Obstacle avoidance method and system for an automatic guided logistics robot

Abstract

The present invention relates to the field of data processing technology, and in particular to an obstacle avoidance method and system for an automatic guided logistics robot. The method comprises: obtaining all feature points of the logistics robot in the environment at the current moment; obtaining the moving direction of the logistics robot, constructing a plane coordinate system in the moving direction, and calculating the uniformity of all feature points; determining the neighborhood feature points among all feature points, and obtaining the uniformity of the neighborhood feature points; calculating the local uniformity of the environment; calculating the feature distance between any two neighborhood feature points; clustering the neighborhood feature points using K-means clustering based on the feature distance, and determining obstacle points to achieve obstacle avoidance for the automatic guided logistics robot. The present invention solves the problem of low accuracy of K-means clustering when feature points are unevenly distributed by introducing multi-dimensional features such as uniformity, local uniformity, and color distance, thereby improving the accuracy of obstacle recognition.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to an obstacle avoidance method and system for an automatic guided logistics robot. Background Art

[0002] Logistics robots can perform real-time positioning through the radar installed on them. Radar obstacle avoidance is an intelligent control technology that uses radar technology to detect and avoid obstacles. Its working principle is to emit high-frequency electromagnetic waves through radar transmitters. When electromagnetic waves encounter obstacles, they will be reflected back and the reflected signals will be received by the radar receiver. The coordinates and color information of the feature points are obtained through the echo energy. The mean clustering algorithm clusters the feature points acquired by the radar, and then the feature points can be used to help the logistics robot avoid obstacles.

[0003] The patent document with announcement number CN110543975B discloses a method for optimizing the evacuation path of a crowd based on a swarm intelligence algorithm and evacuation entropy. The method is executed by a computer, and the specific steps include: step S1: constructing an evacuation environment; step S2: initializing the evacuation crowd and evacuation path parameters; step S3: dividing the crowd into multiple small groups based on the k-means clustering algorithm; step S4: all individuals update their positions based on a hybrid artificial bee colony-bat algorithm; step S5: all individuals correct their positions based on the evacuation entropy; step S6: all individuals avoid obstacles; step S7: if the algorithm does not meet the end condition, return to step S4, otherwise the algorithm ends and the evacuation is completed.

[0004] However, the above patent document does not address the obstacle avoidance problem of logistics robots, and When the mean clustering algorithm clusters the feature points acquired by the radar to identify the feature points of obstacles that affect the movement of the logistics robot, since the feature points acquired by the radar are often sparse and non-uniformly distributed, this data characteristic will cause the identification of obstacle points to be interfered, causing the logistics robot to collide during movement. Summary of the invention

[0005] To solve When mean clustering is performed on the acquired feature points, the feature points are not evenly distributed, which reduces the accuracy of obstacle point recognition and makes it easy for the logistics robot to collide during movement. The present invention provides an obstacle avoidance method and system for an automatically guided logistics robot.

[0006] In a first aspect, the present invention provides an obstacle avoidance method for an automatic guided logistics robot, which adopts the following technical solution: An obstacle avoidance method for an automatic guided logistics robot comprises: obtaining all feature points of the logistics robot in the environment at the current moment, wherein the feature information of the feature points includes horizontal, vertical and vertical coordinates in a spatial coordinate system, and , , Pixel value; obtain the moving direction of the logistics robot, construct a plane coordinate system in the moving direction, and obtain the coordinates of all feature points in the plane coordinate system , Coordinates, calculate the variance of the coordinates of all feature points on each of the same coordinate axes in the space and plane coordinate systems; based on the variance and the size of the circumscribed cube of all feature points on each coordinate axis in the space coordinate system, calculate the uniformity of all feature points; determine the neighborhood feature points among all feature points, and obtain the uniformity of the neighborhood feature points; calculate the local uniformity of the environment based on the uniformity and quantity of all feature points and neighborhood feature points; record the unit vector of the moving direction as the moving vector, and record the unit vector of the neighborhood feature point as the feature vector, based on the local uniformity, the cosine similarity of the feature vectors of any two neighborhood feature points and the moving vector, the Euclidean distance from any two neighborhood feature points to the logistics robot, and use , , The color distance between any two neighborhood feature points obtained by the pixel value is calculated to calculate the feature distance between any two neighborhood feature points; based on the feature distance, use Mean clustering is used to cluster neighborhood feature points and determine obstacle points to achieve obstacle avoidance for the automatic guided logistics robot.

[0007] The beneficial effect is that by introducing multi-dimensional features such as uniformity, local uniformity, and color distance, the The problem of low accuracy of mean clustering when feature points are unevenly distributed is solved, thereby improving the accuracy of obstacle recognition; by calculating the uniformity of feature points and carefully analyzing environmental changes, it is helpful to more reliably cluster obstacles and avoid obstacles; combining color distance and Euclidean distance to calculate feature distance enhances the ability to recognize obstacles and avoids misjudgment, especially in environments with similar colors; through the combination of multiple features, obstacles can be identified in a timely and accurate manner, and obstacles can be avoided quickly, improving the safety and efficiency of logistics robots.

[0008] Furthermore, the uniformity of all the feature points satisfies the following relationship: ; In the formula, is the uniformity of all feature points in the environment where the logistics robot is currently located. are the horizontal, vertical and vertical axes in the spatial coordinate system, is the number of feature points in the spatial coordinate system The variance of the coordinates on the axes, is the circumscribed cube of all feature points in the spatial coordinate system The dimensions on the coordinate axes, They are respectively , Coordinate axes, is the number of all feature points in the plane coordinate system The variance of the coordinates on the axes.

[0009] The beneficial effects are: by comprehensively considering the distribution of feature points in the spatial coordinate system and the plane coordinate system, the uniformity of obstacles in the environment can be accurately evaluated, helping the logistics robot to better analyze the current environment; according to the uniformity of the environment, the logistics robot can judge the dense or sparse areas of obstacles, thereby optimizing the obstacle avoidance strategy and improving the accuracy and safety of decision-making.

[0010] Furthermore, the determination of neighborhood feature points among all feature points includes: clustering all feature points using K-means clustering based on the Euclidean distance from all feature points to the center point of the logistics robot to obtain two clusters, and recording the feature points in the cluster with the shortest average Euclidean distance to the logistics robot as neighborhood feature points.

[0011] The beneficial effect is: through The mean clustering algorithm can accurately identify feature points that are close to the logistics robot, improving the logistics robot's ability to perceive the surrounding environment; focusing on neighborhood feature points helps the robot reduce unnecessary calculations when planning paths, improving the real-time and accuracy of obstacle avoidance decisions; Mean clustering can dynamically adjust according to changes in the environment, helping logistics robots cope with changes in different environments and maintain good obstacle avoidance effects.

[0012] Furthermore, the local uniformity satisfies the following relationship: ; In the formula, is the local uniformity of the logistics robot in the environment at the current moment, is the uniformity of all feature points in the environment where the logistics robot is currently located. is the uniformity of the neighborhood feature points of the logistics robot in the environment at the current moment, is the number of all feature points, is the number of neighborhood feature points, is a natural exponential function.

[0013] The beneficial effects are: by evaluating the uniformity of neighborhood feature points, the logistics robot can more accurately analyze the layout of the current environment and identify areas with dense or sparse obstacles; by calculating local uniformity, it helps the logistics robot to dynamically adjust path planning according to the distribution of environmental feature points, avoid overly crowded or empty areas, optimize obstacle avoidance effects, and improve real-time response capabilities; as the feature points in the environment change, the robot can adjust its behavior strategy in real time, thereby maintaining a stable operating state in a dynamic environment.

[0014] Furthermore, the color distance satisfies the following relationship: ; In the formula, Neighborhood feature points and neighborhood feature points The color distance, and Neighborhood feature points and neighborhood feature points of Pixel value, and Neighborhood feature points and neighborhood feature points of Pixel value, and Neighborhood feature points and neighborhood feature points of Pixel value.

[0015] Furthermore, the characteristic distance satisfies the following relationship: ; In the formula, Neighborhood feature points and neighborhood feature points The characteristic distance, is the local uniformity of the logistics robot in the environment at the current moment, Neighborhood feature points and neighborhood feature points The color distance, Neighborhood feature points The cosine similarity between the feature vector and the motion vector, Neighborhood feature points The cosine similarity between the feature vector and the motion vector, and Neighborhood feature points and neighborhood feature points Euclidean distance to the logistics robot.

[0016] The beneficial effects are: by comprehensively considering the similarity, color distance and Euclidean distance between vectors, the logistics robot can more accurately measure the relationship between neighborhood feature points and improve its understanding of the environment; calculating the feature distance helps the logistics robot identify and analyze the feature differences of different obstacles in the environment, so as to better plan paths and make obstacle avoidance decisions to avoid collisions; as the surrounding feature points change, the robot can update the feature distance in real time to ensure timely and intelligent response in a dynamic environment; by accurately calculating the distance between neighborhood feature points, the robot can identify potential collision risks earlier and improve safety during driving.

[0017] Furthermore, the determination of obstacle points includes: recording the neighborhood feature points in the cluster with the shortest average Euclidean distance to the logistics robot in the two clustering clusters after the neighborhood feature points are clustered as preliminary obstacle points; in response to the ratio of the number of preliminary obstacle points to the number of neighborhood feature points being greater than a preset threshold, determining the preliminary obstacle points as obstacle points.

[0018] Furthermore, the method of achieving obstacle avoidance of the automatic guided logistics robot includes: in response to the logistics robot identifying an obstacle point, adjusting the moving direction of the logistics robot to complete the obstacle avoidance of the automatic guided logistics robot.

[0019] In a second aspect, the present invention provides an obstacle avoidance system for an automatic guided logistics robot, which adopts the following technical solution: An obstacle avoidance system for an automatic guided logistics robot comprises: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned obstacle avoidance method for the automatic guided logistics robot is implemented.

[0020] By adopting the above technical solution, the above-mentioned obstacle avoidance method of the automatic guided logistics robot is generated into a computer program and stored in the memory to be loaded and executed by the processor, so that a terminal device is made according to the memory and the processor for easy use.

[0021] The present invention has the following technical effects: According to the coordinate information of the feature points, the uniformity of all feature points and neighborhood feature points in the environment where the logistics robot is located is obtained, so that obstacle recognition can be applicable to different environments and object placement characteristics; according to the uniformity of all feature points and neighborhood feature points in the environment where the logistics robot is located, the local uniformity of the logistics robot in the current environment is calculated, which narrows the scope of obstacle recognition and weakens the interference of the sparsity and unevenness of the feature point distribution on the feature point clustering; the logistics robot identifies obstacles based on local uniformity and neighborhood feature points, and combines the global information and local information of the feature point distribution to improve the accuracy of obstacle recognition, so that the logistics robot has a higher obstacle avoidance ability during movement. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] By reading the detailed description below with reference to the accompanying drawings, the above and other purposes, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-restrictive manner, and the same or corresponding reference numerals represent the same or corresponding parts.

[0023] Figure 1 This is a flow chart of a method for obstacle avoidance of an automatic guided logistics robot in an embodiment of the present invention.

[0024] Figure 2 It is a first schematic diagram of a plane coordinate system in an obstacle avoidance method for an automatic guided logistics robot according to an embodiment of the present invention.

[0025] Figure 3 It is a second schematic diagram of a plane coordinate system in an obstacle avoidance method for an automatic guided logistics robot according to an embodiment of the present invention.

[0026] Figure 4 It is a schematic diagram of neighborhood feature points in an obstacle avoidance method for an automatic guided logistics robot according to an embodiment of the present invention. DETAILED DESCRIPTION

[0027] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0028] It should be understood that when the terms "first", "second", etc. are used in the claims, descriptions, and drawings of the present invention, they are only used to distinguish different objects, rather than to describe a specific order. The terms "include" and "comprise" used in the description and claims of the present invention indicate the presence of the described features, wholes, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their collections.

[0029] The embodiment of the present invention discloses an obstacle avoidance method based on an automatic guided logistics robot, referring to Figure 1 , comprising steps S1 to S6: S1: Obtain all feature points of the logistics robot in its current environment.

[0030] The feature information of the feature point includes the horizontal, vertical and vertical coordinates in the spatial coordinate system, and , , Pixel value.

[0031] During the movement of the automatically guided logistics robot, the radar installed on the logistics robot is used to obtain the feature points in the environment, and the color information of the feature points is obtained through the echo energy. That is, each feature point has feature information in six dimensions: horizontal coordinate, vertical coordinate, R channel pixel value, G channel pixel value and B channel pixel value.

[0032] S2: Obtain the moving direction of the logistics robot, construct a plane coordinate system in the moving direction, and calculate the uniformity of all feature points.

[0033] Get the moving direction of the logistics robot, construct a plane coordinate system in the moving direction, and project all feature points onto a plane perpendicular to the moving direction of the logistics robot, such as Figure 2 As shown in the figure, the feature points are mapped to the coordinate axis direction of the feature points, but the robot can move in any direction, so it is also possible to map the feature points in any direction, such as Figure 3 As shown in the figure, it is a top view of the projection status, and all feature points are obtained on the plane coordinate system. , Coordinates, calculate the variance of the coordinates of all feature points on each of the same coordinate axes in the space and plane coordinate systems.

[0034] It should be noted that since the feature points acquired by the radar have non-uniform data characteristics, which affect the identification of obstacle points, in order to cope with the impact of this data characteristic on the identification of obstacle points, the present invention obtains the uniformity of all feature points in the environment where the logistics robot is located based on the position information of the feature points.

[0035] Based on the variance and the size of the circumscribed cube of all feature points on each coordinate axis in the spatial coordinate system, the uniformity of all feature points is calculated.

[0036] Specifically, the uniformity of all the feature points satisfies the following relationship: ; In the formula, is the uniformity of all feature points in the environment where the logistics robot is currently located. are the horizontal, vertical and vertical axes in the spatial coordinate system, is the number of feature points in the spatial coordinate system The variance of the coordinates on the axes, is the circumscribed cube of all feature points in the spatial coordinate system The dimensions on the coordinate axes, They are respectively , Coordinate axes, is the number of all feature points in the plane coordinate system The variance of the coordinates on the axes.

[0037] in, is the number of all feature points of the logistics robot in the environment at the current moment. The normalized value of the variance of the coordinates on the coordinate axis. The larger the value, the more likely all feature points are in the The wider the distribution range on the first coordinate axis, the more likely all feature points are evenly distributed in the space, rather than concentrated in certain areas. The greater the uniformity of all feature points, the smaller the value, indicating that all feature points are in the first coordinate axis. The smaller the distribution range on the coordinate axis, the more likely all feature points are concentrated in certain areas of the space, and the smaller the uniformity of all feature points. The larger it is, the greater the uniformity of all feature points. The smaller it is, the less uniform all feature points are; is the number of all feature points of the logistics robot in the environment at the current moment. The normalized value of the variance of the coordinates on the coordinate axis. In the same scene, when the logistics robot moves in different directions, the objects it faces are in different conditions. Therefore, the objects the logistics robot faces during its movement also have a certain significance for the logistics robot's obstacle avoidance. The larger it is, the wider the distribution range of the objects faced by all feature points during the movement of the robot is, which means that the feature points of the objects faced by all feature points during the movement of the logistics robot are more likely to be evenly distributed in space, and the greater the uniformity of all feature points, The smaller it is, the more concentrated the distribution of all feature points in the moving direction of the logistics robot is, which means that during the movement of the logistics robot, the feature points of the objects it faces are more likely to be concentrated in space, and the smaller the uniformity of all feature points is.

[0038] S3: Determine the neighborhood feature points among all feature points, and obtain the uniformity of the neighborhood feature points.

[0039] Specifically, determining the neighborhood feature points among all feature points includes: According to the Euclidean distance from all feature points to the center point of the logistics robot, all feature points are clustered using K-means clustering with a K value of 2 to obtain two clusters. The feature points in the cluster with the shortest average Euclidean distance to the logistics robot (that is, the average value of the Euclidean distance from each feature point in the cluster to the center point of the logistics robot) are recorded as neighborhood feature points, such as Figure 4 shown.

[0040] The method for obtaining the uniformity of the neighborhood feature points is the same as the method for calculating the uniformity of all feature points, which will not be described in detail here.

[0041] S4: Calculate the local uniformity of the environment.

[0042] It should be noted that in the process of obstacle avoidance of the logistics robot, if the feature points are directly clustered to extract obstacle points, the sparsity and non-uniform distribution of the feature points may interfere with the identification of the obstacle points, affecting the obstacle avoidance action of the logistics robot. During the movement of the logistics robot, nearby objects are more of reference significance to the obstacle avoidance action of the logistics robot. In order to make the obstacle identification of the logistics robot more accurate and to perform more accurate obstacle avoidance, the present invention calculates the local uniformity of the environment in which the logistics robot is located based on the uniformity of all feature points in the environment in which the logistics robot is located and the neighborhood feature points in the environment in which the logistics robot is located.

[0043] Based on the uniformity and quantity of all feature points and neighborhood feature points, the local uniformity of the environment is calculated.

[0044] Specifically, the local uniformity satisfies the following relationship: ; In the formula, is the local uniformity of the logistics robot in the environment at the current moment, is the uniformity of all feature points in the environment where the logistics robot is currently located. is the uniformity of the neighborhood feature points of the logistics robot in the environment at the current moment, is the number of all feature points, is the number of neighborhood feature points, is a natural exponential function.

[0045] in, It represents the distribution of feature points near the logistics robot, and nearby feature points, that is, neighborhood feature points, can represent the distribution of feature points in the local range near the logistics robot to a certain extent. The larger it is, the greater the local uniformity of the environment where the logistics robot is located. The smaller it is, the smaller the local uniformity of the environment where the logistics robot is located; It represents the difference between the number of all feature points of the logistics robot excluding the number of neighborhood feature points and the number of neighborhood feature points. The smaller the value, the more likely it is that there are certain objects near the logistics robot, and the objects near the logistics robot are more likely to make the feature points near the logistics robot dense, and the smaller the local uniformity. The larger the value, the more likely it is that there are no certain objects near the logistics robot, and the objects near the robot are likely to make the feature points near the robot sparse, and the greater the local uniformity. It represents the difference between the uniformity of all feature points in the environment where the logistics robot is located and the uniformity of the neighborhood feature points. The smaller the value, the less likely it is that there are locally dense feature points in the environment where the logistics robot is located. The distribution of feature points in the vicinity of the logistics robot and in the environment where the logistics robot is located is more likely to be uniform, and the greater the local uniformity. The larger the value, the more likely it is that the feature points in the vicinity of the logistics robot or outside the vicinity are densely distributed, and the smaller the local uniformity is.

[0046] S5: Calculate the feature distance between any two neighborhood feature points.

[0047] It should be noted that, since the feature points collected during the movement of the logistics robot are sparse and non-uniformly distributed, this data characteristic may interfere with the recognition of the feature points of obstacles. During the obstacle avoidance action of the logistics robot, the feature points that are closer to the logistics robot are more meaningful for obstacle avoidance. Obstacle recognition through the robot's neighborhood feature points can reduce the impact of the inherent characteristics of the feature points acquired by the radar on obstacle recognition. Therefore, the present invention recognizes obstacles based on the local uniformity of the logistics robot's feature points and the logistics robot's neighborhood feature points.

[0048] The unit vector of the moving direction is recorded as the moving vector, and the unit vector of the neighborhood feature point is recorded as the feature vector. Based on the local uniformity, the cosine similarity of the feature vectors and the moving vector of any two neighborhood feature points, the Euclidean distance from any two neighborhood feature points to the center point of the logistics robot, and the , , The color distance between any two neighborhood feature points obtained by pixel values ​​is calculated to calculate the feature distance between any two neighborhood feature points.

[0049] Specifically, the color distance satisfies the following relationship: ; In the formula, Neighborhood feature points and neighborhood feature points The color distance, and Neighborhood feature points and neighborhood feature points of Pixel value, and Neighborhood feature points and neighborhood feature points of Pixel value, and Neighborhood feature points and neighborhood feature points of Pixel value.

[0050] Specifically, the characteristic distance satisfies the following relationship: ; In the formula, Neighborhood feature points and neighborhood feature points The characteristic distance, is the local uniformity of the logistics robot in the environment at the current moment, Neighborhood feature points and neighborhood feature points The color distance, Neighborhood feature points The cosine similarity between the feature vector and the motion vector, Neighborhood feature points The cosine similarity between the feature vector and the motion vector, and Neighborhood feature points and neighborhood feature points Euclidean distance to the logistics robot.

[0051] in, Neighborhood feature points and The color distance between them. Since the objects in the scene where the logistics robot is located have color continuity, that is, the colors of multiple feature points corresponding to an object are the same or similar, and the colors of feature points corresponding to different objects may be different. In order to cluster multiple feature points corresponding to the same object into the same category, The larger is, the larger the characteristic distance between the two neighborhood feature points is. The smaller it is, the smaller the feature distance between the two neighborhood feature points is; is the difference between the Euclidean distances of the two neighborhood feature points and the center point of the logistics robot. Since the Euclidean distances between the feature points of the same object and the logistics robot are similar, in order to better identify obstacles, The larger is, the larger the characteristic distance between the two neighborhood feature points is. The smaller it is, the smaller the feature distance between the two neighborhood feature points is. When there is an object blocking the logistics robot, the object will make the feature points near the logistics robot appear locally dense. When there is no object blocking the logistics robot, the feature points near the logistics robot will appear more evenly distributed. At this time, the data volume of the two types of neighborhood feature points clustered may be similar, which will cause the misidentification of obstacles. Therefore, when the logistics robot is moving, The larger the value, the smaller the The influence on the feature distance, The smaller the time, the better. The influence on the feature distance, so use Divide by , The bigger it is, The smaller the The smaller the impact on feature distance, The more hours, The larger the relative The closer , to maintain The impact on feature distance.

[0052] S6: Based on the characteristic distance, use Mean clustering is used to cluster neighborhood feature points and determine obstacle points to achieve obstacle avoidance for the automatic guided logistics robot.

[0053] Specifically, determining the obstacle point includes: Mean Clustering The value is 2. Among the two clusters after the clustering of the neighborhood feature points is completed, the neighborhood feature points in the cluster with the shortest average Euclidean distance to the logistics robot are recorded as preliminary obstacle points. In response to the ratio of the number of the preliminary obstacle point to the number of the neighborhood feature points being greater than a preset threshold, the preliminary obstacle point is identified as an obstacle point.

[0054] Implementers can set the threshold according to specific implementation circumstances, for example, 0.5.

[0055] Specifically, the method for realizing obstacle avoidance of the automatic guided logistics robot includes: In response to the logistics robot identifying an obstacle point, the moving direction of the logistics robot is adjusted to complete the obstacle avoidance of the automatic guided logistics robot.

[0056] An embodiment of the present invention further discloses an obstacle avoidance system for an automatic guided logistics robot, including a processor and a memory, wherein the memory stores computer program instructions. When the computer program instructions are executed by the processor, an obstacle avoidance method for an automatic guided logistics robot according to the present invention is implemented.

[0057] The above system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface, and their configuration and functions are known in the art, so they will not be described in detail here.

[0058] In the present invention, the aforementioned memory may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium may be any suitable magnetic storage medium or magneto-optical storage medium, or any other medium that can be used to store the required information and can be accessed by an application, module, or both. Any such computer storage medium may be part of the device or accessible or connectable to the device.

[0059] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

[0060] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An obstacle avoidance method for an automatic guided logistics robot, characterized in that: include: Obtain all feature points of the logistics robot in the environment at the current moment, the feature information of the feature points includes the horizontal, vertical and vertical coordinates in the spatial coordinate system, and , , Pixel value; Get the moving direction of the logistics robot, build a plane coordinate system in the moving direction, and get all the feature points in the plane coordinate system. , Coordinates, calculate the variance of the coordinates of all feature points on each of the same coordinate axes in the space and plane coordinate systems; based on the variance and the size of the circumscribed cube of all feature points on each coordinate axis in the space coordinate system, calculate the uniformity of all feature points; determine the neighborhood feature points among all feature points, and obtain the uniformity of the neighborhood feature points; calculate the local uniformity of the environment based on the uniformity and quantity of all feature points and neighborhood feature points; record the unit vector of the moving direction as the moving vector, and record the unit vector of the neighborhood feature point as the feature vector, based on the local uniformity, the cosine similarity of the feature vectors of any two neighborhood feature points and the moving vector, the Euclidean distance from any two neighborhood feature points to the logistics robot, and use , , The color distance between any two neighborhood feature points obtained by the pixel value is calculated to calculate the feature distance between any two neighborhood feature points; based on the feature distance, use Mean clustering is used to cluster neighborhood feature points and determine obstacle points to achieve obstacle avoidance for the automatic guided logistics robot.

2. The obstacle avoidance method of an automatic guided logistics robot according to claim 1, characterized in that: The uniformity of all the feature points satisfies the following relationship: ； In the formula, is the uniformity of all feature points in the environment where the logistics robot is currently located. are the horizontal, vertical and vertical axes in the spatial coordinate system, is the number of all feature points in the spatial coordinate system The variance of the coordinates on the axes, is the circumscribed cube of all feature points in the spatial coordinate system The dimensions on the coordinate axes, They are respectively , Coordinate axes, is the number of all feature points in the plane coordinate system The variance of the coordinates on the axes.

3. The obstacle avoidance method of an automatic guided logistics robot according to claim 1, characterized in that: The step of determining the neighborhood feature points among all the feature points includes: According to the Euclidean distance from all feature points to the center point of the logistics robot, K-means clustering is used to cluster all feature points to obtain two clusters. The feature points in the cluster with the shortest average Euclidean distance to the logistics robot are recorded as neighborhood feature points.

4. The obstacle avoidance method of an automatic guided logistics robot according to claim 1, characterized in that: The local uniformity satisfies the following relationship: ； In the formula, is the local uniformity of the logistics robot in the environment at the current moment, is the uniformity of all feature points in the environment where the logistics robot is currently located. is the uniformity of the neighborhood feature points of the logistics robot in the environment at the current moment, is the number of all feature points, is the number of neighborhood feature points, is a natural exponential function.

5. The obstacle avoidance method of an automatic guided logistics robot according to claim 1, characterized in that: The color distance satisfies the following relationship: ； In the formula, Neighborhood feature points and neighborhood feature points The color distance, and Neighborhood feature points and neighborhood feature points of Pixel value, and Neighborhood feature points and neighborhood feature points of Pixel value, and Neighborhood feature points and neighborhood feature points of Pixel value.

6. The obstacle avoidance method of an automatic guided logistics robot according to claim 1, characterized in that: The characteristic distance satisfies the following relationship: ； In the formula, Neighborhood feature points and neighborhood feature points The characteristic distance of is the local uniformity of the logistics robot in the environment at the current moment, Neighborhood feature points and neighborhood feature points The color distance, Neighborhood feature points The cosine similarity between the feature vector and the motion vector, Neighborhood feature points The cosine similarity between the feature vector and the motion vector, and Neighborhood feature points and neighborhood feature points Euclidean distance to the logistics robot.

7. The obstacle avoidance method of an automatic guided logistics robot according to claim 1, characterized in that: The step of determining the obstacle point comprises: Among the two clusters after the clustering of the neighborhood feature points is completed, the neighborhood feature points in the cluster with the shortest average Euclidean distance to the logistics robot are recorded as preliminary obstacle points; In response to the ratio of the number of the preliminary obstacle point to the number of the neighborhood feature points being greater than a preset threshold, the preliminary obstacle point is identified as an obstacle point.

8. The obstacle avoidance method of an automatic guided logistics robot according to claim 1, characterized in that: The method for realizing obstacle avoidance of the automatic guided logistics robot includes: In response to the logistics robot identifying an obstacle point, the moving direction of the logistics robot is adjusted to complete the obstacle avoidance of the automatic guided logistics robot.

9. An obstacle avoidance system for an automatic guided logistics robot, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, an obstacle avoidance method for an automatic guided logistics robot according to any one of claims 1-8 is implemented.

Images