line is used as the reference direction. The reference direction is used to correct the direction of the coordinate axis in the robot's cruise coordinate system. The corrected cruise coordinate system is used to navigate the robot, so that the robot works in a scientific and reasonable navigation path, which is conducive to improving the robot's work efficiency.

[0049] Furthermore, the image plane of the camera is adjusted to be parallel to the ceiling of the working scene, and the image of the ceiling of the working scene is collected by the camera. When the image plane is parallel to the ceiling, the line structure on the ceiling will not change in the image, and the parallel and perpendicular relationship between the lines will still be maintained. In order to utilize the above characteristics, the image plane of the camera is adjusted to be parallel to the ceiling of the working scene using the homography transformation method (for example, in the case of forward vision), thereby eliminating the calculation amount of image transformation, which is conducive to completing the correction of the robot's cruising coordinate system with higher efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 is a flow chart of a navigation method of a robot in an embodiment of the present invention;

[0051] Figure 2 A sub-step flow chart of a navigation method for a robot in an embodiment of the present invention;

[0052] Figure 3 is another sub-step flow chart of the navigation method of the robot in an embodiment of the present invention;

[0053] Figure 4 is another sub-step flow chart of the navigation method of the robot in an embodiment of the present invention;

[0054] Figure 5 Schematic diagram of a sweeping robot in an embodiment of the present invention. DETAILED DESCRIPTION

[0055] The present invention relates to a sweeping robot (also known as an automatic sweeper, intelligent vacuum cleaner, robot vacuum cleaner, floor robot, etc.), which is a kind of intelligent household appliance. After being turned on, the sweeping robot can automatically complete the cleaning work of the floor in the room by virtue of certain artificial intelligence.

[0056] When a sweeping robot is performing cleaning work, how to scientifically and reasonably plan its cleaning path is an important factor that has a direct impact on work efficiency.

[0057] Early sweeping robots had fool-proof collision sensing and would only turn around after colliding with an obstacle, with a low level of intelligence.

[0058] Some advanced sweeping robots use cameras, lasers, magnetic induction, etc. to sense the surrounding environment, and use certain algorithms to plan a reasonable cleaning path to avoid missed cleaning and repeated cleaning as much as possible.

[0059] A more scientific cleaning path is the "bow" cleaning. The robot's navigation system establishes a cruising coordinate system, cleaning in a straight line along the coordinate axis (x-axis, y-axis) or in a direction parallel to the coordinate axis, and turning at right angles when necessary.

[0060] In the prior art, there are various algorithms and sensors used by sweeping robots to establish a cruising coordinate system. For example, a magnetic sensor can be used to identify an origin magnetic pin, and a single-frame image can be collected based on a starting position to establish a coordinate system.

[0061] Specifically, the sweeping robot is started at the origin magnetic pin, and the sweeping robot identifies the origin magnetic pin through its own magnetic sensor, and determines the position of the origin magnetic pin as the coordinate origin of the cruise coordinate system.

[0062] The above-mentioned existing technical solutions are helpful to plan a better navigation path, but there is still room for further improvement.

[0063] After research, it was found that the direction of the coordinate axis of the cruise coordinate system will also have a significant impact on the robot's work efficiency. In practice, sweeping robots usually work in indoor environments, and sweeping along the direction of the house wall can effectively avoid missing sweeps. To this end, the present invention designs a robot navigation method (and device) based on vision, which can enable the robot to sweep along the direction of the house wall as much as possible.

[0064] The present invention collects images of the working scene when the robot is turned on or in the working process, searches for and selects a group of mutually parallel lines with the largest number of lines parallel or perpendicular to them as reference lines in the collected images of the working scene, uses the direction of the reference line as the reference direction, uses the reference direction to correct the direction of the coordinate axis in the robot's cruising coordinate system, and uses the corrected cruising coordinate system to navigate the robot, so that the robot works in a scientific and reasonable navigation path, which is beneficial to improving the working efficiency of the robot.

[0065] In order to enable those skilled in the art to better understand and implement the present invention, the present invention is described in detail below through specific embodiments with reference to the accompanying drawings.

[0066] Embodiment 1

[0067] As described below, an embodiment of the present invention provides a navigation method for a robot.

[0068] First, the implementation of the robot navigation method in this embodiment requires that the robot has a visual perception function, for example, it can be a robot based on VSLAM visual positioning.

[0069] Specifically, a camera may be installed on the robot.

[0070] In the subsequent steps, the images captured by the camera are needed to calibrate the cruise coordinate system of the robot navigation system.

[0071] In addition, the robot also needs to have certain computing capabilities, such as a built-in processor.

[0072] The present invention is applicable to various household robot products with visual sensors, such as sweeping robots and the like.

[0073] Reference Figure 1 The flowchart of the robot navigation method is shown in the figure, and the following is a detailed description through specific steps:

[0074] S101, collecting images of a working scene through the camera.

[0075] The camera here is the camera installed on the robot mentioned above.

[0076] In some embodiments, after the robot is turned on, sometimes the robot is turned on close to a wall. In this case, most or all of the image captured by the robot's camera is information about the wall, while in the image, there is very little information about the line segments on the ceiling. Therefore, the robot needs to rotate one circle or walk a certain distance (i.e., rotate and / or move) to capture the image, so as to capture useful photos of the ceiling (as images of the working scene), and then (search and) extract useful line segments therefrom.

[0077] In some embodiments, reference Figure 2 As shown in the sub-step flowchart, the step of collecting an image of the working scene by the camera may specifically include the following sub-steps:

[0078] S1011, adjust the camera to a top view state.

[0079] The difference between using top view and front view is:

[0080] When the robot camera looks up, the image plane is parallel to the ceiling, the line structure on the ceiling does not change in the image, and the parallel and perpendicular relationship between the lines remains unchanged. Using this feature, we search for parallel and perpendicular line segments on the image plane, calculate the angle between the line direction and the main axis of the image, and use it as the rotation angle required for the machine to turn straight.

[0081] When the robot's perspective is a front perspective, there is a certain angle between the image plane and the ceiling. Due to the perspective projection relationship of the camera, the line structure on the ceiling will change in the image, and the parallel and perpendicular relationship between the lines is difficult to maintain. The images taken at different perspectives can be converted using the homography transformation method to obtain a transformed image. The image plane of the transformed image is parallel to the ceiling of the work scene (Note: two images of the same plane taken by the same camera at different postures can be converted by homography transformation). For example, the front view image is transformed into a top view image using the homography matrix to restore the parallel and perpendicular relationship between the ceiling lines.

[0082] S1012: Collect an image of the ceiling of the working scene through the camera.

[0083] It can be seen from the above description of the technical solution that: in this embodiment, when the robot is turned on or in the working process, an image of the working scene is collected, and in the collected image of the working scene, a group of mutually parallel lines with the largest number of lines parallel or perpendicular to them is searched and selected as reference lines, and the direction of the reference line is used as the reference direction, and the reference direction is used to correct the direction of the coordinate axis in the robot's cruise coordinate system, and the corrected cruise coordinate system is used to navigate the robot, so that the robot works with a scientific and reasonable navigation path, which is conducive to improving the work efficiency of the robot.

[0084] In some embodiments, the timing of executing the acquisition of the image of the working scene by the camera may be when the robot is turned on, or may be during the operation of the robot.

[0085] For example, after powering on, the robot is controlled to rotate and / or move so that the shooting area of ​​the robot's camera rotates one circle (one rotation can avoid the inability to extract the linear structure of the ceiling due to the limited field of view caused by the robot's initial position), and images of the working scene are collected at predetermined time intervals.

[0086] For another example, during the operation of the robot, the camera collects images of the working scene at predetermined time intervals, and executes subsequent steps based on the images, thereby performing real-time correction of the cruise coordinate system during the operation of the robot.

[0087] The predetermined time interval may be 200 ms.

[0088] In the method provided in this embodiment, before the method starts, the robot can be in any orientation.

[0089] S102: Calculate a reference direction of the working scene according to the collected image of the working scene.

[0090] After the image of the working scene is obtained, a reference direction of the working scene is calculated based on the collected image of the working scene.

[0091] In some embodiments, based on the robot posture data, the vector expression of the direction of the reference line in the robot's initial coordinate system can be calculated, and then the perpendicular and parallel directions in all images are counted, and the direction of the reference line with the largest number is taken as the direction of the room wall (i.e., the reference direction of the working scene), and then the direction of the cruise coordinate system is corrected to make it consistent.

[0092] In some embodiments, reference Figure 3 As shown in the sub-step flowchart, the reference direction of the working scene is calculated based on the collected image of the working scene, which may specifically include the following sub-steps:

[0093] S1021, searching in the collected image of the work scene to identify line segments on the ceiling of the work scene.

[0094] In another embodiment, in the previous step, a homography transformation is performed on the captured image of the work scene to obtain a transformed image, and the image plane of the transformed image is parallel to the ceiling of the work scene; a search is performed in the transformed image to identify line segments on the ceiling of the work scene.

[0095] Specifically, a length threshold may be preset, line segments whose length is less than the length threshold may be ignored, and line segments whose length reaches the length threshold may be searched on the ceiling of the work scene.

[0096] S1022: Select a line segment that meets a preset condition from among the searched line segments as a reference line, and determine a reference direction according to the direction of the reference line.

[0097] In some embodiments, the preset condition is: being parallel to other searched line segments.

[0098] In some embodiments, the preset condition is: being perpendicular to other searched line segments.

[0099] In some embodiments, the preset condition is: the reference score is the highest.

[0100] The reference score of the line segment is: the weighted total score of other line segments parallel or perpendicular to the line segment, wherein the longer the length of other line segments parallel or perpendicular to the line segment, the greater the weight.

[0101] For example, among the searched line segments, there are 4 line segments that are parallel or perpendicular to line segment A. According to the lengths of these 4 line segments, the weights of these 4 line segments are 5, 3, 2, and 2 respectively (the longer the length, the greater the weight);

[0102] Then the weighted total score of line segment A is 1×(5+3+2+2)=12;

[0103] There are 3 line segments parallel or perpendicular to line segment B. According to the lengths of these 3 line segments, the weights of these 3 line segments are 7, 6, and 4 respectively (the longer the length, the greater the weight);

[0104] Then the weighted total score of line segment B is 1×(7+6+4)=17;

[0105] Compare the weighted total scores of each line segment. The weighted total score of line segment A is less than the weighted total score of line segment B (of course, there may also be line segment C, line segment D, etc., which are not repeated here). In this case, select line segment B as the reference line.

[0106] In some embodiments, the direction vectors of these reference lines in the initial world system may be calculated, and whether the lines are parallel or perpendicular may be determined based on the direction vectors.

[0107] S103: Correct the directions of the coordinate axes in the robot's cruising coordinate system using the reference direction.

[0108] In some embodiments, reference Figure 4 As shown in the sub-step flow chart, the step of correcting the direction of the coordinate axis in the robot's cruise coordinate system with the reference direction may specifically include the following sub-steps:

[0109] S1031, calculating a first angle.

[0110] The first angle is the angle between the main axis direction of the image and the direction of the coordinate axis in the cruising coordinate system of the robot before correction.

[0111] In some embodiments, the main axis direction of the image is the same as the direction of the coordinate axis in the cruise coordinate system of the robot before correction (for example, the direction of the coordinate axis in the initial cruise coordinate system), such as the case of a top-view robot. In this case, the first angle is 0.

[0112] In some other embodiments, there are also cases where the main axis direction of the image is different from the direction of the coordinate axis in the cruising coordinate system of the robot before correction, such as when looking forward.

[0113] S1032, calculating a second angle.

[0114] The second angle is the angle between the reference direction and the main axis direction of the image.

[0115] S1033: Calculate a third angle according to the first angle and the second angle.

[0116] The third angle is the angle between the reference direction and the direction of the coordinate axis in the cruising coordinate system of the robot before correction.

[0117] S1034: Correct the directions of the coordinate axes in the cruising coordinate system of the robot according to the third angle.

[0118] Subsequently, the orientation of the robot can be rotated according to the third angle.

[0119] S104, using the corrected cruise coordinate system to navigate the robot.

[0120] In some embodiments, the robot can be arranged to perform cleaning and other tasks along a "bow"-shaped path.

[0121] It can be seen from the above description of the technical solution that: in this embodiment, when the robot is turned on or in the working process, an image of the working scene is collected, and in the collected image of the working scene, a group of mutually parallel lines with the largest number of lines parallel or perpendicular to them is searched and selected as reference lines, and the direction of the reference line is used as the reference direction, and the reference direction is used to correct the direction of the coordinate axis in the robot's cruise coordinate system, and the corrected cruise coordinate system is used to navigate the robot, so that the robot works with a scientific and reasonable navigation path, which is conducive to improving the work efficiency of the robot.

[0122] In addition, the present invention does not bring additional hardware costs, for example, there is no need to add new hardware sensors, etc., and the problem of robot cruise coordinate system correction is solved by using an algorithm, which is simple and convenient and does not increase costs.

[0123] Embodiment 2

[0124] As described below, an embodiment of the present invention provides a robot.

[0125] In some embodiments, Figure 5 As shown, the robot may be a sweeping robot, or other household robot products with visual sensors.

[0126] The robot includes: an image acquisition unit, a calculation unit, a coordinate correction unit and a navigation unit; the main functions of each unit are as follows:

[0127] An image acquisition unit, which acquires images of the working scene through the camera;

[0128] A calculation unit, which calculates a reference direction of the working scene according to the collected image of the working scene;

[0129] A coordinate correction unit, which corrects the direction of the coordinate axis in the robot's cruising coordinate system with the reference direction;

[0130] The navigation unit uses the corrected cruise coordinate system to navigate the robot.

[0131] It can be seen from the above description of the technical solution that: in this embodiment, when the robot is turned on or in the working process, an image of the working scene is collected, and in the collected image of the working scene, a group of mutually parallel lines with the largest number of lines parallel or perpendicular to them is searched and selected as reference lines, and the direction of the reference line is used as the reference direction, and the reference direction is used to correct the direction of the coordinate axis in the robot's cruise coordinate system, and the corrected cruise coordinate system is used to navigate the robot, so that the robot works with a scientific and reasonable navigation path, which is conducive to improving the work efficiency of the robot.

[0132] In some embodiments, Figure 2 , Figure 3 As shown, the calculation unit calculates the reference direction of the working scene according to the collected image of the working scene, including:

[0133] Performing homography transformation on the collected image of the working scene to obtain a transformed image, wherein an image plane of the transformed image is parallel to the ceiling of the working scene;

[0134] A search is performed in the transformed image to identify line segments on the ceiling of the working scene.

[0135] It can be seen from the above description of the technical solution that: in this embodiment, when the robot is turned on or in the working process, an image of the working scene is collected, and in the collected image of the working scene, a group of mutually parallel lines with the largest number of lines parallel or perpendicular to them is searched and selected as reference lines, and the direction of the reference line is used as the reference direction, and the reference direction is used to correct the direction of the coordinate axis in the robot's cruise coordinate system, and the corrected cruise coordinate system is used to navigate the robot, so that the robot works with a scientific and reasonable navigation path, which is conducive to improving the work efficiency of the robot.

[0136] In some embodiments, Figure 3 As shown, the calculation unit calculates the reference direction of the working scene according to the collected image of the working scene, including:

[0137] A search is performed in the collected image of the work scene to identify line segments on the ceiling of the work scene.

[0138] In some embodiments, the calculation unit calculates the reference direction of the working scene according to the collected image of the working scene, including:

[0139] Among the searched line segments, a line segment that meets a preset condition is selected as a reference line, and a reference direction is determined according to the direction of the reference line.

[0140] In some embodiments, the preset condition is: being parallel to other searched line segments.

[0141] In some embodiments, the preset condition is: being perpendicular to other searched line segments.

[0142] In some embodiments, Figure 4 As shown, the coordinate correction unit corrects the direction of the coordinate axis in the robot's cruise coordinate system with the reference direction, including:

[0143] Calculating a first angle, where the first angle is the angle between the main axis direction of the image and the direction of the coordinate axis in the cruising coordinate system of the robot before correction;

[0144] Calculating a second angle, where the second angle is the angle between the reference direction and the main axis direction of the image;

[0145] Calculate a third angle according to the first angle and the second angle, wherein the third angle is the angle between the reference direction and the direction of the coordinate axis in the cruising coordinate system of the robot before correction;

[0146] The directions of the coordinate axes in the robot's cruising coordinate system are corrected according to the third angle.

[0147] In some embodiments, the camera of the image acquisition unit acquires the image of the working scene when the robot is turned on or when the robot is working.

[0148] In some embodiments, the robot is controlled to rotate and / or move so that the shooting area of ​​the robot's camera rotates one circle, and the image acquisition unit captures images of the working scene through the camera at predetermined time intervals.

[0149] A person skilled in the art can understand that, in the various methods of the above embodiments, all or part of the steps can be completed by hardware related to program instructions, and the program can be stored in a computer-readable storage medium, which may include: ROM, RAM, disk or CD, etc.

[0150] Although the present invention is disclosed as above, the present invention is not limited thereto. Any person skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the scope defined by the claims.

Claims

1. A robot navigation method, characterized in that: The robot is equipped with a camera, and the navigation method includes: Collecting images of the working scene through the camera; Searching in the collected image of the work scene to identify line segments on the ceiling of the work scene; Among the searched line segments, a group of mutually parallel lines with the largest number of mutually parallel or mutually perpendicular lines to the other searched line segments is selected as a reference line, and a reference direction is determined according to the direction of the reference line; Correcting the direction of the coordinate axis in the robot's cruising coordinate system using the reference direction; The corrected cruise coordinate system is used to navigate the robot.

2. The robot navigation method according to claim 1, characterized in that: Searching in the collected image of the work scene to identify line segments on the ceiling of the work scene includes: Performing homography transformation on the collected image of the working scene to obtain a transformed image, wherein an image plane of the transformed image is parallel to the ceiling of the working scene; A search is performed in the transformed image to identify line segments on the ceiling of the working scene.

3. The robot navigation method according to claim 1, characterized in that: Searching in the collected image of the work scene to identify line segments on the ceiling of the work scene includes: A search is performed in the collected image of the work scene to identify line segments on the ceiling of the work scene.

4. The robot navigation method according to claim 1, characterized in that: The method of correcting the direction of the coordinate axis in the robot's cruise coordinate system by using the reference direction includes: Calculating a first angle, where the first angle is the angle between the main axis direction of the image and the direction of the coordinate axis in the cruising coordinate system of the robot before correction; Calculating a second angle, where the second angle is the angle between the reference direction and the main axis direction of the image; Calculate a third angle according to the first angle and the second angle, wherein the third angle is the angle between the reference direction and the direction of the coordinate axis in the cruising coordinate system of the robot before correction; The directions of the coordinate axes in the robot's cruising coordinate system are corrected according to the third angle.

5. The robot navigation method according to claim 1, characterized in that: The camera collects images of the working scene when the robot is turned on or when the robot is working.

6. The robot navigation method according to claim 1, characterized in that: The robot is controlled to rotate and / or move so that the shooting area of ​​the robot's camera rotates one circle, and the camera collects images of the working scene at predetermined time intervals.

7. A robot, characterized in that: include: Image acquisition unit, calculation unit, coordinate correction unit and navigation unit; wherein: An image acquisition unit, which acquires images of the working scene through a camera; The computing unit searches in the collected image of the working scene to identify line segments on the ceiling of the working scene; among the searched line segments, a group of parallel lines with the largest number of parallel or perpendicular lines to other searched line segments is selected as a reference line, and a reference direction is determined according to the direction of the reference line; A coordinate correction unit, which corrects the direction of the coordinate axis in the robot's cruising coordinate system with the reference direction; The navigation unit uses the corrected cruise coordinate system to navigate the robot.

8. The robot according to claim 7, characterized in that: The computing unit searches in the collected image of the working scene to identify line segments on the ceiling of the working scene, including: Performing homography transformation on the collected image of the working scene to obtain a transformed image, wherein an image plane of the transformed image is parallel to the ceiling of the working scene; A search is performed in the transformed image to identify line segments on the ceiling of the working scene.

9. The robot according to claim 7, characterized in that: The computing unit searches in the collected image of the working scene to identify line segments on the ceiling of the working scene, including: A search is performed in the collected image of the work scene to identify line segments on the ceiling of the work scene.

10. The robot according to claim 7, characterized in that: The coordinate correction unit corrects the direction of the coordinate axis in the robot's cruising coordinate system with the reference direction, including: Calculating a first angle, where the first angle is the angle between the main axis direction of the image and the direction of the coordinate axis in the cruising coordinate system of the robot before correction; Calculating a second angle, where the second angle is the angle between the reference direction and the main axis direction of the image; Calculate a third angle according to the first angle and the second angle, wherein the third angle is the angle between the reference direction and the direction of the coordinate axis in the cruising coordinate system of the robot before correction; The directions of the coordinate axes in the robot's cruising coordinate system are corrected according to the third angle.

11. The robot according to claim 7, characterized in that: The timing for the camera of the image acquisition unit to acquire the image of the working scene is when the robot is turned on or when the robot is working.

12. The robot according to claim 7, characterized in that: The robot is controlled to rotate and / or move so that the shooting area of ​​the camera of the robot rotates one circle, and the image acquisition unit acquires images of the working scene through the camera at predetermined time intervals.

Images