CN111582352B

CN111582352B - Object-based perception method, object-based perception device, robot and storage medium

Info

Publication number: CN111582352B
Application number: CN202010363102.5A
Authority: CN
Inventors: 张伟义; 沈孝通; 秦宝星; 程昊天
Original assignee: Shanghai Gaussian Automation Technology Development Co Ltd
Current assignee: Shanghai Gaussian Automation Technology Development Co Ltd
Priority date: 2020-04-30
Filing date: 2020-04-30
Publication date: 2023-06-27
Anticipated expiration: 2040-04-30
Also published as: CN111582352A

Abstract

The invention discloses an object-based perception method, a device, a robot and a storage medium, wherein the method comprises the following steps: establishing a graph structure according to the collected at least one point cloud data; clustering the point cloud data according to the weight value of the edge in the graph structure, and obtaining at least one class of mark semantic tags after clustering; determining the number of objects with the distance from the robot smaller than a preset first distance threshold according to the clustered class of at least one mark semantic label; controlling the movement of the robot according to the number of objects whose distance to the robot is smaller than a preset first distance threshold. According to the method, when the object perception is carried out, the collected point cloud data are clustered according to the graph structure established by the point cloud data, so that semantic segmentation is realized, the movement of the robot is controlled, the clustering process is simple and efficient, the object perception efficiency is improved, and the safety of the robot and the surrounding environment in the moving process is ensured.

Description

Object-based perception method, object-based perception device, robot and storage medium

Technical Field

The embodiment of the invention relates to the technical field of robot perception, in particular to an object-based perception method, an object-based perception device, a robot and a storage medium.

Background

With the continuous development of computer technology, the development of robot sensor devices is perfect and the popularization of robot operating systems, and the trend of robot replacement is coming. The sensing technology in the movement of the robot is to analyze and calculate a series of sensor data to obtain the understanding of the robot to the surrounding environment. Currently, the sensing technology is successfully applied to various fields of unmanned driving, port wharfs, home services and the like. The robust and accurate perception capability is one of the core technologies of the robot and is a basic requirement for realizing autonomous navigation of the robot. When the robot is executing the task, the robot should respond to the surrounding environment in time according to the information of the sensor, so as to meet the requirement of safety.

At present, when the robot senses in the moving process, the robot can cluster according to the characteristic attribute of the collected point cloud data, and whether an object affecting the operation of the robot exists is determined according to a clustering result. The characteristic attributes of the point cloud data may include texture, normal vector, color information, euclidean distance, and the like. The specific process can be as follows: and determining characteristic attributes of the point cloud data, clustering according to the attributes, and dividing the point cloud data with different attributes.

However, in the above process, the sensing efficiency is low and the safety in the moving process of the robot is low because the characteristic attribute of the point cloud data needs to be determined first.

Disclosure of Invention

The invention provides an object-based sensing method, an object-based sensing device, a robot and a storage medium, and aims to solve the technical problem that the object sensing efficiency is low due to the existing sensing method.

In a first aspect, an embodiment of the present invention provides an object-based sensing method, including:

establishing a graph structure according to the collected at least one point cloud data;

clustering the point cloud data according to the weight value of the edge in the graph structure to obtain at least one class of mark semantic tags after clustering;

determining the number of objects with the distance from the robot smaller than a preset first distance threshold according to the clustered class of at least one mark semantic tag;

controlling the robot according to the number of objects with the distance to the robot smaller than a preset first distance threshold.

In the method as described above, after controlling the robot according to the number of objects whose distance from the robot is smaller than a preset first distance threshold, the method further includes:

determining the ratio of the number of the point cloud data with the distance from the robot smaller than a preset second distance threshold value to the total number in one frame of the point cloud data according to the point cloud data;

And controlling the movement of the robot according to the ratio.

In the implementation mode, the two judgment conditions are combined with the judgment mode, so that the misjudgment probability can be reduced, the accuracy of object perception is improved, and further, the motion of the robot can be accurately controlled.

In the method shown above, the vertex set in the graph structure is a set of coordinate values of the point cloud data, the edge set in the graph structure is a set of edges connecting adjacent point cloud data, and weights of the edges of the adjacent point cloud data are determined according to the distance between the adjacent point cloud data;

clustering the point cloud data according to the weight value of the edge in the graph structure, including:

sequencing the edges according to the sequence from small weight to large weight, and obtaining a sequenced edge set;

if the dissimilarity between the two sides which do not contain the common vertexes in the ordered edge set is smaller than a preset dissimilarity threshold, gathering the point cloud data connected with the two sides into one class, and obtaining at least one class after primary clustering;

and re-clustering each class after the at least one primary clustering according to the furthest distance between the point cloud data included in each class after the at least one primary clustering to form a class marking the semantic label.

In the implementation mode, based on a graph structure, each class after at least one primary clustering is acquired first, then each class after the primary clustering is clustered again, and finally the acquired class is the class of the mark semantic label. On one hand, the clustering process is simple and efficient, and the object perception efficiency is improved; on the other hand, according to the furthest distance between the point cloud data included in each class after primary clustering, clustering each class after at least one primary clustering again, and the obtained class marked with the semantic label can facilitate object perception of a robot, so that the accuracy of object perception is improved, the motion of the robot is better controlled, and the accuracy of object perception is improved.

In the method shown above, the clustering the at least one first clustered class again according to the farthest distance between the point cloud data included in the at least one first clustered class to form a class of marking semantic labels includes:

determining the furthest distance between the point cloud data included in each class after the primary clustering;

determining the primary clustered class with the farthest distance between preset distance ranges as an effective class;

And clustering the effective classes again based on a semantic segmentation mode to form classes marking semantic tags.

In the implementation mode, effective classes are screened out and clustered to form classes of the mark semantic tags, so that the accuracy of the clustered classes of the mark semantic tags can be improved, and further, the accuracy of robot perception can be improved.

In the method shown above, the clustering the valid classes again based on the semantic segmentation mode to form classes marking semantic tags includes:

taking the minimum number of the point cloud data included in the effective class as the number of the effective class; the serial numbers of the point cloud data are serial numbers formed by sequentially numbering at least one acquired point cloud data from left to right;

sorting the effective classes according to the sequence from the small number to the large number of the effective classes, and obtaining a sorted effective class set;

determining the distance between adjacent effective classes in the ordered effective class set;

if the distance between the adjacent effective classes is smaller than a preset third distance threshold value, the adjacent effective classes are gathered into a class marking the semantic tags;

and if the distance between the adjacent effective classes is greater than or equal to the preset third distance threshold, using the adjacent effective classes as classes of two marking semantic tags.

In the implementation mode, the classes after primary clustering are clustered again to form the classes of the mark semantic tags, on one hand, the clustering process is simple and efficient, the object perception efficiency is further improved, and on the other hand, the acquired classes of the mark semantic tags can facilitate the object perception of the robot, so that the object perception accuracy is improved.

In the method shown above, the object is a person, the class of the labeled semantic tag is a class of a person, the effective class is a class of a person leg, the preset distance range is a preset width range of the person leg, and the preset third distance threshold is a preset width between two legs.

The realization mode can realize the perception of the robot to the crowd's bustle, and when detecting to be bustle by crowd, in time stop the motion to personal safety and self safety around the protection.

In the method as shown above, the determining, according to the clustered class of the at least one labeled semantic label, the number of objects whose distance from the robot is smaller than a preset first distance threshold includes:

and determining the number of classes of the marking semantic tags, the distance from the robot of which is smaller than a preset first distance threshold, as the number of objects, the distance from the robot of which is smaller than the preset first distance threshold.

The implementation mode can determine the surrounding environment of the robot according to the number of the classes of the marking semantic tags, wherein the distance between the classes and the robot is smaller than the preset first distance threshold, the implementation process is simple and efficient, and the perception efficiency of the robot is improved.

In the method as shown above, the controlling the movement of the robot according to the number of objects whose distance from the robot is smaller than a preset first distance threshold includes:

and if the number of the objects with the distance from the robot smaller than the preset first distance threshold is larger than or equal to the preset object number threshold, controlling the robot to stop moving.

In the implementation manner, when the number of objects with the distance from the robot smaller than the preset first distance threshold is larger than or equal to the preset object number threshold, the robot is controlled to stop moving, so that the safety of the robot and the safety of the surrounding environment are protected.

In the method as described above, if the number of objects whose distance from the robot is smaller than the preset first distance threshold is greater than or equal to the preset object number threshold, controlling the robot to stop moving includes:

if the number of objects with the distance from the robot smaller than a preset first distance threshold is larger than or equal to a preset object number threshold, setting an object flag bit of a frame corresponding to the at least one point cloud data as an effective value;

And if the number of frames with the object zone bits continuously being the effective value is larger than a preset frame threshold value, controlling the robot to stop moving.

According to the realization mode, the motion of the robot can be controlled by combining the sensing results of multiple frames, and the accuracy of object sensing is improved, so that the motion of the robot can be accurately controlled.

In the method as shown above, before the building of the graph structure according to the at least one point cloud data collected by the robot, the method further includes:

acquiring at least one piece of acquired original point cloud data;

and deleting the original point cloud data, of which the distance between the original point cloud data and the robot is not in a preset robot detection distance range, from the at least one original point cloud data to form the filtered at least one point cloud data.

By the implementation mode, invalid point cloud data can be deleted, so that accuracy of object perception is improved.

In a second aspect, an embodiment of the present invention provides an object-based sensing apparatus, including:

the building module is used for building a graph structure according to the collected at least one point cloud data;

the clustering module is used for clustering the point cloud data according to the weight value of the edge in the graph structure, and obtaining at least one class of mark semantic tags after clustering;

The determining module is used for determining the number of objects with the distance from the robot smaller than a preset first distance threshold according to the class of the clustered at least one mark semantic tag;

and the control module is used for controlling the movement of the robot according to the number of objects with the distance to the robot being smaller than a preset first distance threshold value.

In the apparatus as shown above, the apparatus further comprises:

the determining control module is used for determining the ratio of the number of the point cloud data with the distance from the robot smaller than a preset second distance threshold value to the total number of one frame of point cloud data according to the point cloud data; and controlling the movement of the robot according to the ratio.

In the above-mentioned device, the vertex set in the graph structure is a set of coordinate values of the point cloud data, the edge set in the graph structure is a set of edges connecting adjacent point cloud data, weights of the edges of the adjacent point cloud data are determined according to the distance between the adjacent point cloud data, and the clustering module specifically includes:

the acquisition sub-module is used for sequencing the edges according to the sequence from small weight to large weight, and acquiring a sequenced edge set;

the first clustering sub-module is used for gathering the point cloud data connected with the two sides into one class if the dissimilarity between the two sides which do not contain the common vertexes in the ordered edge set is smaller than a preset dissimilarity threshold value, and acquiring each class after at least one primary clustering;

And the second clustering sub-module is used for clustering each class after the at least one primary clustering again according to the furthest distance between the point cloud data included in each class after the at least one primary clustering to form a class marking the semantic label.

In the apparatus as shown above, the second aggregation sub-module is specifically configured to:

In the device shown above, in terms of clustering the valid classes again based on the semantic segmentation mode to form classes marking the semantic tags, the second clustering sub-module is specifically configured to:

In the device shown above, the class of the labeled semantic tag is a human class, the effective class is a human leg class, the preset distance range is a preset width range of human legs, and the preset third distance threshold is a preset width between two legs.

In the device shown above, the determining module is specifically configured to determine the number of classes of the labeled semantic tags whose distance from the robot is smaller than a preset first distance threshold as the number of objects whose distance from the robot is smaller than the preset first distance threshold.

In the apparatus as shown above, the control module includes:

and the control sub-module is used for controlling the robot to stop moving if the number of objects with the distance smaller than the preset first distance threshold value is larger than or equal to the preset object number threshold value.

In the apparatus as shown above, the control submodule is specifically configured to:

In the apparatus as above, before the building of the graph structure according to the at least one point cloud data collected by the robot, the apparatus further includes:

the acquisition module is used for acquiring the acquired at least one original point cloud data;

and the deleting module is used for deleting the original point cloud data, of which the distance between the original point cloud data and the robot is not in a preset robot detection distance range, to form the filtered at least one point cloud data.

In a third aspect, an embodiment of the present invention further provides a robot, including:

one or more processors;

a memory for storing one or more programs;

the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the object-based perception method as provided in the first aspect.

In a fourth aspect, embodiments of the present invention also provide a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the object-based perception method as provided in the first aspect.

The embodiment of the invention provides an object-based sensing method, an object-based sensing device, a robot and a storage medium, wherein the method comprises the following steps: establishing a graph structure according to the collected at least one point cloud data; clustering the point cloud data according to the weight value of the edge in the graph structure, and obtaining at least one class of mark semantic tags after clustering; determining the number of objects with the distance from the robot smaller than a preset first distance threshold according to the clustered class of at least one mark semantic label; controlling the movement of the robot according to the number of objects whose distance to the robot is smaller than a preset first distance threshold. When the method is used for object perception, the acquired point cloud data can be clustered according to the graph structure established by the point cloud data, semantic segmentation is realized, the number of objects with the distance smaller than a preset first distance threshold value from the robot is determined, so that the movement of the robot is controlled, on one hand, the clustering process is simple and efficient, the object perception efficiency is improved, on the other hand, the acquired classes marked with the semantic tags can facilitate the object perception of the robot, the accuracy of the object perception is improved, the movement of the robot is better controlled, and the accuracy of the object perception is improved. Therefore, the object-based perception method provided by the embodiment ensures the safety of the robot in the moving process, and simultaneously ensures the safety of the surrounding environment in the moving process of the robot.

Drawings

Fig. 1 is a schematic diagram of an application scenario of an object-based perception method according to an embodiment of the present invention;

FIG. 2 is a flow chart of an object-based sensing method according to an embodiment of the present invention;

fig. 3 is a schematic flow chart of primary clustering of point cloud data in an object-based sensing method according to an embodiment of the present invention;

fig. 4 is a schematic diagram of a process of primary clustering of point cloud data in an object-based perception method according to an embodiment of the present invention;

FIG. 5 is a flowchart illustrating a method for determining classes for marking semantic tags in an object-based perception method according to an embodiment of the present invention;

FIG. 6 is a schematic diagram illustrating a structure of an object-based sensing device according to an embodiment of the present invention;

fig. 7 is a schematic structural diagram of a robot according to an embodiment of the present invention;

fig. 8 is a schematic structural diagram of a computer readable storage medium according to an embodiment of the present invention.

Detailed Description

The invention is described in further detail below with reference to the drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting thereof. It should be further noted that, for convenience of description, only some, but not all of the structures related to the present invention are shown in the drawings.

Fig. 1 is a schematic diagram of an application scenario of an object-based perception method according to an embodiment of the present invention. As shown in fig. 1, during movement of the robot 11, for example, when the robot 11 performs autonomous navigation in a known map, it is necessary to sense an object in the surrounding environment that affects the movement of the robot, so as to control the movement of the robot according to the sensing result. The object in this embodiment may be a stationary object in the environment, such as a table 12, or a moving object in the environment, such as a person 13, or other living being, such as a dog, etc. According to the object-based perception method, the point cloud data acquired by the robot are clustered, semantic segmentation is achieved, the number of objects with the distance smaller than the preset first distance threshold value to the robot is determined, so that the movement of the robot is controlled, the clustering process is simple and efficient, object perception efficiency is improved, safety in the moving process of the robot is guaranteed, and meanwhile safety of surrounding environments in the moving process of the robot is guaranteed.

Fig. 2 is a flow chart of an object-based sensing method according to an embodiment of the invention. The method and the device are suitable for a scene of sensing objects in the surrounding environment in the moving process of the robot. The object-based perception method may be performed by an object-based perception device, which may be implemented in software and/or hardware, which may be integrated in a robot. As shown in fig. 2, the object-based sensing method provided in this embodiment includes the following steps:

Step 201: and establishing a graph structure according to the acquired at least one point cloud data.

Specifically, in the present embodiment, the robot may acquire point cloud data by a laser sensor provided thereon. The point cloud data in this embodiment may be two-dimensional or three-dimensional point cloud data.

Optionally, the vertex set in the graph structure is a set of coordinate values of the point cloud data, the edge set in the graph structure is a set of edges connecting adjacent point cloud data, and weights of the edges of the adjacent point cloud data are determined according to distances between the adjacent point cloud data.

The coordinate values of the point cloud data may be coordinate values of the collected points in a robot coordinate system. In the scene of the two-dimensional point cloud data, the robot coordinate system is a coordinate system constructed with a point on the robot as an origin, the forward direction of the robot as the X-axis direction, and the left direction of the robot as the Y-axis direction. In the three-dimensional point cloud data scene, the robot coordinate system is a coordinate system constructed with a point on the robot as an origin, the forward direction of the robot as an X-axis direction, the left direction of the robot as a Y-axis direction, and a direction perpendicular to the X-Y plane as a Z-axis direction. The point on the robot may be any point on the robot, for example, the middle point of the drive wheel connection.

In one implementation manner, in order to improve accuracy of object perception, before step 201, acquired at least one piece of original point cloud data may be acquired, and original point cloud data, of which a distance from the robot is not in a preset detection distance range of the robot, in the at least one piece of original point cloud data may be deleted, so as to form filtered at least one piece of point cloud data involved in step 201. Namely, the original point cloud data exceeding the detection distance range of the robot is deleted, and the deleted element point cloud data are invalid point cloud data.

The range of the detection distance of the robot referred to herein may be (range_min, range_max), wherein range_min is the closest detection distance of the robot and range_max is the farthest detection distance of the robot. The range_min and range_max may be set according to parameters of a laser sensor provided on the robot. Illustratively, range_min may be a value between 3 cm and 5 cm.

The original point cloud data, the distance from the robot of which is not in the preset robot detection distance range, refers to point cloud data, the distance from the robot of which is smaller than range_min or larger than range_max.

In this embodiment, the distance between the point cloud data and the robot refers to the euclidean distance between the point cloud data and the robot. Assuming that the coordinate value of the point cloud data 1 is (x) ₁ ，y ₁ ) The distance between the point cloud data 1 and the robot is

The distance between adjacent point cloud data in the graph structure refers to the euclidean distance between the adjacent point cloud data. Assuming that adjacent point cloud data are point cloud data 2 and point cloud data 3, respectively, the coordinate value of the point cloud data 2 is (x ₂ ，y ₂ ) The coordinate value of the point cloud data 3 is (x ₃ ，y ₃ ) The distance between the point cloud data 2 and the point cloud data 3 is

When the point cloud data are collected, a corresponding sequence is arranged among the plurality of point cloud data based on the scanning sequence of the laser sensor. In this embodiment, adjacent point cloud data in at least one point cloud data are connected, and these connecting lines are edges in the graph structure. Assuming that the number of the point cloud data collected by the robot is n, in the graph structure, the number of the vertexes is n, and the number of the edges is n-1. Each edge has a weight that is the distance between the two point cloud data that this edge connects. The meaning of the weight is the dissimilarity between the two point cloud data connected by the edge.

Step 202: and clustering the point cloud data according to the weight value of the edge in the graph structure, and obtaining at least one class of the mark semantic label after clustering.

Specifically, after determining the graph structure, clustering the point cloud data according to the weight value of the edge in the graph structure, realizing semantic segmentation, and obtaining at least one class marking the semantic label after clustering. Illustratively, the class of tagged semantic tags may be a class of people.

In one implementation, point cloud data of edge connections with weight values within the same threshold range may be grouped into one class based on the weight values of the edges.

In another implementation, the point cloud data may be clustered based on a graph partitioning (Graphseg) algorithm, based on weight values of edges. An important feature of the Graphseg algorithm is that it is able to preserve regions in low-variability image regions while ignoring regions in high-variability regions.

In another implementation manner, the clustering can be performed twice based on the Graphseg algorithm, and the specific process can be as follows: ordering edges in the graph structure according to the order of the weights from small to large, and obtaining an ordered edge set; if the dissimilarity between the two sides which do not contain the common vertexes in the ordered edge set is smaller than a preset dissimilarity threshold value, gathering the point cloud data connected with the two sides into one class, and obtaining each class after at least one primary clustering; and re-clustering the at least one primary clustered class according to the furthest distance between the point cloud data included in the at least one primary clustered class to form a class marking the semantic label.

More specifically, in performing the re-clustering, it may be realized based on the following manner: determining the furthest distance between the point cloud data included in each class after primary clustering; determining the class which is the farthest distance between the preset distance ranges and is subjected to primary clustering as an effective class; and clustering the effective classes again based on the semantic segmentation mode to form classes marked with semantic tags. Effective classes are screened out and clustered to form classes of the mark semantic tags, so that the accuracy of the clustered classes of the mark semantic tags can be improved, and further, the accuracy of robot perception can be improved.

The distances between each point cloud data and other point cloud data in the class after primary clustering can be calculated in an enumeration mode, and the farthest distance in the distances is selected. In one particular scenario, when the object is a person, the active class may be a human leg class. The preset distance range may be a preset width range of the human leg. The preset distance range may be leg_width±thres, where leg_width may be 0.2 meter and thres may be 0.05 meter.

In the implementation mode, based on a graph structure, at least one class after primary clustering is acquired first, then each class after primary clustering is clustered again, and finally the acquired class is the class of the mark semantic label. On one hand, the clustering process is simple and efficient, and the object perception efficiency is improved; on the other hand, according to the furthest distance between the point cloud data included in each class after primary clustering, clustering each class after at least one primary clustering again, and the obtained class marked with the semantic label can facilitate object perception of a robot, so that the accuracy of object perception is improved, the motion of the robot is better controlled, and the accuracy of object perception is improved. The process of forming the class of the tag semantic tags based on the clustering method of the Graphseg algorithm will be described in detail later.

At least one point cloud data can be included in each class after primary clustering, and the number of the point cloud data included in each class can be the same or different. The respective classes of the tag semantic tags may include at least one point cloud data, and the number of the point cloud data included in each class may be the same or different.

Step 203: and determining the number of objects with the distance from the robot smaller than a preset first distance threshold according to the clustered class of at least one mark semantic label.

Specifically, after clustering, the number of objects whose distance to the robot is less than a preset first distance threshold may be determined based on the class of the labeled semantic tags. More specifically, the number of classes of labeled semantic tags having a distance to the robot that is less than a preset first distance threshold may be determined as the number of objects having a distance to the robot that is less than the preset first distance threshold.

The distance between the class of the tag semantic tag and the robot may be the distance between any one of the point cloud data in the class of the tag semantic tag and the robot, or the maximum value or the minimum value of the distances between all the point cloud data in the class of the tag semantic tag and the robot, which is not limited in this embodiment.

Step 204: controlling the movement of the robot according to the number of objects whose distance to the robot is smaller than a preset first distance threshold.

In one implementation, if the number of objects whose distance from the robot is smaller than the preset first distance threshold is greater than or equal to the preset object number threshold, the robot is controlled to stop moving.

In another implementation, in order to improve accuracy of object sensing and avoid false sensing, an object flag bit may be set in the robot. Each frame corresponds to an object flag bit. If the number of objects with the distance from the robot smaller than the preset first distance threshold is larger than or equal to the preset object number threshold, setting the object zone bit of the frame corresponding to at least one point cloud data as an effective value; and if the number of frames with the object zone bits continuously being the effective value is larger than the preset frame threshold value, controlling the robot to stop moving. For example, the preset frame threshold may be 10, in other words, if the object flag bits of the consecutive 10 frames are all valid values, the robot is controlled to stop moving.

In still another implementation manner, after step 204, the object-based sensing method provided in this embodiment may further determine, according to the point cloud data, a ratio of the number of point cloud data having a distance from the robot less than the preset second distance threshold to the total number of point cloud data in one frame, and control the movement of the robot according to the ratio. Note that, the total number of one frame of point cloud data in the present embodiment refers to the total number of point cloud data included in one frame of point cloud data.

In this implementation, more specifically, the robot is controlled to stop moving when the ratio of the number of point cloud data having a distance from the robot smaller than a preset second distance threshold to the total number of one frame of point cloud data is greater than a preset ratio threshold.

Alternatively, the second distance threshold may be less than the first distance threshold.

In this implementation, when the ratio of the number of point cloud data with a distance from the robot smaller than the preset second distance threshold to the total number of one frame of point cloud data is greater than the preset ratio threshold, it is indicated that the laser sensor is blocked to a greater extent. If the robot is controlled only depending on the condition that the laser sensor is blocked or only depending on the number of objects with the distance from the robot smaller than a preset first distance threshold value, the movement of the robot is controlled, and the probability of false judgment is high. For example, in the case that the laser sensor is only blocked, the degree of blocking of the laser sensor is light, that is, the ratio of the number of point cloud data with the distance from the robot smaller than the preset second distance threshold to the total number of one frame of point cloud data is smaller than or equal to the preset ratio threshold, but when the number of objects with the distance from the robot smaller than the preset first distance threshold is greater than or equal to the preset object number threshold, erroneous judgment occurs, so that the robot continues to move under the condition that the objects exist, and the safety of the surrounding environment and the robot is reduced. In another example, in the case of controlling the movement of the robot only depending on the number of objects having a distance from the robot smaller than the preset first distance threshold, if the number of objects having a distance from the robot smaller than the preset first distance threshold is smaller than the preset object number threshold, but the ratio of the number of point cloud data having a distance from the robot smaller than the preset second distance threshold to the total number of one frame of point cloud data is larger than the preset ratio threshold, erroneous judgment occurs, resulting in continued movement of the robot under the condition that the laser sensor is seriously blocked, and the safety of the surrounding environment and itself is reduced.

Therefore, the combination of the two judging conditions can reduce the false judgment probability, improve the accuracy of object perception, and further realize accurate control of the movement of the robot.

In a more specific scenario, the object in this embodiment may be a person. The embodiment can realize the perception of the robot to the crowd around, and when the robot detects the crowd around, the robot stops moving in time, thereby protecting the personal safety around and the self safety.

The object-based perception method provided in the embodiment includes: establishing a graph structure according to the collected at least one point cloud data; clustering the point cloud data according to the weight value of the edge in the graph structure, and obtaining at least one class of mark semantic tags after clustering; determining the number of objects with the distance from the robot smaller than a preset first distance threshold according to the clustered class of at least one mark semantic label; controlling the movement of the robot according to the number of objects whose distance to the robot is smaller than a preset first distance threshold. When the method is used for object perception, the acquired point cloud data can be clustered according to the graph structure established by the point cloud data, semantic segmentation is realized, the number of objects with the distance smaller than a preset first distance threshold value from the robot is determined, so that the movement of the robot is controlled, on one hand, the clustering process is simple and efficient, the object perception efficiency is improved, on the other hand, the acquired classes marked with the semantic tags can facilitate the object perception of the robot, the accuracy of the object perception is improved, the movement of the robot is better controlled, and the accuracy of the object perception is improved. Therefore, the object-based perception method provided by the embodiment ensures the safety of the robot in the moving process, and simultaneously ensures the safety of the surrounding environment in the moving process of the robot.

The present embodiment also provides another object-based perception method. The embodiment describes in detail a process of primary clustering of point cloud data based on the embodiment shown in fig. 1 and various alternatives. Fig. 3 is a schematic flow chart of primary clustering of point cloud data in an object-based perception method according to an embodiment of the present invention. As shown in fig. 3, if the dissimilarity between two sides not including a common vertex in the ordered edge set is smaller than a preset dissimilarity threshold, the point cloud data connected to the two sides are gathered into one class, and each class after at least one primary clustering is acquired, including the following steps:

step 301: and acquiring the first edge of the first bit, which is not clustered, of the connected point cloud data in the ordered edge set.

Step 302: and acquiring a second edge which is arranged at a second position and is different from the point cloud data connected with the first edge, wherein the point cloud data connected with the second edge are not clustered in the ordered edge set.

Step 303: the dissimilarity of the first edge with the second edge is determined.

Step 304: if the dissimilarity between the first edge and the second edge is smaller than a preset dissimilarity threshold, the point cloud data connected with the second edge and the point cloud data connected with the first edge are gathered into one type.

Step 305: and returning to the step of determining the dissimilarity between the first edge and the second edge by taking the edge which is arranged behind the second edge and is not clustered by the connected point cloud data and is different from the point cloud data connected to the first edge in the ordered edge set as a new second edge until the edge which is not clustered by the connected point cloud data and is arranged behind the second edge and is different from the point cloud data connected to the first edge does not exist in the ordered edge set, and stopping returning to the step.

Step 306: and returning the edges which are arranged behind the first edge and are not clustered by the connected point cloud data in the ordered edge set to execute the steps of acquiring the second edges which are arranged at the second position and are different from the connected point cloud data of the first edge in the ordered edge set until the first edges which are not clustered by the connected point cloud data are not arranged at the first position in the ordered edge set, or the number of the edges which are not clustered by the connected point cloud data in the ordered edge set is 1, or the second edges which are not clustered by the connected point cloud data and are connected with the first edge and are arranged at the second position in the ordered edge set are not exist in the ordered edge set, stopping executing, and acquiring all the classes after primary clustering.

Fig. 4 is a schematic diagram of a process of primary clustering of point cloud data in an object-based perception method according to an embodiment of the present invention. The specific implementation of steps 301-306 is described in detail below in conjunction with fig. 4. As shown in fig. 4, it is assumed that 7 point cloud data are acquired, denoted as P1-P7 from left to right, respectively. The edge between P1 and P2 is denoted as e12, the edge between P2 and P3 is denoted as e23, the edge between P3 and P4 is denoted as e34, the edge between P4 and P5 is denoted as e45, the edge between P5 and P6 is denoted as e56, and the edge between P6 and P7 is denoted as e67. That is, in the graph structure created from the point cloud data, the vertex sets are (P1, P2, P3, P4, P5, P6, P7), and the edge sets are (e 12, e23, e34, e45, e56, e 67).

Suppose the ordered set of edges is (e 45, e34, e23, e12, e56, e 67). The ordered edge sets may also be renumbered, e.g., (e 1, e2, e3, e4, e5, e 6), with e45 corresponding to e1, e34 corresponding to e2, e23 corresponding to e3, e12 corresponding to e4, e56 corresponding to e5, and e67 corresponding to e 6.

Step 301 is performed. At the beginning of clustering, the first edge of the first bit of the row of connected point cloud data, which is not clustered, is e1, i.e., e45.

Step 302 is performed again. At the beginning of clustering, the connected point cloud data is not clustered, and the connected point cloud data is different from the point cloud data connected by the first edge, and the second edge arranged at the second position is e3, i.e., e23 (because e2 arranged after e1 in the ordered edge set, i.e., e34, has the same vertex P4 as e 1).

Step 303 is next performed. Illustratively, the dissimilarity of the first edge with the second edge may be determined by: respectively determining the distance between two point cloud data connected with the first side and each point cloud data connected with the second side; the minimum distance among the distances is determined as the dissimilarity of the first side and the second side. And determining the minimum distance as the dissimilarity between the first edge and the second edge in the distance between the two point cloud data connected by the first edge and each point cloud data connected by the second edge, so that the accuracy and the clustering efficiency of subsequent clustering can be improved. In this example, the two point cloud data connected by the first edge e45 are P4 and P5, and the two point cloud data connected by the second edge e23 are P2 and P3. The distances of P4 and P2, and P4 and P3 are determined, respectively, and the distances of P5 and P2, and P5 and P3 are determined, respectively. The minimum distance among these four distances is taken as the dissimilarity between the first edge e45 and the second edge e 23.

Step 304 is then performed. If the dissimilarity between the first edge and the second edge is smaller than a preset dissimilarity threshold, the point cloud data connected with the second edge and the point cloud data connected with the first edge are gathered into one type. The preset dissimilarity threshold here may be determined from the performance of the laser sensor. The better the performance of the laser sensor, the smaller the preset dissimilarity threshold. Assuming that the dissimilarity between the first edge e45 and the second edge e23 is smaller than the preset dissimilarity threshold, the point cloud data connected to the second edge and the point cloud data connected to the first edge are clustered into one type, that is, P2, P3, P4 and P5 are clustered into one type.

In another case, if the dissimilarity between the first edge and the second edge is greater than or equal to a preset dissimilarity threshold, the point cloud data connected to the second edge and the point cloud data connected to the first edge are not clustered into one type.

Thereafter, step 305 is performed. And taking the edge which is arranged behind the second edge and is not clustered in the connected point cloud data in the ordered edge set and is different from the point cloud data connected with the first edge as a new second edge. After step 304 is performed, P2, P3, P4 and P5 have already been clustered, and P2 of the e12 connections in the ordered edge set that are arranged after e23 have already been clustered, so e12 cannot be taken as a new second edge. e56 has the same vertex as the first edge e45, and thus e67 is taken as a new second edge. Step 303 is performed back, i.e. the dissimilarity of e45 and e67 is determined. Step 304 is then performed. Assuming that the dissimilarity between the first edge e45 and the second edge e67 is greater than or equal to the preset dissimilarity threshold, e67 and e45 do not converge into one type. Then step 305 is executed, and the return execution is stopped, because there is no edge in which the connected point cloud data is not clustered and the connected point cloud data is different from the point cloud data connected to the first edge and is arranged behind the second edge.

Step 306 is then performed. And taking the edge, which is arranged behind the first edge and is not clustered by the connected point cloud data, in the ordered edge set as a new first edge. In the ordered edge set, the point cloud data P3 and P4 connected by e34 are clustered, the point cloud data P2 connected by e12 are clustered, and the point cloud data P5 connected by e56 are clustered. Thus, e67 is taken as the new first edge. However, the number of edges of the connected point cloud data that are not clustered is 1. The return stop condition of step 306 is satisfied, so the return execution is stopped, and each class after the primary clustering is acquired. The primary clustered classes are (P2, P3, P4, P5), (P1), (P6) and (P7), i.e., 4 classes are obtained.

Further, in the specific application, when the step 306 is executed, if the return stop condition of the step 306 is not satisfied, the step 302 needs to be returned.

In step 304, if the dissimilarity between the first edge and the second edge is smaller than the preset dissimilarity threshold, the number of the second edge may be updated to the number of the first edge, and finally, when each class after the primary clustering is obtained, the point cloud data connected by the edges with the same number is used as a class.

More specifically, a flag bit indicating whether or not clustering has been performed may also be set for each point cloud data. And after clustering is realized, the bit position valid bit of the point cloud data is used. When the first edge and the second edge are obtained, the zone bit of each point cloud data can be queried first, and then whether the point cloud data are clustered or not is determined according to the zone bit.

In this embodiment, the edges of the connected point cloud data that are not clustered refer to edges of the connected point cloud data that are not clustered.

According to the object-based perception method, the point cloud data is subjected to primary clustering through the Graphseg algorithm, and the areas in the low-variability image areas can be reserved through the clustering method based on the characteristics of the Graphseg algorithm, so that accurate clustering is achieved, and further, the object perception accuracy is improved.

The present embodiment also provides yet another object-based perception method. This embodiment describes in detail how the classes of markup semantic tags are formed based on the embodiments shown in fig. 1 and 3 and various alternatives. Fig. 5 is a schematic flow chart of determining a class of marking semantic tags in an object-based perception method according to an embodiment of the present invention. As shown in fig. 5, based on the semantic segmentation mode, the valid classes are clustered again to form classes marking semantic tags, which comprises the following steps:

step 501: and taking the minimum number of the point cloud data included in the effective class as the number of the effective class.

The serial numbers of the point cloud data are serial numbers formed by sequentially numbering at least one point cloud data acquired by the robot from left to right.

Step 502: and sequencing the effective classes according to the sequence from the small number to the large number of the effective classes, and obtaining a sequenced effective class set.

Step 503: and determining the distance between adjacent effective classes in the ordered effective class set.

Step 504: if the distance between the adjacent effective classes is smaller than a preset third distance threshold value, the adjacent effective classes are gathered into a class of marking the semantic tags.

Step 505: and if the distance between the adjacent effective classes is greater than or equal to a preset fourth distance threshold, taking the adjacent effective classes as the classes of the two marking semantic tags.

Specifically, in step 503, the distance of the intermediate point cloud data of the adjacent valid classes may be used as the distance between the adjacent valid classes. The intermediate point cloud data of the adjacent effective class refers to point cloud data located at an intermediate position after the point cloud data are ordered according to the number of the point cloud data in the effective class.

When the object is a person, the preset third distance threshold may be a preset width between the two legs.

And (5) performing secondary clustering on the effective classes through steps 501-505 to obtain at least one class marking the semantic tags.

Then, the distance between the class of each tag semantic label and the robot is determined. Determining the number of classes of the marking semantic tags with the distance from the robot being smaller than a preset first distance threshold value as the number of objects with the distance from the robot being smaller than the preset first distance threshold value, and controlling the movement of the robot according to the number of objects with the distance from the robot being smaller than the preset first distance threshold value.

According to the object-based perception method provided by the embodiment, clustering is performed again on each class after primary clustering to form the class of the mark semantic tag, on one hand, the clustering process is simple and efficient, the object perception efficiency is further improved, and on the other hand, the obtained class of the mark semantic tag can facilitate the object perception of a robot, so that the object perception accuracy is improved.

Fig. 6 is a schematic structural diagram of an object-based sensing device according to an embodiment of the invention. As shown in fig. 6, the object-based sensing device provided in this embodiment includes the following modules: the building module 61, the clustering module 62, the determining module 63 and the control module 64.

The establishing module 61 is configured to establish a graph structure according to at least one point cloud data collected by the robot.

And the clustering module 62 is used for clustering the point cloud data according to the weight value of the edge in the graph structure, and obtaining at least one class of the label semantic label after clustering.

A determining module 63, configured to determine, according to the clustered class of the at least one labeled semantic label, a number of objects whose distance from the robot is smaller than a preset first distance threshold.

Optionally, the determining module 63 is specifically configured to: and determining the number of classes of the marking semantic tags with the distance from the robot being smaller than a preset first distance threshold as the number of objects with the distance from the robot being smaller than the preset first distance threshold.

The control module 64 is configured to control the movement of the robot according to the number of objects whose distance to the robot is smaller than a preset first distance threshold.

Optionally, in one implementation, the control module 64 includes: and the control sub-module is used for controlling the robot to stop moving if the number of objects with the distance smaller than the preset first distance threshold value is larger than or equal to the preset object number threshold value.

More specifically, the control submodule is specifically configured to: if the number of objects with the distance from the robot smaller than the preset first distance threshold is larger than or equal to the preset object number threshold, setting the object zone bit of the frame corresponding to at least one point cloud data as an effective value; and if the number of frames with the object zone bits continuously being the effective value is larger than the preset frame threshold value, controlling the robot to stop moving.

In another implementation, the apparatus further includes: and the determining control module is used for determining the ratio of the number of the point cloud data with the distance from the robot smaller than a preset second distance threshold value to the total number in one frame of the point cloud data according to the point cloud data, and controlling the movement of the robot according to the ratio.

Optionally, the apparatus may further include: the device comprises an acquisition module and a deletion module.

The acquisition module is used for acquiring the acquired at least one original point cloud data.

And the deleting module is used for deleting the original point cloud data, of which the distance between the original point cloud data and the robot is not in the preset robot detection distance range, to form at least one filtered point cloud data.

The object-based sensing device provided by the embodiment of the invention can execute the object-based sensing method provided by any of the embodiments and various optional modes of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

The invention also provides another object-based sensing device. The specific structure of the clustering module 62 will be described in detail based on the embodiment shown in fig. 6 and various alternatives. With continued reference to fig. 6, in the object-based sensing device provided in this embodiment, the clustering module 62 specifically includes: the acquisition sub-module 621, the first clustering sub-module 622, and the second clustering sub-module 623.

The obtaining sub-module 621 is configured to sort the edges in order from small to large, and obtain the sorted edge set.

The first clustering sub-module 622 is configured to, if the dissimilarity between two sides, which do not include common vertices, in the ordered edge set is smaller than a preset dissimilarity threshold, group the point cloud data connected to the two sides into one class, and obtain each class after at least one primary clustering.

Optionally, the first clustering sub-module 622 is specifically configured to:

acquiring a first edge of the first bit, which is not clustered, of the connected point cloud data in the ordered edge set;

acquiring a second edge which is arranged at a second position and is different from the point cloud data connected with the first edge, wherein the point cloud data connected with the second edge are not clustered in the ordered edge set;

determining dissimilarity between the first edge and the second edge;

if the dissimilarity between the first edge and the second edge is smaller than a preset dissimilarity threshold, the point cloud data connected with the second edge and the point cloud data connected with the first edge are gathered into one type;

returning to the step of determining the dissimilarity between the first edge and the second edge by taking the edge which is arranged behind the second edge and is not clustered by the connected point cloud data and is different from the point cloud data connected to the first edge in the ordered edge set as a new second edge until the edge which is not clustered by the connected point cloud data and is arranged behind the second edge and is different from the point cloud data connected to the first edge does not exist in the ordered edge set, and stopping returning to the step;

And returning the edges which are arranged behind the first edge and are not clustered by the connected point cloud data in the ordered edge set to execute the steps of acquiring the second edges which are arranged at the second position and are different from the connected point cloud data of the first edge in the ordered edge set until the first edges which are not clustered by the connected point cloud data are not arranged at the first position in the ordered edge set, or the number of the edges which are not clustered by the connected point cloud data in the ordered edge set is 1, or the second edges which are not clustered by the connected point cloud data and are connected with the first edge and are arranged at the second position in the ordered edge set are not exist in the ordered edge set, stopping executing, and acquiring all the classes after primary clustering.

A second clustering sub-module 623, configured to re-cluster each class after at least one primary clustering according to the furthest distance between the point cloud data included in each class after at least one primary clustering, to form a class of marking semantic tags.

Optionally, the second aggregation sub-module 623 is specifically configured to: determining the furthest distance between the point cloud data included in each class after primary clustering; determining the class which is the farthest distance between the preset distance ranges and is subjected to primary clustering as an effective class; and clustering the effective classes again based on the semantic segmentation mode to form classes marked with semantic tags.

More specifically, in terms of clustering valid classes again based on semantic segmentation, forming classes that label semantic tags, the second aggregation sub-module 623 is specifically configured to:

taking the minimum number of the point cloud data included in the effective class as the number of the effective class, wherein the number of the point cloud data is a number formed by numbering at least one acquired point cloud data according to a left-to-right sequence;

sorting the effective classes according to the sequence from small to large of the serial numbers of the effective classes, and obtaining a sorted effective class set;

and if the distance between the adjacent effective classes is greater than or equal to a preset third distance threshold, taking the adjacent effective classes as the classes of the two marking semantic tags.

Optionally, the class of the semantic tag is a class of human, the effective class is a class of human legs, the preset distance range is a preset width range of human legs, and the preset third distance threshold is a preset width between two legs.

Fig. 7 is a schematic structural diagram of a robot according to an embodiment of the present invention. As shown in fig. 7, the robot includes a processor 70 and a memory 71. The number of processors 70 in the robot may be one or more, one processor 70 being taken as an example in fig. 7; the processor 70 and the memory 71 of the robot may be connected by a bus or otherwise, in fig. 7 by way of example.

The memory 71 is a computer readable storage medium, and may be used to store software programs, computer executable programs, and modules, such as program instructions and modules corresponding to the object-based sensing method in the embodiment of the present invention (e.g., the creation module 61, the clustering module 62, the determination module 63, and the control module 64 in the object-based sensing device). The processor 70 performs various functional applications of the robot and data processing, i.e., implements the above-described object-based perception method, by running software programs, instructions and modules stored in the memory 71.

The memory 71 may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, at least one application program required for functions; the storage data area may store data created according to the use of the robot, etc. In addition, memory 71 may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, memory 71 may further include memory remotely located with respect to processor 70, which may be connected to the robot via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

Optionally, the robot may further include: a power supply assembly 72, an audio assembly 73, a communication assembly 74, and a sensor assembly 75. The power supply assembly 72, the audio assembly 73, the communication assembly 74, and the sensor assembly 75 may all be coupled to the processor 70 via a bus.

The power supply assembly 72 provides power to the various components of the robot. The power components 72 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the robot.

The audio component 73 is configured to output and/or input audio signals. For example, the audio component 73 includes a microphone configured to receive external audio signals when the robot is in an operational mode, such as a recording mode and a speech recognition mode. The received audio signals may be further stored in the memory 71 or transmitted via the communication component 74. In some embodiments, audio component 73 further comprises a speaker for outputting audio signals.

The communication component 74 is configured to facilitate communication between the robot and other devices in a wired or wireless manner. The robot may access a wireless network based on a communication standard. In one exemplary embodiment, the communication component 74 receives broadcast signals or broadcast related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, the communication component 74 further includes a near field communication (Near Field Communication, NFC) module to facilitate short range communications. For example, the NFC module may be implemented based on radio frequency identification (Radio Frequency Identification, RFID) technology, infrared data association technology, ultra wideband technology, bluetooth technology, and other technologies.

The sensor assembly 75 includes one or more sensors for providing status assessment of various aspects of the robot. The sensor assembly 75 may include a laser sensor for collecting point cloud data. In some embodiments, the sensor assembly 75 may also include an acceleration sensor, a magnetic sensor, a pressure sensor, a temperature sensor, or the like.

Fig. 8 is a schematic structural diagram of a computer readable storage medium according to an embodiment of the present invention. As shown in fig. 8, the present invention also provides a computer-readable storage medium 82 containing computer-executable instructions 81, which computer-executable instructions 81, when executed by a processor 83, are for performing an object-based perception method, the method comprising:

Of course, the storage medium containing computer executable instructions provided in the embodiments of the present invention is not limited to the above-described method operations, but may also perform related operations in the object-based perception method provided in any embodiment of the present invention.

From the above description of embodiments, it will be clear to a person skilled in the art that the present invention may be implemented by means of software and necessary general purpose hardware, but of course also by means of hardware, although in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as a floppy disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a FLASH Memory (FLASH), a hard disk or an optical disk of a computer, etc., including several instructions for causing a robot (which may be a personal computer, a vehicle, a network device, etc.) to perform the object-based perception method according to the embodiments of the present invention.

It should be noted that, in the embodiment of the object-based sensing device, each unit and module included are only divided according to the functional logic, but not limited to the above-mentioned division, so long as the corresponding functions can be implemented; in addition, the specific names of the functional units are also only for distinguishing from each other, and are not used to limit the protection scope of the present invention.

Note that the above is only a preferred embodiment of the present invention and the technical principle applied. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, while the invention has been described in connection with the above embodiments, the invention is not limited to the embodiments, but may be embodied in many other equivalent forms without departing from the spirit or scope of the invention, which is set forth in the following claims.

Claims

1. An object-based perception method, comprising:

controlling the robot according to the number of objects with the distance to the robot smaller than a preset first distance threshold;

after the robot is controlled according to the number of objects with the distance to the robot smaller than a preset first distance threshold, the method further comprises:

controlling the movement of the robot according to the ratio;

2. The method of claim 1, wherein a vertex set in the graph structure is a set of coordinate values of the point cloud data, an edge set in the graph structure is a set of edges connecting adjacent point cloud data, and weights of the edges of the adjacent point cloud data are determined according to distances between the adjacent point cloud data.

3. The method of claim 2, wherein the reclustering the at least one primary clustered classes according to the furthest distance between the point cloud data included in the at least one primary clustered class, forming a class of labeled semantic labels, comprises:

4. A method according to claim 3, wherein said clustering the valid classes again based on semantic segmentation to form classes that label semantic tags comprises:

5. The method of claim 4, wherein the class of the labeled semantic tag is a human class, the valid class is a human leg class, the preset distance range is a preset human leg width range, and the preset third distance threshold is a preset width between two legs.

6. The method of claim 1, wherein determining the number of objects having a distance to the robot less than a preset first distance threshold from the clustered class of at least one labeled semantic label comprises:

7. The method according to any one of claims 1-6, wherein controlling the robot according to the number of objects having a distance to the robot smaller than a preset first distance threshold comprises:

8. The method of claim 7, wherein controlling the robot to stop moving if the number of objects having a distance to the robot less than a preset first distance threshold is greater than or equal to a preset object number threshold comprises:

9. The method of any of claims 1-6, wherein prior to establishing a graph structure from the collected at least one point cloud data, the method further comprises:

acquiring at least one piece of acquired original point cloud data;

10. An object-based perception device, comprising:

the control module is used for controlling the movement of the robot according to the number of objects with the distance to the robot being smaller than a preset first distance threshold;

The determining control module is used for determining the ratio of the number of the point cloud data with the distance from the robot being smaller than a preset second distance threshold value to the total number in one frame of point cloud data according to the point cloud data, and controlling the movement of the robot according to the ratio;

the first clustering sub-module is used for gathering the point cloud data connected with the two sides into one class if the dissimilarity between the two sides which do not contain the common vertexes in the ordered edge set is smaller than a preset dissimilarity threshold value, and acquiring at least one class after primary clustering;

11. A robot, the robot comprising:

one or more processors;

a memory for storing one or more programs;

the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the object-based perception method of any of claims 1-9.

12. A computer readable storage medium, on which a computer program is stored, which program, when being executed by a processor, implements an object-based perception method as claimed in any one of claims 1-9.