Door traversal recognition method of visual robot and chip

Abstract

The invention discloses a door traversal recognition method of a visual robot and a chip, and the method comprises the steps: 1, controlling the visual robot to collect an environment visual image in real time in a process of moving along a wall, executing a first door matching process for the environment visual image to determine whether a door may exist in the environment visual image, and if yes, executing a second door matching process for the environment visual image; if so, entering step 2; 2, dividing a first room door area based on a frame of environment visual image in which a room door possibly exists, executing a room door area detection process on the first room door area, and determining the existence condition of the room door in the first room door area; and ending the room door area detection process according to the existence condition of the room door, and returning to the step 1. According to the invention, traversal identification of all doors in the environment is realized, the door identification precision is improved through a multiple identification detection mode, and the visual robot is ensured to subsequently realize high-precision mapping, navigation and other functions based on the existence condition of the doors.

Description

Technical Field

[0001] The present application relates to the field of robot door recognition, and specifically to a door traversal recognition method and chip for a visual robot. Background Art

[0002] At present, the way in which visual robots identify doors in the existing technology is mainly to determine the existence of doors by analyzing and comparing the features of the collected visual images. This single-time recognition method of randomly collecting visual images has the defects of low recognition efficiency and poor recognition accuracy. It cannot accurately identify the existence of all doors in the environment, resulting in the subsequent mapping and navigation accuracy of the visual robot based on the existence of doors being affected. Summary of the invention

[0003] This application provides a door traversal recognition method and chip for a visual robot. The specific technical solution is as follows: A door traversal and recognition method for a visual robot specifically comprises: step 1: controlling the visual robot to collect environmental visual images in real time while moving along a wall, and executing a first door matching process in real time for each frame of the collected environmental visual image to determine whether the environmental visual image meets the condition that a door may exist; if the first door matching result is that the current frame of the environmental visual image meets the condition that a door may exist, then entering step 2; if the first door matching result is that the current frame of the environmental visual image does not meet the condition that a door may exist, then continuing to execute step 1; step 2: dividing a first door area based on a frame of the environmental visual image that meets the condition that a door may exist, controlling the visual robot to execute a door area detection process for the first door area, determining the existence of a door in the first door area, terminating the door area detection process corresponding to the frame of the environmental visual image that meets the condition that a door may exist according to the existence of the door in the first door area, and returning to step 1.

[0004] Furthermore, the first door matching process is executed in real time for each frame of the collected environmental visual image to determine whether the environmental visual image meets the condition that a door may exist, specifically including: obtaining point features and line features in the current frame of the environmental visual image; performing point feature matching on the point features in the current frame of the environmental visual image and the point features in the preset door visual image to obtain the point feature matching degree of the current frame of the environmental visual image; performing line feature matching on the line features in the current frame of the environmental visual image and the line features in the preset door visual image to obtain the line feature matching degree of the current frame of the environmental visual image; calculating the possibility of the door existing in the current frame of the environmental visual image based on the point feature matching degree of the current frame of the environmental visual image and the line feature matching degree of the current frame of the environmental visual image; and determining whether the current frame of the environmental visual image meets the condition that a door may exist based on the door existence possibility degree of the current frame of the environmental visual image.

[0005] Furthermore, the method for dividing the first door area based on a frame of environmental visual image that meets the condition that a door may exist specifically includes: obtaining the position of the visual robot when collecting a frame of environmental visual image that meets the condition that a door may exist as the marked navigation position; obtaining the moving direction of the visual robot when collecting a frame of environmental visual image that meets the condition that a door may exist as the marked navigation direction; using the marked navigation position as the origin of the marked coordinate system, and using the marked navigation direction as the positive direction of the Y axis of the marked coordinate system; using the marked navigation position as the midpoint of the first area edge of the first door area, and the first area edge and the marked coordinate system are the origin ... The first area edge is parallel to the X-axis of the marking coordinate system, and the length of the first area edge is equal to twice the preset door length; based on the first area edge, a distance of twice the preset door length is extended in the positive direction of the Y-axis of the marking coordinate system to obtain a second area edge related to the first door area, the second area edge is parallel to the first area edge, and the vertical distance between the second area edge and the first area edge is equal to twice the preset door length; the two ends of the first area edge are respectively connected to the two ends of the second area edge to form a third area edge and a fourth area edge; wherein the first area edge, the second area edge, the third area edge and the fourth area edge constitute the first door area, and the lengths of the four area edges are equal.

[0006] Furthermore, the method for executing the door area detection process for the first door area specifically includes: controlling the visual robot to switch from the wall walking mode to the bow-shaped walking mode at the marked navigation position; controlling the visual robot to perform a bow-shaped movement in the first door area from the marked navigation position based on the bow-shaped walking mode to traverse the first door area, while collecting environmental visual images; controlling the visual robot to perform a second door matching process based on the environmental visual images collected in the first door area to detect the existence of the door in the first door area.

[0007] Furthermore, the control of the visual robot based on the bow-shaped walking mode starts from the marked navigation position to perform a bow-shaped movement in the first door area to traverse the first door area, specifically including: controlling the visual robot to use the midpoint of the first area edge of the first door area as the first starting point, and the connection point of the first area edge and the third area edge or the fourth area edge as the second starting point; controlling the visual robot to move in a straight line from the first starting point to the second starting point; controlling the visual robot to adjust the moving direction of the visual robot to the marked navigation direction at the second starting point; controlling the visual robot to perform a bow-shaped movement between the first area edge and the second area edge from the second starting point, so that the visual robot traverses the first door area in a bow shape.

[0008] Furthermore, the door area detection process corresponding to a frame of environmental visual image that meets the condition that a door may exist is ended according to the existence of a door in the first door area, specifically including: when the existence of a door in the first door area is determined to exist, controlling the visual robot to switch from a bow-shaped walking mode to a wall-walking mode, controlling the visual robot to move to a marked navigation position, adjusting the orientation of the visual robot to the marked navigation orientation, and ending the door area detection process corresponding to a frame of environmental visual image where a door may exist.

[0009] Furthermore, the door area detection process corresponding to a frame of environmental visual image that meets the condition that a door may exist according to the existence of a door in the first door area is terminated, and also includes: controlling the visual robot to start timing when starting to execute a bow-shaped movement in the first door area; when the duration of the bow-shaped movement of the visual robot in the first door area reaches a preset detection duration, and the existence of the door in the first door area obtained by the visual robot is that the existence of a door is not determined, controlling the visual robot to switch from a bow-shaped walking mode to a wall walking mode, controlling the visual robot to move to the marked navigation position, adjusting the orientation of the visual robot to the marked navigation orientation, and terminating the door area detection process for the first door area.

[0010] Furthermore, the door area detection process corresponding to a frame of environmental visual image that meets the condition that a door may exist according to the existence of a door in the first door area is terminated, and also includes: after controlling the visual robot to start performing a bow-shaped movement in the first door area, determining in real time whether the first door area has been traversed in a bow-shaped manner; when the visual robot has traversed the first door area in a bow-shaped manner, controlling the visual robot to switch from a bow-shaped walking mode to a wall walking mode, controlling the visual robot to move to a marked navigation position, adjusting the orientation of the visual robot to the marked navigation orientation, and terminating the door area detection process for the first door area.

[0011] Furthermore, the door traversal and recognition method of the visual robot also includes: when the visual robot determines that there is a door in the first door area, recording the first door area information as historical first door area information.

[0012] Furthermore, the door traversal and recognition method of the visual robot also includes: if the current frame of the environmental visual image meets the conditions that a door may exist, after dividing the first door area based on the current frame of the environmental visual image, the first door area is matched with the historical first door area information to determine whether the first door area divided by the current frame of the environmental visual image belongs to the historical first door area; if the first door area divided by the current frame of the environmental visual image belongs to the historical first door area, there is no need to control the visual robot to execute the door area detection process for the first door area, and the existence of the door in the first door area is determined based on the historical first door area information.

[0013] The present application also discloses a chip having a computer program stored therein. When the computer program stored therein is executed by a processor, the door traversal and recognition method of the visual robot as described above is executed.

[0014] The door traversal recognition method and chip of the visual robot described in the present application, because the setting of the doors in the environment needs to rely on the walls, the visual robot can accurately find the location of all the doors in the environment by moving along the walls, and there are door detection opportunities in the environmental visual images collected when the visual robot enters a room by following the walls, and there are door detection opportunities in the environmental visual images collected when the visual robot leaves a room by following the walls. Therefore, based on the movement along the walls, the visual robot can perform door detection on the same door when entering and leaving the room, effectively improving the door recognition rate and reducing the possibility of missed recognition. Based on the result of the first door matching process that the door may exist, the first door area is divided, so as to control the visual robot to perform further door area detection process for the first door area, effectively improving the door detection accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A flowchart of a door traversal and recognition method of a visual robot according to an embodiment of the present application. Implementation

[0016] The following will describe the embodiments of the present application in detail in conjunction with the accompanying drawings. It should be understood that the specific embodiments described below are only used to explain the present application and are not used to limit the present application.

[0017] This application provides a door traversal recognition method for a visual robot, which aims to realize traversal recognition of all doors in an environment, improve the door recognition accuracy through multiple recognition detection methods, and ensure that the visual robot can subsequently realize high-precision mapping, navigation and other functions based on the existence of doors. Specifically, Figure 1 As shown, the door traversal and recognition method of the visual robot includes: Step 1: Control the visual robot to collect environmental visual images in real time while moving along the wall, and perform the first door matching process in real time for each frame of environmental visual images collected to determine whether the environmental visual image meets the condition that a door may exist. If the first door matching result is that the current frame of environmental visual image meets the condition that a door may exist, then enter step 2. If the first door matching result is that the current frame of environmental visual image does not meet the condition that a door may exist, then continue to execute step 1. In this step, the visual robot is controlled to collect environmental visual images while moving along the wall to perform the first door matching process. Since the setting of doors in the environment needs to rely on walls, the visual robot can accurately find all the doors in the environment by moving along the wall. Moreover, there is a door detection opportunity in the environmental visual image collected when the visual robot enters a room by following the wall. There is a door detection opportunity in the environmental visual image collected when the visual robot leaves a room by following the wall. Therefore, based on the movement along the wall, the visual robot can perform door detection on the same door when entering and leaving the room, effectively improving the door recognition rate and reducing the possibility of missed recognition.

[0018] Step 2: Divide the first door area based on a frame of environmental visual image that meets the condition that a door may exist; control the visual robot to perform the door area detection process on the first door area to determine the existence of the door in the first door area; end the door area detection process corresponding to the frame of environmental visual image that meets the condition that a door may exist according to the existence of the door in the first door area, and return to step 1. In this step, based on the result of the first door matching process that meets the condition that a door may exist, the first door area is divided, so as to control the visual robot to perform a further door area detection process for the first door area, effectively improving the door detection accuracy.

[0019] As one of the implementation modes, the first door matching process is executed in real time for each frame of environmental visual image collected to determine whether the environmental visual image meets the condition that a door may exist, specifically including: obtaining point features and line features in the current frame of environmental visual image; performing point feature matching on the point features in the current frame of environmental visual image and the point features in the preset door visual image to obtain the point feature matching degree of the current frame of environmental visual image; performing line feature matching on the line features in the current frame of environmental visual image and the line features in the preset door visual image to obtain the line feature matching degree of the current frame of environmental visual image; specifically, since the point feature extraction, line feature extraction, point feature matching and line feature matching based on visual images can be achieved through conventional image feature extraction and matching algorithms, they will not be elaborated here.

[0020] The possibility of the door existing in the current frame of the environmental visual image is calculated based on the point feature matching degree of the current frame of the environmental visual image and the line feature matching degree of the current frame of the environmental visual image; specifically, the calculation of the possibility of the door existing in the current frame of the environmental visual image based on the point feature matching degree of the current frame of the environmental visual image and the line feature matching degree of the current frame of the environmental visual image refers to adding the product of the point feature matching degree and the preset point feature matching index and the product of the line feature matching degree and the preset line feature matching index, and the addition result is used as the possibility of the door existing; wherein the preset point feature matching index and the preset line feature matching index are index values ​​for adjusting the influencing factors of the point features and the line features in the possibility of the door existing. Since the door is more often expressed in the form of line features in the environmental visual image, in the calculation of the possibility of the door existing in this step, the preset line feature matching index is configured to be greater than the preset point feature matching index, which may be but is not limited to configuring the preset line feature matching index to be twice the preset point feature matching index.

[0021] Based on the door existence possibility of the current frame of the environmental visual image, it is determined whether the current frame of the environmental visual image meets the condition that a door may exist; specifically, when the door existence possibility reaches a preset door existence possibility threshold, it is determined that the current frame of the environmental visual image meets the condition that a door may exist, and the preset door existence possibility threshold is set based on the preset line feature matching index and the preset point feature matching index.

[0022] As one of the implementation modes, the method for dividing the first door area based on a frame of environmental visual image that meets the condition that a door may exist specifically includes: obtaining the position of the visual robot when collecting a frame of environmental visual image that meets the condition that a door may exist as the marked navigation position; obtaining the moving direction of the visual robot when collecting a frame of environmental visual image that meets the condition that a door may exist as the marked navigation direction; using the marked navigation position as the origin of the marked coordinate system, and using the marked navigation direction as the positive direction of the Y axis of the marked coordinate system; using the marked navigation position as the midpoint of the first area edge of the first door area, and the first area edge and the marked navigation direction are the origin of the marked coordinate system; using the marked navigation position as the midpoint of the first area edge of the first door area, and the first area edge and the marked navigation direction are the positive direction of the Y axis ... first door area. The X-axis of the coordinate system is parallel, and the length of the first area edge is equal to twice the preset door length; based on the first area edge extending twice the preset door length in the positive direction of the Y-axis of the marking coordinate system, the second area edge with the first door area is obtained, the second area edge is parallel to the first area edge, and the vertical distance between the second area edge and the first area edge is equal to twice the preset door length; the two ends of the first area edge are connected to the two ends of the second area edge respectively to form the third area edge and the fourth area edge; wherein the first area edge, the second area edge, the third area edge and the fourth area edge constitute the first door area, and the lengths of the four area edges are equal. Specifically, by recording the position of the visual robot and the moving direction of the visual robot, it can assist the visual robot to return to the end point of the wall mode after completing the door area detection process, so that the visual robot can continue the wall door recognition work.

[0023] As one implementation manner, the method for performing the door area detection process on the first door area specifically includes: The visual robot is controlled to switch from the wall walking mode to the bow-shaped walking mode at the marked navigation position; the visual robot is controlled to perform a bow-shaped movement in the first door area based on the bow-shaped walking mode from the marked navigation position to traverse the first door area, while collecting environmental visual images; the visual robot is controlled to perform a second door matching process based on the environmental visual images collected in the first door area to detect the existence of the door in the first door area. Specifically, the second door matching process and the first door matching process can be but are not limited to the same door matching process, and can also be a fusion of multiple environmental visual images and matching with a preset door visual image to calculate the possibility of the existence of the door. In this embodiment, the door area detection process is performed in the first door area near the visual robot that meets the conditions for the possible existence of a door. By controlling the visual robot to move in a bow shape in the first door area based on the bow-shaped walking mode, the visual robot can collect environmental visual images from multiple angles, and more accurately identify the existence of the door area based on the multi-angle environmental visual images, further improving the detection accuracy of the existence of the door.

[0024] As one of the implementation modes, the visual robot is controlled to perform a bow-shaped movement in the first door area from the marked navigation position based on the bow-shaped walking mode to traverse the first door area, specifically including: controlling the visual robot to use the midpoint of the first area edge of the first door area as the first starting point, and the connection point of the first area edge and the third area edge or the fourth area edge as the second starting point; controlling the visual robot to move in a straight line from the first starting point to the second starting point; controlling the visual robot to adjust the moving direction of the visual robot to the marked navigation direction at the second starting point; controlling the visual robot to perform a bow-shaped movement between the first area edge and the second area edge from the second starting point, so that the visual robot traverses in a bow shape within the first door area. Specifically, the control of the visual robot to trigger the execution of a bow-shaped movement between the first area edge and the second area edge at the second location means that the visual robot is controlled to move from the first area edge to the second area edge. When it moves to the second area edge, the visual robot is controlled to turn ninety degrees in the direction away from the second starting point and advance a preset bow-shaped interval distance, and then turn ninety degrees in the direction of the first area edge, and the visual robot is controlled to move from the second area edge to the first area edge. When it moves to the first area edge, the visual robot is controlled to turn ninety degrees in the direction away from the second starting point and advance a preset bow-shaped interval distance, and then turn ninety degrees in the direction of the second area edge, and the visual robot is controlled to move from the first area edge to the second area edge, and so on, so as to realize the visual robot to execute a bow-shaped movement between the first area edge and the second area edge. This embodiment controls the visual robot to perform bidirectional and multi-angle collection of environmental visual images in the first door area in a bow-shaped manner, and assists the visual robot in observing the existence of the door in a more diverse and accurate manner.

[0025] As one of the implementation modes, the door area detection process corresponding to a frame of environmental visual image that meets the condition that a door may exist is terminated according to the existence of a door in the first door area, specifically including: when the existence of a door in the first door area is determined to exist, the visual robot is controlled to switch from a bow-shaped walking mode to a wall-walking mode, the visual robot is controlled to move to the marked navigation position, the orientation of the visual robot is adjusted to the marked navigation orientation, and the door area detection process corresponding to a frame of environmental visual image where a door may exist is terminated. This implementation mode is based on the recorded marked navigation position and marked navigation orientation, so that the visual robot can return to the original position and restart walking along the wall after completely detecting the existence of doors in the first door area, ensuring that the visual robot does not miss every door in the environment, the door detection result is highly credible, and there is no need for the visual robot to perform wall search again in order to achieve wall walking, thereby improving the door traversal and recognition efficiency of the visual robot.

[0026] As one of the implementation modes, the door area detection process corresponding to a frame of environmental visual image that meets the condition that a door may exist according to the existence of a door in the first door area is terminated, and further includes: controlling the visual robot to start timing when the bow-shaped movement is started in the first door area; when the duration of the bow-shaped movement of the visual robot in the first door area reaches a preset detection duration, and the existence of the door in the first door area obtained by the visual robot is that the existence of the door is not determined, then controlling the visual robot to switch from the bow-shaped walking mode to the wall walking mode, controlling the visual robot to move to the marked navigation position, adjusting the direction of the visual robot to the marked navigation direction, and terminating the door area detection process of the first door area. Specifically, this implementation mode limits the duration of the bow-shaped movement of the visual robot in the first door area, and when the duration of the visual robot failing to detect the existence of the door in the first door area reaches the preset detection duration, then controlling the visual robot to terminate the door area detection process, thereby avoiding the visual robot from performing the bow-shaped movement in the same first door area for a long time to perform the door area detection process, and ensuring that the visual robot performs the door area detection process within a reasonable time range.

[0027] As one of the implementation modes, the door area detection process corresponding to a frame of the environment visual image where a door may exist is terminated according to the existence of the door in the first door area, and further includes: after the visual robot is controlled to start the bow-shaped movement in the first door area, it is judged in real time whether the first door area has been traversed in a bow-shaped manner; when the visual robot has traversed the first door area in a bow-shaped manner, the visual robot is controlled to switch from the bow-shaped walking mode to the wall-walking mode, the visual robot is controlled to move to the marked navigation position, the direction of the visual robot is adjusted to the marked navigation direction, and the door area detection process of the first door area is terminated. Specifically, the method for judging whether the first door area has been traversed in a bow-shaped manner may be, but is not limited to, judging whether the visual robot has horizontally moved the length of the first area side in the X-axis direction of the marked coordinate system starting from the marked navigation position in the first door area, and when the visual robot has horizontally moved the length of the first area side in the X-axis direction of the marked coordinate system starting from the marked navigation position in the first door area, it is determined that the visual robot has traversed the first door area in a bow-shaped manner. This embodiment limits the traversal situation of the first door area as a condition for ending the door area detection process, so that the visual robot only performs traversal once in the first door area. If the existence of a door cannot be determined after one traversal, the visual robot is controlled to end the door area detection process, thereby avoiding wasting detection computing power.

[0028] As one of the implementation modes, the door traversal and recognition method of the visual robot further includes: when the visual robot determines that there is a door in the first door area, recording the area coordinate information of the first door area as the historical first door area information. In this implementation mode, the area coordinate information of the first door area where the door is determined to exist is recorded as the historical first door area information, and by recording the historical first door area, the visual robot can have an archived record of the coordinates of the door.

[0029] As one of the implementation modes, the door traversal and recognition method of the visual robot also includes: if there may be a door in the current frame of the environmental visual image, after dividing the first door area based on the current frame of the environmental visual image, the first door area is matched with the historical first door area information to determine whether the first door area divided by the current frame of the environmental visual image belongs to the historical first door area; if the first door area divided by the current frame of the environmental visual image belongs to the historical first door area, there is no need to control the visual robot to perform the door area detection process on the first door area, and the existence of the door in the first door area is determined based on the historical first door area information. This implementation mode is based on matching the first door area with the historical first door area information, so that the visual robot does not repeatedly perform the door area detection process for the area where the door is determined to exist, thereby improving the execution rationality of the door area detection process and improving the door recognition efficiency of the visual robot.

[0030] In one embodiment of the present application, a chip is provided, in which a computer program is stored. When the computer program stored in the chip is executed by a processor, the door traversal and recognition method of a visual robot as described in any of the above embodiments is executed.

[0031] Obviously, the above embodiments are only some embodiments of the present invention, not all embodiments, and the technical solutions between the various embodiments can be combined with each other. If the terms "first", "second", "third" and the like appear in the embodiments, they are for the convenience of distinguishing the relevant features and cannot be understood as indicating or implying their relative importance, order or number of technical features.

[0032] Those skilled in the art will understand that all or part of the steps in the above-mentioned implementation methods can be completed by instructing the relevant hardware through a program, and the program is stored in a storage medium, including several instructions for making a single-chip microcomputer, a chip or a processor (processor) execute all or part of the steps of the method described in each implementation of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0033] It should be noted that any process or method description in the flowchart or otherwise described herein may be understood to represent a module, fragment or portion of a code comprising one or more executable instructions for implementing the steps of a specific logical function or process, and that the scope of the preferred embodiments of the present invention includes alternative implementations in which the functions may not be performed in the order described or discussed, including performing the functions in a substantially simultaneous manner or in the reverse order according to the functions involved, which should be understood by those skilled in the art to which the embodiments of the present invention belong.

[0034] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A door traversal recognition method for a visual robot, characterized in that: The door traversal and recognition method of the visual robot specifically includes: Step 1: Control the visual robot to collect environmental visual images in real time while moving along the wall, and perform the first door matching process in real time for each frame of environmental visual images collected to determine whether the environmental visual image meets the condition that a door may exist. If the first door matching result is that the current frame of environmental visual image meets the condition that a door may exist, then proceed to step 2; if the first door matching result is that the current frame of environmental visual image does not meet the condition that a door may exist, then continue to execute step 1; Step 2: Divide the first door area based on a frame of environmental visual image that meets the condition that a door may exist, control the visual robot to execute the door area detection process for the first door area, determine the existence of the door in the first door area, end the door area detection process corresponding to the frame of environmental visual image that meets the condition that a door may exist according to the existence of the door in the first door area, and return to step 1.

2. The door traversal recognition method of a visual robot according to claim 1, characterized in that: The first door matching process is executed in real time on each frame of the collected environment visual image to determine whether the environment visual image meets the condition that a door may exist, specifically including: Obtain point features and line features in the current frame of environmental visual image; Performing point feature matching on the point features in the current frame of the environment visual image and the point features in the preset door visual image to obtain the point feature matching degree of the current frame of the environment visual image; Perform line feature matching on the line features in the current frame of the environment visual image and the line features in the preset door visual image to obtain the line feature matching degree of the current frame of the environment visual image; Calculate the possibility of the door existing in the current frame of the environmental visual image based on the point feature matching degree of the current frame of the environmental visual image and the line feature matching degree of the current frame of the environmental visual image; Based on the door existence possibility of the current frame of the environmental visual image, it is determined whether the current frame of the environmental visual image meets the condition that a door may exist.

3. The door traversal recognition method of a visual robot according to claim 1, characterized in that: The method for dividing the first door area based on a frame of environmental visual image that meets the condition that a door may exist specifically includes: The position of the visual robot when acquiring a frame of environmental visual image that meets the condition that a door may exist is used as a marked navigation position; When acquiring a frame of environment visual image that meets the condition that a door may exist, the moving direction of the visual robot is the marked navigation direction; The marked navigation position is taken as the origin of the marked coordinate system, and the marked navigation direction is taken as the positive direction of the Y axis of the marked coordinate system; The marked navigation position is used as the midpoint of the first area edge of the first door area, the first area edge is parallel to the X-axis of the marked coordinate system, and the length of the first area edge is equal to twice the preset door length; Based on the first area edge extending in the positive direction of the Y axis of the marking coordinate system by twice the preset door length, a second area edge in the first door area is obtained, the second area edge is parallel to the first area edge, and the vertical distance between the second area edge and the first area edge is equal to twice the preset door length; Connecting two ends of the first region edge and two ends of the second region edge respectively to form a third region edge and a fourth region edge; The first area edge, the second area edge, the third area edge and the fourth area edge constitute the first door area, and the lengths of the four area edges are equal.

4. The door traversal recognition method of a visual robot according to claim 3 is characterized in that: The method for performing a door area detection process on a first door area specifically includes: Control the visual robot to switch from the wall-walking mode to the bow-shaped walking mode at the marked navigation position; Control the visual robot to perform a bow-shaped movement in the first door area from the marked navigation position based on the bow-shaped walking mode to traverse the first door area, while collecting environmental visual images; The visual robot is controlled to perform a second door matching process based on the environmental visual image collected in the first door area to detect the existence of the door in the first door area.

5. The door traversal recognition method of a visual robot according to claim 4, characterized in that: The controlling the visual robot to perform a bow-shaped movement in the first door area from the marked navigation position based on the bow-shaped walking mode to traverse the first door area specifically includes: Control the visual robot to use the midpoint of the first area edge of the first door area as the first starting point, and use the connection point of the first area edge and the third area edge or the fourth area edge as the second starting point; Control the visual robot to move in a straight line from the first starting point to the second starting point; Controlling the visual robot to adjust the moving direction of the visual robot to the marked navigation direction at the second starting point; The visual robot is controlled to start from the second starting point and perform a bow-shaped movement between the first area edge and the second area edge, so that the visual robot traverses in the first door area in a bow shape.

6. The door traversal recognition method of a visual robot according to claim 4, characterized in that: The method of terminating the door area detection process corresponding to a frame of environmental visual image that meets the condition that a door may exist according to the existence of a door in the first door area specifically includes: when the existence of a door in the first door area is determined to be a door, controlling the visual robot to switch from a bow-shaped walking mode to a wall-walking mode, controlling the visual robot to move to a marked navigation position, adjusting the orientation of the visual robot to the marked navigation orientation, and terminating the door area detection process corresponding to a frame of environmental visual image where a door may exist.

7. The door traversal recognition method of a visual robot according to claim 6, characterized in that: The process of terminating the door area detection process corresponding to a frame of the ambient visual image that meets the condition that a door may exist according to the existence of the door in the first door area also includes: Control the visual robot to start timing when it starts to perform bow-shaped movement in the first door area; When the duration of the bow-shaped movement performed by the visual robot in the first door area reaches the preset detection duration, and the door existence status in the first door area obtained by the visual robot is that the existence of a door is not determined, the visual robot is controlled to switch from the bow-shaped walking mode to the wall walking mode, the visual robot is controlled to move to the marked navigation position, the direction of the visual robot is adjusted to the marked navigation direction, and the door area detection process of the first door area is ended.

8. The door traversal recognition method of a visual robot according to claim 6, characterized in that: The process of terminating the door area detection process corresponding to a frame of the ambient visual image that meets the condition that a door may exist according to the existence of the door in the first door area also includes: After the visual robot starts to move in a bow shape in the first door area, it is determined in real time whether the first door area has been traversed in a bow shape; When the visual robot has traversed the first door area in a bow shape, the visual robot is controlled to switch from the bow-shaped walking mode to the wall walking mode, the visual robot is controlled to move to the marked navigation position, and the direction of the visual robot is adjusted to the marked navigation direction, thereby ending the door area detection process of the first door area.

9. The door traversal recognition method of a visual robot according to claim 1, characterized in that: The door traversal and identification method of the visual robot also includes: when the visual robot determines that there is a door in the first door area, recording the first door area information as historical first door area information.

10. The door traversal recognition method of a visual robot according to claim 9, characterized in that: The door traversal and recognition method of the visual robot also includes: If the current frame of the environmental visual image meets the condition that there may be a door, after dividing the first door area based on the current frame of the environmental visual image, the first door area is matched with the historical first door area information to determine whether the first door area divided by the current frame of the environmental visual image belongs to the historical first door area; If the first door area divided by the current frame of environmental visual image belongs to the historical first door area, there is no need to control the visual robot to execute the door area detection process for the first door area, and the existence of the door in the first door area can be determined based on the historical first door area information.

11. A chip having a computer program stored therein, characterized in that: The computer program stored in the chip is executed by the processor when it is run by the processor to execute the door traversal and recognition method of the visual robot as described in any one of claims 1 to 10.

Images