Noise point determination method and device, robot and storage medium
By screening and calculating the centroid and offset distance of the noise set, the accuracy of noise determination is improved, the problem of low accuracy of noise determination in the prior art is solved, and the accuracy of robot obstacle avoidance is enhanced.
Patent Information
- Application Number
- CN202411165783.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-22
- Publication Date
- 2024-12-03
AI Technical Summary
In the prior art, the accuracy of noise determination in the obstacle point cloud is low, resulting in misjudgment in the robot's obstacle avoidance, and it is impossible to effectively distinguish between real obstacles and noise.
By obtaining the obstacle point cloud twice, the candidate noise set with a number smaller than the threshold is selected, and the straight line distance and offset distance between the centroids are calculated. If the offset distance is greater than the threshold, it is determined as noise, which improves the accuracy of noise determination.
It effectively improves the accuracy of noise determination, avoids misjudging isolated points or obstacles with smaller sizes as noise, and enhances the accuracy of robot obstacle avoidance.
Smart Images

Figure CN119063729A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of artificial intelligence technology, and in particular to a noise determination method, device, robot and storage medium. Background Art
[0002] With the development of artificial intelligence technology, robots are widely used in life and production. During operation, robots need to avoid obstacles based on the obstacle point cloud detected by sensors. In order to improve the accuracy of obstacle avoidance, the obstacle point cloud needs to be filtered to eliminate noise in the obstacle point cloud.
[0003] In the traditional technology, the shortest straight-line distance between an obstacle point in the obstacle point cloud and other obstacle points is calculated. When the shortest straight-line distance is greater than a distance threshold, the obstacle point is determined as a noise point, resulting in low accuracy in noise point determination. Summary of the invention
[0004] Based on this, it is necessary to provide a noise determination method, device, robot and storage medium that can improve the accuracy of noise determination in order to solve the above technical problems.
[0005] In a first aspect, the present application provides a noise point determination method. The method comprises:
[0006] Acquire at least one first candidate noise point set corresponding to a first moment, and at least one second candidate noise point set corresponding to a second moment; the first moment is before the second moment; the number of obstacle points in the first candidate noise point set is less than a quantity threshold, and the number of obstacle points in the second candidate noise point set is less than the quantity threshold;
[0007] Determine, based on a straight-line distance between a centroid of the first candidate noise point set and a centroid of the second candidate noise point set, a first candidate noise point set corresponding to the second candidate noise point set;
[0008] Determine an offset distance between a centroid of the second candidate noise point set and a centroid of the first candidate noise point set corresponding to the second candidate noise point set;
[0009] When the offset distance is greater than the distance threshold, the obstacle point in the second candidate noise point set corresponding to the offset distance is determined as the target noise point.
[0010] In a second aspect, the present application also provides a noise point determination device. The device comprises:
[0011] an acquisition module, configured to acquire at least one first candidate noise point set corresponding to a first moment, and at least one second candidate noise point set corresponding to a second moment; the first moment is before the second moment; the number of obstacle points in the first candidate noise point set is less than a quantity threshold, and the number of obstacle points in the second candidate noise point set is less than the quantity threshold;
[0012] a matching module, configured to determine a first candidate noise point set corresponding to the second candidate noise point set based on a straight-line distance between a centroid of the first candidate noise point set and a centroid of the second candidate noise point set;
[0013] a calculation module, configured to determine an offset distance between a centroid of the second candidate noise point set and a centroid of the first candidate noise point set corresponding to the second candidate noise point set;
[0014] The determination module is used to determine the obstacle point in the second candidate noise point set corresponding to the offset distance as the target noise point when the offset distance is greater than the distance threshold.
[0015] In a third aspect, the present application further provides a robot comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the methods described in the first aspect when executing the computer program.
[0016] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of any one of the methods described in the first aspect.
[0017] In a fifth aspect, the present application further provides a computer program product, comprising a computer program, which, when executed by a processor, implements the steps of any one of the methods described in the first aspect.
[0018] In this embodiment, at least one first candidate noise point set is determined by the first obstacle point cloud collected at the first moment, and at least one second candidate noise point set is determined by the second obstacle point cloud collected at the second moment. The number of obstacle points in the first candidate noise point set is less than the number threshold, indicating that the obstacle points in the first candidate noise point set are isolated points, which may be noise points formed by the sensor or real obstacle points corresponding to obstacles of smaller sizes. Similarly, the obstacle points in the second candidate noise point set may also be noise points formed by the sensor or real obstacle points corresponding to obstacles of smaller sizes, thereby achieving preliminary screening of noise points. On this basis, based on the straight-line distance between the centroid of the first candidate noise point set and the centroid of the second candidate noise point set, the first candidate noise point set corresponding to the second candidate noise point set is determined. The second candidate noise point set and the corresponding first candidate noise point set respectively represent the same object detected by the robot at different times, that is, the first candidate noise point set represents the object detected by the robot at the first moment, and the second candidate noise point set represents the object detected by the robot at the second moment. If the offset distance between the second candidate noise point set and the corresponding first candidate noise point set is greater than the distance threshold, it indicates that the appearance of the object is random. In view of the randomness of the appearance of noise points, it means that the object is not a real obstacle, that is, the obstacle points in the second candidate noise point set are noise points; if the offset distance between the second candidate noise point set and the corresponding first candidate noise point set is equal to or less than the distance threshold, it indicates that the appearance of the object is fixed and is a real obstacle. The reason why the obstacle points in the second candidate noise point set are isolated points is that the size of the object is small, that is, the obstacle points in the second candidate noise point set are not noise points, and the noise points are screened again. Through the above two screenings, it is avoided to directly determine isolated points as noise points, that is, it is avoided to determine obstacle points corresponding to obstacles of smaller sizes as noise points, thereby improving the accuracy of target noise point determination. BRIEF DESCRIPTION OF THE DRAWINGS
[0019] Figure 1 A diagram showing an application environment of a noise determination method in an embodiment;
[0020] Figure 2 is a flow chart of a noise point determination method in one embodiment;
[0021] Figure 3 It is a flowchart of the step of obtaining the first candidate noise point set in one embodiment;
[0022] Figure 4 is a schematic diagram of a first obstacle point cloud in another embodiment;
[0023] Figure 5 A schematic diagram of a flow chart of a step of determining a first candidate noise point set in an embodiment;
[0024] Figure 6 A schematic diagram of a noise point determination process in one embodiment;
[0025] Figure 7 is a structural block diagram of a noise point determination device in one embodiment;
[0026] Figure 8 1 is a diagram showing the internal structure of a robot in one embodiment. DETAILED DESCRIPTION
[0027] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0028] The noise determination method provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the terminal 102 communicates with the robot 104 through the network. The terminal 102 and the robot 104 can be used together to execute the noise determination method provided in the embodiment of the present application, and the robot 104 can also be used alone to execute the noise determination method provided in the embodiment of the present application. Among them, the robot 104 can be various self-moving devices, such as: cleaning robots, delivery robots, guide robots, automatic guided vehicles (Automatic Guided Vehicle, AGV) and sweepers. The terminal 102 can be but is not limited to various personal computers, laptops, smart phones, tablets, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart TVs and smart car-mounted devices, etc., and the portable wearable devices can be smart watches, etc.
[0029] The robot 104 can also cooperate with the server to implement the noise determination method provided in the embodiment of the present application, wherein the server can be an independent physical server or a service node in the blockchain system, and each service node in the blockchain system forms a peer-to-peer (P2P) network, and the P2P protocol is an application layer protocol running on the Transmission Control Protocol (TCP) protocol. In addition, the server can also be a server cluster composed of multiple physical servers, which can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content distribution networks (Content Delivery Network, CDN), and big data and artificial intelligence platforms. The robot 104 can be connected to the server through Bluetooth, USB (Universal Serial Bus) or network communication connection methods, and this application does not limit this.
[0030] In one embodiment, Figure 2 As shown, a noise point determination method is provided. This embodiment is described by taking a robot executing the noise point determination method as an example, including steps 202 to 208.
[0031] Step 202, obtaining at least one first candidate noise point set corresponding to a first moment, and at least one second candidate noise point set corresponding to a second moment; the first moment is before the second moment; the number of obstacle points in the first candidate noise point set is less than a quantity threshold, and the number of obstacle points in the second candidate noise point set is less than a quantity threshold.
[0032] Among them, the first moment refers to the moment when the robot collects the first obstacle point cloud through the sensor, and the second moment refers to the moment when the robot collects the second obstacle point cloud through the sensor. The first moment is before the second moment, and the first moment and the second moment can be two adjacent moments or non-adjacent moments. The first candidate noise point set refers to a set of obstacle points that may be noise points in the first obstacle point cloud, and the number of the first candidate noise point set is at least one, that is, the number of the first candidate noise point set can be one or more. The second candidate noise point set refers to a set of obstacle points that may be noise points in the second obstacle point cloud, and the number of the second candidate noise point set is at least one, that is, the number of the second candidate noise point set can be one or more. The positions of the obstacle points in the first candidate noise point set and the positions of the obstacle points in the second candidate noise point set are located in the same coordinate system. The quantity threshold refers to the minimum number used to determine the candidate noise point set, for example, the quantity threshold can be 2 or 3.
[0033] Exemplarily, during the operation of the robot, based on a preset time interval, an obstacle point cloud is acquired at a first moment, the obstacle point cloud acquired at the first moment is determined as a first obstacle point cloud, at least one first candidate noise point set is determined based on the first obstacle point cloud, an obstacle point cloud is acquired at a second moment, the obstacle point cloud acquired at the second moment is determined as a second obstacle point cloud, at least one second candidate noise point set is determined based on the second obstacle point cloud, and the positions of the obstacle points in the first candidate noise point set and the second candidate noise point set are located in the same coordinate system. The second moment is equal to the first moment plus the preset time interval. The preset time interval can be determined according to the running speed of the robot. The faster the running speed of the robot, the smaller the preset time interval is, so as to ensure that the first obstacle point cloud and the second obstacle point cloud have point cloud data of the same obstacle.
[0034] Step 204: Determine the first candidate noise point set corresponding to the second candidate noise point set based on the straight-line distance between the centroid of the first candidate noise point set and the centroid of the second candidate noise point set.
[0035] Among them, the centroid refers to the geometric center of all points in a set, the centroid of the first candidate noise point set refers to the geometric center of all obstacle points in the first candidate noise point set, and the centroid of the second candidate noise point set refers to the geometric center of all obstacle points in the second candidate noise point set. It can be understood that when there is only one obstacle point in the candidate noise point set, the obstacle point is the centroid of the candidate noise point set; when there are multiple obstacle points in the candidate noise point set, the geometric centers of the multiple obstacle points are the centroid of the candidate noise point set.
[0036] The first candidate noise point set corresponding to the second candidate noise point set refers to the first candidate noise point set that matches the second candidate noise point set when the distance cost is minimized. It can be understood that one-to-one matching is performed on at least one first candidate noise point set and at least one second candidate noise point set to obtain multiple sets of matched first candidate noise point sets and second candidate noise sets, and the straight-line distances between the centroid of each set of matched first candidate noise point sets and the centroid of the second candidate noise set are summed to obtain the distance costs of the multiple sets of matched first candidate noise point sets and second candidate noise sets. When the distance cost is minimized, the first candidate noise set that matches the second candidate noise set is the first candidate noise set corresponding to the second candidate noise set. For example, the Hungarian matching algorithm can be used to determine the first candidate noise set corresponding to each second candidate noise set.
[0037] It should be noted that, when there are multiple first candidate noise point sets and multiple second candidate noise point sets, the object represented by a first candidate noise point set and the object represented by a second candidate noise point set may be the same object detected by the robot at different times. In order to match the first candidate noise point set and the second candidate noise set representing the same object, the first candidate noise set representing the same object as the second candidate noise set is determined at the cost of the straight-line distance between the first candidate noise set and the second candidate noise set, that is, the first candidate noise set corresponding to the second candidate noise set is determined.
[0038] Exemplarily, the robot determines the center of mass of the first candidate noise point set based on the positions of obstacle points in the first candidate noise point set to obtain the first center of mass, determines the center of mass of the second candidate noise point set based on the positions of obstacle points in the second candidate noise point set to obtain the second center of mass, calculates the straight-line distance between the first center of mass and the second center of mass, uses the straight-line distance as the matching cost, matches at least one first candidate noise point set and at least one second candidate noise point set, and obtains a first candidate noise point set corresponding to each second candidate noise point set.
[0039] In one embodiment, the straight-line distance between the centroid of the first candidate noise point set and the centroid of the second candidate noise point set is determined, and the straight-line distance is used as the matching cost, and the first candidate noise point set corresponding to each second candidate noise point set is determined using the Hungarian matching algorithm. This embodiment uses the Hungarian algorithm to determine the optimal matching of multiple first candidate noise point sets and multiple second candidate noise point sets, so that each second candidate noise point set can determine a corresponding first candidate noise point set under the condition of the lowest matching cost, which provides basic data for the subsequent determination of whether the objects represented by the second candidate noise set and the corresponding first candidate noise set are random.
[0040] In one embodiment, the method for determining the centroid of the first candidate noise point set includes: obtaining the first initial noise point set; obtaining the posture corresponding to the first moment to obtain the first posture, and obtaining the posture corresponding to the second moment to obtain the second posture; determining the centroid corresponding to the first initial noise point set to obtain the initial centroid; based on the first posture, converting the initial centroid to the corresponding intermediate centroid in the world coordinate system; based on the second posture, converting the intermediate centroid to the target centroid; the target centroid is located in the robot coordinate system with the second posture as the origin; determining the first initial noise point set as the first candidate noise point set, and determining the target centroid corresponding to the first initial noise point set as the centroid corresponding to the corresponding first candidate noise point set. Wherein, the method for determining the first initial noise point set includes: obtaining the obstacle point cloud corresponding to the first moment to obtain the first obstacle point cloud; clustering the first obstacle point cloud to obtain the first cluster cluster; counting the obstacle points in the first cluster cluster to obtain the first statistical quantity; comparing the first statistical quantity with the quantity threshold, and when the first statistical quantity is less than the quantity threshold, determining the first cluster cluster corresponding to the first statistical quantity as the first initial noise point set. That is, by converting the initial center of mass of the first initial noise point set located in the robot coordinate system with the first posture as the origin into the target center of mass of the robot coordinate system with the second posture as the origin, the target center of mass corresponding to the first initial noise point set is determined as the center of mass corresponding to the corresponding first candidate noise point set. The center of mass corresponding to the first candidate noise point set and the center of mass of the second candidate noise point set are located in the same coordinate system, which lays the foundation for subsequent calculations.
[0041] Step 206: determine the offset distance between the centroid of the second candidate noise point set and the centroid of the first candidate noise point set corresponding to the second candidate noise point set.
[0042] The offset distance refers to the straight-line distance between the centroid of the second candidate noise point set and the centroid of the first candidate noise point set corresponding to the second candidate noise point set.
[0043] Exemplarily, for each second candidate noise point set, the robot calculates the straight-line distance between the centroid of the second candidate noise point set and the centroid of the first candidate noise point set corresponding to the second candidate noise point set, and obtains the offset distance between the second candidate noise point set and the first candidate noise point set corresponding to the second candidate noise point set.
[0044] Step 208: When the offset distance is greater than the distance threshold, the obstacle point in the second candidate noise point set corresponding to the offset distance is determined as the target noise point.
[0045] The distance threshold refers to the minimum offset distance used to determine the target noise point, and the distance threshold can be determined according to the measurement error of the sensor.
[0046] Exemplarily, the robot compares the offset distance corresponding to the second candidate noise point set with the distance threshold. If the offset distance is greater than the distance threshold, the obstacle point in the second candidate noise point set is determined as the target noise point; if the offset distance is less than or equal to the distance threshold, the obstacle point in the second candidate noise point set is determined as the real obstacle point.
[0047] In this embodiment, at least one first candidate noise point set is determined by the first obstacle point cloud collected at the first moment, and at least one second candidate noise point set is determined by the second obstacle point cloud collected at the second moment. The number of obstacle points in the first candidate noise point set is less than the number threshold, indicating that the obstacle points in the first candidate noise point set are isolated points, which may be noise points formed by the sensor or real obstacle points corresponding to obstacles of smaller sizes. Similarly, the obstacle points in the second candidate noise point set may also be noise points formed by the sensor or real obstacle points corresponding to obstacles of smaller sizes, thereby achieving preliminary screening of noise points. On this basis, based on the straight-line distance between the centroid of the first candidate noise point set and the centroid of the second candidate noise point set, the first candidate noise point set corresponding to the second candidate noise point set is determined. The second candidate noise point set and the corresponding first candidate noise point set respectively represent the same object detected by the robot at different times, that is, the first candidate noise point set represents the object detected by the robot at the first moment, and the second candidate noise point set represents the object detected by the robot at the second moment. If the offset distance between the second candidate noise point set and the corresponding first candidate noise point set is greater than the distance threshold, it indicates that the appearance of the object is random. In view of the randomness of the appearance of noise points, it means that the object is not a real obstacle, that is, the obstacle points in the second candidate noise point set are noise points; if the offset distance between the second candidate noise point set and the corresponding first candidate noise point set is equal to or less than the distance threshold, it indicates that the appearance of the object is fixed and is a real obstacle. The reason why the obstacle points in the second candidate noise point set are isolated points is that the size of the object is small, that is, the obstacle points in the second candidate noise point set are not noise points, and the noise points are screened again. Through the above two screenings, it is avoided to directly determine isolated points as noise points, that is, it is avoided to determine obstacle points corresponding to obstacles of smaller sizes as noise points, thereby improving the accuracy of target noise point determination.
[0048] In one embodiment, Figure 3 As shown, obtaining at least one first candidate noise point set corresponding to the first moment includes:
[0049] Step 302: Obtain an obstacle point cloud corresponding to a first moment to obtain a first obstacle point cloud.
[0050] The obstacle point cloud refers to a collection of multiple obstacle points. The first obstacle point cloud refers to the obstacle point cloud detected by the robot at the first moment, for example, Figure 4 The set of multiple obstacle points detected at the first moment is the first obstacle point cloud. The first obstacle point cloud is detected by the robot's sensor, which includes but is not limited to at least one of a laser radar sensor, a depth camera, and an infrared sensor, etc., which are not limited here. The positions of the obstacle points in the first obstacle point cloud are located in the historical robot coordinate system, and the historical robot coordinate system takes the robot's posture at the first moment as the origin.
[0051] Exemplarily, the robot acquires an obstacle point cloud through a sensor at a first moment to obtain a first obstacle point cloud.
[0052] Step 304: cluster the first obstacle point cloud to obtain a first cluster.
[0053] The cluster refers to a point cloud cluster composed of at least one obstacle point. The first cluster refers to a point cloud cluster composed of multiple obstacle points in the first obstacle point cloud. It can be understood that the first cluster is a subset of the first obstacle point cloud.
[0054] Exemplarily, the robot selects a target obstacle point from the first obstacle point cloud, calculates the straight-line distance between the target obstacle point and other obstacle points respectively, and determines the obstacle point whose straight-line distance is less than the clustering distance threshold as the target obstacle point. For each determined target obstacle point, the robot repeats the above steps of calculating the straight-line distance between the target obstacle point and other obstacle points respectively, and determining the obstacle point whose straight-line distance is less than the clustering distance threshold as the target obstacle point, until there is no obstacle point whose straight-line distance is less than the clustering distance threshold, and determines multiple target obstacle points as the first clustering cluster. Repeat the above steps for the remaining obstacle points in the first obstacle point cloud to obtain multiple first clusters. The clustering distance threshold refers to the maximum straight-line distance for determining an obstacle point in the first obstacle point cloud as a target obstacle point.
[0055] Step 306: Count the obstacle points in the first cluster to obtain a first statistical quantity.
[0056] The first statistical quantity refers to the number of obstacle points included in the first cluster.
[0057] Exemplarily, for each first cluster, the robot counts the obstacle points in the first cluster to obtain a first statistical quantity.
[0058] Step 308: compare the first statistical quantity with a quantity threshold, and when the first statistical quantity is less than the quantity threshold, determine the first cluster corresponding to the first statistical quantity as the first initial noise point set.
[0059] The first initial noise point set refers to the first cluster whose first statistical quantity is less than the quantity threshold. The initial positions of the obstacle points in the first initial noise point set are located in the historical robot coordinate system.
[0060] Exemplarily, the robot compares the first statistical quantity with the quantity threshold. If the first statistical quantity is less than the quantity threshold, the first cluster corresponding to the first statistical quantity is determined as the first initial noise point set, that is, the cluster with fewer obstacle points is selected as the initial noise point set.
[0061] Step 310: Based on the first initial noise point set, obtain a first candidate noise point set corresponding to the first initial noise point set.
[0062] Among them, the first candidate noise point set refers to a set composed of obstacle points corresponding to the obstacle points in the first initial noise point set in the current robot coordinate system. It can be understood that the target position of the obstacle point in the first candidate noise point set is located in the current robot coordinate system, and the current robot coordinate system takes the robot's posture at the second moment as the origin.
[0063] Exemplarily, for each first initial noise point set, based on the posture corresponding to the first moment and the posture corresponding to the second moment, the initial position of the obstacle point in the first initial noise point set is converted to obtain the first candidate noise point set corresponding to the first initial noise point set. In this embodiment, by clustering the first obstacle point cloud, multiple first clusters are obtained, that is, multiple adjacent first obstacle points are combined into a first cluster. If the statistical number of obstacle points in the first cluster is less than the number threshold, it means that the obstacle point in the first cluster is an isolated point, and the isolated point may be a noise point. Therefore, the above first cluster is determined as the first initial noise point set. It can be understood that based on the statistical number of obstacle points in the first cluster, the obstacle points that may be noise points in the first obstacle point cloud are preliminarily screened. Based on the first initial noise point set, the first candidate noise point set corresponding to the first initial noise point set is obtained, that is, since the first initial noise point set and the second candidate noise set are located in different coordinate systems, the first initial noise point set is converted into a first candidate noise point set located in the same coordinate system as the second candidate noise set, providing accurate basic data for the subsequent determination of the offset distance.
[0064] In one embodiment, Figure 5 As shown, based on the first initial noise point set, a first candidate noise point set corresponding to the first initial noise point set is obtained, including:
[0065] Step 502, obtaining the posture corresponding to the first moment to obtain the first posture, and obtaining the posture corresponding to the second moment to obtain the second posture.
[0066] Among them, the first pose refers to the pose of the robot in the world coordinate system at the first moment, and the second pose refers to the pose of the robot in the world coordinate system at the second moment. The pose refers to the position and posture of the robot. The position represents the specific position of the robot in space and can be represented by two-dimensional coordinates or three-dimensional coordinates. The pose represents the rotation state of the robot in space, usually represented by the rotation angle. For example, if the position is (X, Y) and the rotation angle is θ, then the pose is (X, Y, θ).
[0067] Exemplarily, the robot obtains a posture corresponding to a first moment to obtain a first posture, and obtains a posture corresponding to a second moment to obtain a second posture.
[0068] Step 504: for the initial position of the obstacle point in the first initial noise point set, based on the first pose, convert the initial position into a corresponding intermediate position in the world coordinate system.
[0069] Among them, the world coordinate system refers to the global coordinate system, which can be understood as a global reference system that does not change with the movement of the robot or other objects. The initial position refers to the position of the obstacle point in the historical robot coordinate system, and the intermediate position refers to the position of the obstacle point in the world coordinate system.
[0070] Exemplarily, for each obstacle point in the first initial noise point set, the robot obtains the initial position of its obstacle point from the first initial noise point set, calculates the initial position and the first pose, and obtains the intermediate position corresponding to its obstacle point.
[0071] In one embodiment, the initial position of the obstacle point in the first initial noise point set is (x1, y1), and the first pose is (x A ,Y A ,θ A ), then the middle position corresponding to the obstacle point (x M ,y M ) is shown below:
[0072] Formula (1)
[0073] Step 506: based on the second posture, convert the intermediate position into a target position; the target position is located in the robot coordinate system with the second posture as the origin.
[0074] The target position refers to the position of the obstacle point in the first initial noise point set in the current robot coordinate system, and the current robot coordinate system is a coordinate system with the second posture as the origin.
[0075] Exemplarily, the robot performs calculations on the intermediate position and the second posture to obtain a target position corresponding to the obstacle point in the first initial noise point set.
[0076] In one embodiment, the middle position (x M ,y M ), the second pose is (X B ,Y B ,θ B ), then the target position (x2, y2) corresponding to the obstacle point is as follows:
[0077] Formula (2)
[0078] Step 508: Based on the target positions corresponding to the obstacle points in the first initial noise point set, a first candidate noise point set corresponding to the first initial noise point set is obtained.
[0079] Exemplarily, the robot obtains a first candidate noise point set based on a target position corresponding to each obstacle point in the first initial noise point set.
[0080] In this embodiment, the robot converts the first initial noise point set into a first candidate noise point set based on the posture at the first moment and the posture at the second moment. The first candidate noise point set and the second candidate noise point set are located in the same coordinate system, that is, in the current robot coordinate system with the second posture as the origin, providing accurate basic data for the subsequent determination of the offset distance.
[0081] In one embodiment, obtaining at least one second candidate noise point set corresponding to the second moment includes:
[0082] Obtain an obstacle point cloud corresponding to the second moment to obtain a second obstacle point cloud; cluster the second obstacle point cloud to obtain a second cluster cluster; count the obstacle points in the second cluster cluster to obtain a second statistical quantity; compare the second statistical quantity with a quantity threshold, and when the second statistical quantity is less than the quantity threshold, determine the second cluster cluster corresponding to the second statistical quantity as a second candidate noise point set.
[0083] Among them, the obstacle point cloud refers to a set of multiple obstacle points acquired by the robot. It can be understood that the second obstacle point cloud refers to a set of obstacle points detected by the robot at the second moment. The second cluster refers to a set consisting of at least one obstacle point in the second obstacle point cloud. It can be understood that the second cluster is a subset of the second obstacle point cloud. The second statistical quantity refers to the number of second obstacle points included in the second cluster. The second candidate noise point set refers to a second cluster whose second statistical quantity is less than the quantity threshold.
[0084] Exemplarily, the robot obtains the obstacle point cloud corresponding to the second moment to obtain the second obstacle point cloud, selects the target obstacle point from the second obstacle point cloud, calculates the straight-line distance between the target obstacle point and other obstacle points respectively, and determines the obstacle point whose straight-line distance is less than the clustering distance threshold as the target obstacle point. For each determined target obstacle point, the above steps of respectively calculating the straight-line distance between the target obstacle point and other obstacle points and determining the obstacle point whose straight-line distance is less than the clustering distance threshold as the target obstacle point are repeatedly performed until there is no obstacle point whose straight-line distance is less than the clustering distance threshold, and multiple target obstacle points are determined as second cluster clusters; the above steps are repeated for the remaining obstacle points in the second obstacle point cloud to obtain multiple second cluster clusters. For each second cluster cluster, the obstacle points in the second cluster cluster are counted to obtain a second statistical quantity, and the second statistical quantity is compared with the quantity threshold. If the second statistical quantity is less than the quantity threshold, the second cluster cluster corresponding to the second statistical quantity is determined as the second candidate noise point set.
[0085] In this embodiment, the second obstacle point cloud is clustered to obtain a second cluster, that is, multiple adjacent obstacle points in the second obstacle point cloud are combined into a second cluster. If the statistical number of obstacle points in the second cluster is less than the quantity threshold, it means that the obstacle points in the second cluster are isolated points, and the isolated points may be noise points. Therefore, the above second cluster is determined as the second candidate noise point set. It can be understood that based on the statistical number of the second cluster, the obstacle points that may be noise points in the second obstacle point cloud are preliminarily screened, which provides basic data for subsequent re-screening.
[0086] In one embodiment, determining the first candidate noise point set corresponding to the second candidate noise point set based on the straight-line distance between the centroid of the first candidate noise point set and the centroid of the second candidate noise point set includes:
[0087] For each second candidate noise point set, determine a first candidate noise point set that matches the second candidate noise point set; determine the straight-line distance between the centroid of the second candidate noise point set and the centroid of the first candidate noise point set that matches the second candidate noise point set; sum the straight-line distances corresponding to the second candidate noise point sets to obtain the distance cost; when the distance cost is minimized, determine the first candidate noise point set that matches the second candidate noise point set as the first candidate noise point set corresponding to the second candidate noise point set.
[0088] The first candidate noise point set matching the second candidate noise point set refers to the first candidate noise point set whose straight-line distance is determined with the second candidate noise point set, and the first candidate noise point set matching the second candidate noise point set may be any first candidate noise point set in at least one first candidate noise point set. The distance cost refers to the sum of the straight-line distances corresponding to all second candidate noise point sets.
[0089] Exemplarily, for each second candidate noise point set, the robot determines a first candidate noise point set that matches the second candidate noise point set, calculates the straight-line distance between the centroid of the second candidate noise point set and the centroid of the matched first candidate noise point set, sums the straight-line distances corresponding to all second candidate noise point sets to obtain a distance cost; repeats the above steps to obtain multiple distance costs; compares the multiple distance costs to obtain the minimum distance cost; in the case of the minimum distance cost, determines the first candidate noise point set that matches the second candidate noise point set as the first candidate noise point set corresponding to the second candidate noise point set, and obtains the first candidate noise point set corresponding to each second candidate noise point set.
[0090] For example, there are three second candidate noise point sets B1, B2 and B3 and three first candidate noise point sets A1, A2 and A3. There are the following six possible matching combinations between the three second candidate noise point sets and the three first candidate noise point sets. B1 matches A2, B2 matches A3, B3 matches A1, the straight-line distance between B1 and A2 is S12, the straight-line distance between B2 and A3 is S23, and the straight-line distance between B3 and A1 is S31. Then the distance cost S1=S12+S23+S31; B1 matches A2, B2 matches A1, B3 matches A3, the straight-line distance between B1 and A2 is S12, the straight-line distance between B2 and A1 is S21, and the straight-line distance between B3 and A3 is S31. 3, then the distance cost S2=S12+S21+S33; B1 matches A3, B2 matches A1, B3 matches A2, the straight-line distance between B1 and A3 is S13, the straight-line distance between B2 and A1 is S21, and the straight-line distance between B3 and A2 is S32, then the distance cost S3=S13+S21+S32; B1 matches A3, B2 matches A2, B3 matches A1, the straight-line distance between B1 and A3 is S13, the straight-line distance between B2 and A2 is S21, and the straight-line distance between B3 and A2 is S32. The straight-line distance between B1 and A1 is S22, and the straight-line distance between B3 and A1 is S31, then the distance cost S4=S13+S22+S31; B1 matches A1, B2 matches A2, B3 matches A3, the straight-line distance between B1 and A1 is S11, the straight-line distance between B2 and A2 is S22, and the straight-line distance between B3 and A3 is S33, then the distance cost S5=S11+S22+S33; B1 matches A1, B2 matches A3, B3 matches A2, and ... The straight-line distance between B1 and A2 is S11, the straight-line distance between B2 and A3 is S23, and the straight-line distance between B3 and A2 is S32, then the distance cost S6=S11+S23+S32; compare the distance costs S1, S2, S3, S4, S5 and S6 of the above 6 matching combinations. If S1 is the smallest distance cost, it is finally determined that the first candidate noise point set corresponding to B1 is A2, the first candidate noise point set corresponding to B2 is A3, and the first candidate noise point set corresponding to B3 is A1.
[0091] If the number of the second candidate noise point set is different from the number of the first candidate noise point set, the number of the second candidate noise point set is made the same as the number of the first candidate noise point set by adding an auxiliary candidate noise point set, the number of auxiliary candidate noise points in the auxiliary candidate noise point set is less than the number threshold, and the auxiliary candidate noise points are far away from the first candidate noise and the second candidate noise.
[0092] In this embodiment, by determining the first candidate noise point set that matches the second candidate noise point set as the first candidate noise point set corresponding to the second candidate noise point set when the distance cost is minimized, it can be understood that the minimum distance cost means that the straight-line distance between all second candidate noise point sets and the matched first candidate noise point sets is as minimum as possible. When the distance cost is minimized, the first candidate noise point set that matches the second candidate noise point set is determined as the first candidate noise point set corresponding to the second candidate noise set. The second candidate noise set and the corresponding first candidate noise set respectively represent the same object detected by the robot at different times, that is, the first candidate noise set represents the object detected by the robot at the first time, and the second candidate noise set represents the object detected by the robot at the second time, which provides basic data for candidate determination of target noise points.
[0093] In one embodiment, after determining the obstacle point in the second candidate noise point set corresponding to the offset distance as the target noise point, the method further includes:
[0094] The target noise point is removed from the second obstacle point cloud to obtain the target obstacle point cloud corresponding to the second moment.
[0095] The target obstacle point cloud refers to an obstacle point cloud obtained by filtering the second obstacle point cloud, that is, the obstacle point cloud obtained by removing target noise points in the second obstacle point cloud.
[0096] Exemplarily, after determining the target noise point, the target noise point is removed from the second obstacle point cloud to obtain the target obstacle point cloud corresponding to the second moment.
[0097] In one embodiment, the target obstacle point cloud determination method includes: acquiring an obstacle point cloud corresponding to a second moment to obtain a second obstacle point cloud; clustering the second obstacle point cloud to obtain a second cluster cluster; counting obstacle points in the second cluster cluster to obtain a second statistical quantity; comparing the second statistical quantity with a quantity threshold, and when the second statistical quantity is equal to or greater than the quantity threshold, determining the second cluster cluster corresponding to the second statistical quantity as an obstacle set; when the second statistical quantity is less than the quantity threshold, determining the second cluster cluster corresponding to the second statistical quantity as a second candidate noise point set; and after step 206, also including: when the offset distance is equal to or less than the distance threshold, determining the second candidate noise point set corresponding to the offset distance as an obstacle set; based on the above obstacle set, obtaining the target obstacle point cloud corresponding to the second moment.
[0098] It can be understood that if the statistical number of obstacle points in the second cluster is equal to or greater than the number threshold, it means that the obstacle points in the second cluster are non-isolated points, and the obstacle points in the second cluster are real obstacle points, not noise points. Therefore, the second cluster whose second statistical number is equal to or greater than the number threshold is determined as the obstacle set, and the second cluster whose second statistical number is less than the number threshold is determined as the second candidate noise point set; based on the straight-line distance between the centroid of the first candidate noise point set and the centroid of the second candidate noise point set, the first candidate noise point set corresponding to the second candidate noise point set is determined, and the second candidate noise set and the corresponding first candidate noise set are respectively represented by The first candidate noise point set represents the same object detected by the robot at different times, that is, the first candidate noise point set represents the object detected by the robot at the first time, and the second candidate noise point set represents the object detected by the robot at the second time. If the offset distance between the second candidate noise point set and the corresponding first candidate noise point set is equal to or less than the distance threshold, it indicates that the appearance of the object is fixed and is a real obstacle. The reason why the obstacle points in the second candidate noise point set are isolated points is that the size of the object is small, that is, the obstacle points in the second candidate noise point set are not noise points. Therefore, the second candidate noise point set with an offset distance equal to or less than the distance threshold is determined as the obstacle set. Based on the obstacle set, the target obstacle point cloud corresponding to the second time is obtained, and the obstacle points corresponding to the obstacles with smaller sizes are retained, thereby improving the accuracy of the target obstacle point cloud.
[0099] In this embodiment, the target obstacle point cloud corresponding to the second moment is obtained by removing the target noise points from the second obstacle point cloud. The target obstacle point cloud retains obstacle points corresponding to obstacles of smaller size, thereby improving the accuracy of the target obstacle point cloud.
[0100] In an exemplary embodiment, the noise determination process is as follows: Figure 6 As shown, including:
[0101] The robot obtains the obstacle point cloud corresponding to the first moment to obtain the first obstacle point cloud, performs Euclidean clustering on the first obstacle point cloud to obtain a first cluster cluster; for each first cluster cluster, counts the obstacle points in the first cluster cluster to obtain a first statistical quantity, compares the first statistical quantity with a quantity threshold, and if the first statistical quantity is less than the quantity threshold, determines the first cluster cluster corresponding to the first statistical quantity as the first initial noise point set.
[0102] The robot obtains the posture corresponding to the first moment, and obtains the first posture, and the posture corresponding to the second moment, and obtains the second posture. For each obstacle point in the first initial noise point set, the initial position of the obstacle point is obtained from the first initial noise point set. Formula (1) is used to calculate the initial position and the first posture corresponding to the obstacle point in the first initial noise point set to obtain the middle position corresponding to the obstacle point in the first initial noise point set. Formula (2) is used to calculate the middle position and the second posture corresponding to the obstacle point in the first initial noise point set to obtain the target position corresponding to the obstacle point in the first initial noise point set. Based on the target position corresponding to the obstacle point in the first initial noise point set, the first candidate noise point set is obtained.
[0103] The robot obtains the obstacle point corresponding to the second moment to obtain the second obstacle point cloud, performs Euclidean clustering on the second obstacle point cloud to obtain a second cluster cluster; for each second cluster cluster, counts the obstacle points in the second cluster cluster to obtain a second statistical quantity, compares the second statistical quantity with a quantity threshold, and if the second statistical quantity is less than the quantity threshold, determines the second cluster cluster corresponding to the second statistical quantity as the second candidate noise point set.
[0104] The robot uses the Hungarian matching algorithm to match the first candidate noise point set and the second candidate noise point set to obtain the first candidate noise point set corresponding to each second candidate noise point set; based on the position of the obstacle point in the first candidate noise point set, the centroid of the first candidate noise point set is determined, and based on the position of the obstacle point in the second candidate noise point set, the centroid of the second candidate noise point set is determined; for each second candidate noise point set, the straight-line distance between the centroid of the second candidate noise point set and the centroid of the first candidate noise point set corresponding to the second candidate noise point set is calculated to obtain the offset distance between the second candidate noise point set and the first candidate noise point set corresponding to the second candidate noise point set; the offset distance corresponding to the second candidate noise point set is compared with the distance threshold, and if the offset distance is greater than the distance threshold, the obstacle point in the above second candidate noise point set is determined as the target noise point.
[0105] The robot removes the target noise point from the second obstacle point cloud to obtain the target obstacle point cloud corresponding to the second moment.
[0106] The above-mentioned noise point determination method determines at least one first candidate noise point set through the first obstacle point cloud collected at the first moment, and determines at least one second candidate noise point set through the second obstacle point cloud collected at the second moment. The number of obstacle points in the first candidate noise point set is less than the number threshold, indicating that the obstacle points in the first candidate noise point set are isolated points, which may be noise points formed by the sensor or real obstacle points corresponding to obstacles of smaller sizes; similarly, the obstacle points in the second candidate noise point set may also be noise points formed by the sensor or real obstacle points corresponding to obstacles of smaller sizes, thereby achieving preliminary screening of noise points. On this basis, based on the straight-line distance between the centroid of the first candidate noise point set and the centroid of the second candidate noise point set, the first candidate noise point set corresponding to the second candidate noise point set is determined. The second candidate noise point set and the corresponding first candidate noise point set respectively represent the same object detected by the robot at different times, that is, the first candidate noise point set represents the object detected by the robot at the first moment, and the second candidate noise point set represents the object detected by the robot at the second moment. If the offset distance between the second candidate noise point set and the corresponding first candidate noise point set is greater than the distance threshold, it indicates that the appearance of the object is random. In view of the randomness of the appearance of noise points, it means that the object is not a real obstacle, that is, the obstacle points in the second candidate noise point set are noise points; if the offset distance between the second candidate noise point set and the corresponding first candidate noise point set is equal to or less than the distance threshold, it indicates that the appearance of the object is fixed and is a real obstacle. The reason why the obstacle points in the second candidate noise point set are isolated points is that the size of the object is small, that is, the obstacle points in the second candidate noise point set are not noise points, and the noise points are screened again. Through the above two screenings, it is avoided to directly determine isolated points as noise points, that is, it is avoided to determine obstacle points corresponding to obstacles of smaller sizes as noise points, thereby improving the accuracy of target noise point determination.
[0107] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.
[0108] Based on the same inventive concept, the embodiment of the present application also provides a noise point determination device for implementing the noise point determination method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in the one or more noise point determination device embodiments provided below can refer to the limitations of the noise point determination method above, and will not be repeated here.
[0109] In one embodiment, Figure 7 As shown, a noise point determination device is provided, including: an acquisition module 702, a matching module 704, a calculation module 706 and a determination module 708, wherein:
[0110] An acquisition module 702 is used to acquire at least one first candidate noise point set corresponding to a first moment, and at least one second candidate noise point set corresponding to a second moment; the first moment is before the second moment; the number of obstacle points in the first candidate noise point set is less than a number threshold, and the number of obstacle points in the second candidate noise point set is less than a number threshold;
[0111] A matching module 704 is used to determine the first candidate noise point set corresponding to the second candidate noise point set based on the straight-line distance between the centroid of the first candidate noise point set and the centroid of the second candidate noise point set;
[0112] A calculation module 706 is used to determine an offset distance between the centroid of the second candidate noise point set and the centroid of the first candidate noise point set corresponding to the second candidate noise point set;
[0113] The determination module 708 is used to determine the obstacle point in the second candidate noise point set corresponding to the offset distance as the target noise point when the offset distance is greater than the distance threshold.
[0114] In one embodiment, the acquisition module 702 is also used to: acquire the obstacle point cloud corresponding to the first moment to obtain a first obstacle point cloud; cluster the first obstacle point cloud to obtain a first cluster; count the obstacle points in the first cluster to obtain a first statistical quantity; compare the first statistical quantity with a quantity threshold, and when the first statistical quantity is less than the quantity threshold, determine the first cluster corresponding to the first statistical quantity as a first initial noise point set; based on the first initial noise point set, obtain a first candidate noise point set corresponding to the first initial noise point set.
[0115] In one embodiment, the acquisition module 702 is also used to: acquire the posture corresponding to the first moment to obtain the first posture, and acquire the posture corresponding to the second moment to obtain the second posture; for the initial position of the obstacle point in the first initial noise point set, based on the first posture, convert the initial position into the corresponding intermediate position in the world coordinate system; based on the second posture, convert the intermediate position into the target position; the target position is located in the robot coordinate system with the second posture as the origin; based on the target position corresponding to the obstacle point in the first initial noise point set, obtain the first candidate noise point set corresponding to the first initial noise point set.
[0116] In one embodiment, the acquisition module 702 is also used to: acquire the obstacle point cloud corresponding to the second moment to obtain a second obstacle point cloud; cluster the second obstacle point cloud to obtain a second cluster cluster; count the obstacle points in the second cluster cluster to obtain a second statistical quantity; compare the second statistical quantity with a quantity threshold, and when the second statistical quantity is less than the quantity threshold, determine the second cluster cluster corresponding to the second statistical quantity as the second candidate noise point set.
[0117] In one embodiment, the matching module 704 is also used to: determine, for each second candidate noise point set, a first candidate noise point set that matches the second candidate noise point set; determine the straight-line distance between the centroid of the second candidate noise point set and the centroid of the first candidate noise point set that matches the second candidate noise point set; sum the straight-line distances corresponding to the second candidate noise point sets to obtain the distance cost; and when the distance cost is minimized, determine the first candidate noise point set that matches the second candidate noise point set as the first candidate noise point set corresponding to the second candidate noise point set.
[0118] In one embodiment, the noise point determination device further includes a filtering module, and the filtering module is used to remove the target noise point from the second obstacle point cloud to obtain the target obstacle point cloud corresponding to the second moment.
[0119] Each module in the above noise point determination device can be implemented in whole or in part by software, hardware, or a combination thereof. Each module can be embedded in or independent of the processor in the robot in hardware form, or can be stored in the memory in the robot in software form, so that the processor can call and execute the operations corresponding to each module.
[0120] In one embodiment, a robot is provided. The robot may be a terminal, and its internal structure diagram may be as follows: Figure 8As shown. The robot includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. The processor of the robot is used to provide computing and control capabilities. The memory of the robot includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the robot is used to exchange information between the processor and an external device. The communication interface of the robot is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a noise determination method is implemented. The display unit of the robot is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the robot can be a touch layer covering the display screen, or a button, trackball or touchpad set on the robot shell, or an external keyboard, touchpad or mouse.
[0121] Those skilled in the art will understand that Figure 8 The structure shown in the figure is merely a block diagram of a partial structure related to the scheme of the present application, and does not constitute a limitation on the robot to which the scheme of the present application is applied. The specific robot may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.
[0122] In one embodiment, a robot is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.
[0123] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.
[0124] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.
[0125] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.
[0126] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.
[0127] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0128] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.
Claims
1. A noise point determination method, characterized in that: The method is used for a robot, comprising: Acquire at least one first candidate noise point set corresponding to a first moment, and at least one second candidate noise point set corresponding to a second moment; the first moment is before the second moment; the number of obstacle points in the first candidate noise point set is less than a quantity threshold, and the number of obstacle points in the second candidate noise point set is less than the quantity threshold; Determine, based on a straight-line distance between a centroid of the first candidate noise point set and a centroid of the second candidate noise point set, a first candidate noise point set corresponding to the second candidate noise point set; Determine an offset distance between a centroid of the second candidate noise point set and a centroid of the first candidate noise point set corresponding to the second candidate noise point set; When the offset distance is greater than the distance threshold, the obstacle point in the second candidate noise point set corresponding to the offset distance is determined as the target noise point.
2. The method according to claim 1, characterized in that: The obtaining of at least one first candidate noise point set corresponding to the first moment includes: Obtaining an obstacle point cloud corresponding to the first moment to obtain a first obstacle point cloud; Clustering the first obstacle point cloud to obtain a first cluster; Counting the obstacle points in the first cluster to obtain a first statistical quantity; Comparing the first statistical quantity with the quantity threshold, and when the first statistical quantity is less than the quantity threshold, determining a first cluster corresponding to the first statistical quantity as a first initial noise point set; Based on the first initial noise point set, a first candidate noise point set corresponding to the first initial noise point set is obtained.
3. The method according to claim 2, characterized in that The obtaining, based on the first initial noise point set, a first candidate noise point set corresponding to the first initial noise point set includes: Obtaining the posture corresponding to the first moment to obtain a first posture, and obtaining the posture corresponding to the second moment to obtain a second posture; For an initial position of an obstacle point in the first initial noise point set, converting the initial position into a corresponding intermediate position in a world coordinate system based on the first posture; Based on the second posture, converting the intermediate position into a target position; the target position is located in a robot coordinate system with the second posture as the origin; Based on the target position corresponding to the obstacle point in the first initial noise point set, a first candidate noise point set corresponding to the first initial noise point set is obtained.
4. The method according to claim 1, characterized in that The obtaining of at least one second candidate noise point set corresponding to the second moment includes: Obtaining an obstacle point cloud corresponding to the second moment to obtain a second obstacle point cloud; Clustering the second obstacle point cloud to obtain a second cluster; Counting the obstacle points in the second cluster to obtain a second statistical quantity; The second statistical quantity is compared with the quantity threshold, and when the second statistical quantity is smaller than the quantity threshold, a second cluster corresponding to the second statistical quantity is determined as a second candidate noise point set.
5. The method according to claim 1, characterized in that The determining, based on the straight-line distance between the centroid of the first candidate noise point set and the centroid of the second candidate noise point set, the first candidate noise point set corresponding to the second candidate noise point set includes: For each of the second candidate noise point sets, determining a first candidate noise point set that matches the second candidate noise point set; Determine a straight-line distance between a centroid of the second candidate noise point set and a centroid of a first candidate noise point set that matches the second candidate noise point set; Sum the straight-line distances corresponding to the second candidate noise point set to obtain a distance cost; When the distance cost is the minimum, the first candidate noise point set matching the second candidate noise point set is determined as the first candidate noise point set corresponding to the second candidate noise point set.
6. The method according to claim 1, characterized in that After determining the obstacle point in the second candidate noise point set corresponding to the offset distance as the target noise point, the method further includes: The target noise point is removed from the second obstacle point cloud to obtain a target obstacle point cloud corresponding to the second moment.
7. A noise point determination device, characterized in that: The device is used for a robot and comprises: an acquisition module, configured to acquire at least one first candidate noise point set corresponding to a first moment, and at least one second candidate noise point set corresponding to a second moment; the first moment is before the second moment; the number of obstacle points in the first candidate noise point set is less than a quantity threshold, and the number of obstacle points in the second candidate noise point set is less than the quantity threshold; a matching module, configured to determine a first candidate noise point set corresponding to the second candidate noise point set based on a straight-line distance between a centroid of the first candidate noise point set and a centroid of the second candidate noise point set; a calculation module, configured to determine an offset distance between a centroid of the second candidate noise point set and a centroid of the first candidate noise point set corresponding to the second candidate noise point set; The determination module is used to determine the obstacle point in the second candidate noise point set corresponding to the offset distance as the target noise point when the offset distance is greater than the distance threshold.
8. A robot comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.
10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.
