CN111723866A - Point cloud clustering method and device, unmanned vehicle and readable storage medium - Google Patents
Point cloud clustering method and device, unmanned vehicle and readable storage medium Download PDFInfo
- Publication number
- CN111723866A CN111723866A CN202010570828.6A CN202010570828A CN111723866A CN 111723866 A CN111723866 A CN 111723866A CN 202010570828 A CN202010570828 A CN 202010570828A CN 111723866 A CN111723866 A CN 111723866A
- Authority
- CN
- China
- Prior art keywords
- points
- core
- distance
- point
- clustering
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Optical Radar Systems And Details Thereof (AREA)
- Traffic Control Systems (AREA)
Abstract
The application provides a point cloud clustering method and device, an unmanned vehicle and a readable storage medium. The method for point cloud clustering comprises the following steps: acquiring point cloud data of a laser radar; the point cloud data comprises a plurality of points which are sequentially arranged according to the rotation sequence of the laser radar; calculating the intervals among the plurality of points, and screening out core points from the plurality of points according to the intervals among the plurality of points; and sequentially clustering the plurality of core points according to the distance between two adjacent core points based on the rotation sequence. When the method is applied to obstacle detection of an unmanned vehicle which is unmanned (namely, automatically driven), the calculation amount of point cloud clustering can be reduced, and the efficiency of point cloud clustering is further improved.
Description
Technical Field
The application relates to the technical field of laser radar positioning, in particular to a point cloud clustering method and device, an unmanned vehicle and a readable storage medium.
Background
The laser radar perception technology mainly comprises main links of segmentation/detection, tracking, identification, reconstruction and the like. The segmentation and detection processes are usually performed by a Clustering algorithm, including a Density Clustering algorithm (DBSCAN) of application with Noise (Density-Based Spatial Clustering algorithm).
However, the DBSCAN algorithm is suitable for disordered points, and the algorithm needs to traverse all the points for many times, so that the calculation amount is large, and the efficiency of point cloud clustering is low.
Disclosure of Invention
The embodiment of the application aims to provide a point cloud clustering method and device, an unmanned vehicle and a readable storage medium, which are used for reducing the calculated amount of point cloud clustering and further improving the efficiency of point cloud clustering.
In a first aspect, an embodiment of the present application provides a method for point cloud clustering, including: acquiring point cloud data of the laser radar, wherein the point cloud data comprises a plurality of points which are sequentially arranged according to the rotation sequence of the laser radar; calculating the intervals among the plurality of points, and screening out core points from the plurality of points according to the intervals among the plurality of points; and sequentially clustering the plurality of core points according to the distance between two adjacent core points based on the rotation sequence.
In the embodiment of the application, the acquired point cloud data are a plurality of points which are sequentially arranged according to the rotation sequence of the laser radar, that is, the points are ordered points; when clustering is carried out, core points are screened out firstly, and then only the core points need to be clustered sequentially. Compared with the prior art, the points are ordered, and after the core points are screened out, the core points are also ordered, the clustering is performed without calculating the distances between all the core points, and the clustering is performed only according to the distances between two adjacent core points in sequence, so that the defect that the conventional DBSCAN algorithm needs to traverse all the points for many times and has large calculation amount is overcome, and the point cloud clustering efficiency is improved.
As a possible implementation manner, the point coordinates are polar coordinates, and the calculating the distance between the plurality of points includes: when the distance between two points is calculated, the distance between the two points is calculated according to the polar diameters of the two points and the polar angle difference between the two points, and the distance calculation satisfies the following formula:
wherein s is1-2Is the distance between the two points, r1、r2Δ θ is the polar diameter of the two points and the polar angle difference between the two points.
In the embodiment of the application, the point cloud data of the laser radar is polar coordinate points, and compared with the DBSCAN algorithm (which needs to convert polar coordinates into XY coordinates first and then calculate) adopted in the prior art, the distance between two points is directly calculated according to the polar diameter of the two points and the polar angle difference between the two points, so that the conversion of point coordinates is avoided, a large amount of calculation is saved, and the efficiency of point cloud clustering is improved.
As a possible implementation manner, the screening the core points from the plurality of points according to the intervals between the plurality of points includes: aiming at each point, determining the number of adjacent points of which the distance from the point is smaller than a first preset distance value according to the distance between the plurality of points; and screening out points with the number of the adjacent points larger than a preset value from the plurality of points, and determining the points as the core points.
In the embodiment of the application, when the core point is screened, the number of adjacent points which are provided by the point and have a distance with the point smaller than a first preset distance value is determined according to the distance between the points for each point, and then the points with the number of the adjacent points larger than a preset value are screened from the points, so that the core point is rapidly screened, and the point cloud clustering efficiency is improved.
As a possible implementation manner, the sequentially clustering the plurality of core points according to the distance between two adjacent core points includes: if the distance between two adjacent core points is smaller than a second preset distance value, the two adjacent core points are gathered into one type; and if the distance between two adjacent core points is not smaller than the second preset distance value, taking the core point with the next arrangement sequence as the initial core point of the next class.
In this embodiment of the application, when clustering core points is performed, if a distance between two adjacent core points is smaller than a second preset distance value, it is described that the two adjacent core points can be classified into one class, and if the distance between two adjacent core points is not smaller than the second preset distance value, new clustering can be started with the core points after the core points in the arrangement sequence as a start point, so that fast clustering of the core points is realized, and the efficiency of point cloud clustering is improved.
As a possible implementation manner, the clustering the plurality of core points according to the distance between two adjacent core points includes: when the last core point clustering is completed, judging whether the distance between the last core point and the core points in the first clustering is smaller than a second preset distance value or not; and if the distance between the last core point and the core points in the first cluster is smaller than the second preset distance value, the last cluster and the first cluster are classified into one cluster.
In the embodiment of the application, when the clustering of the last core point is completed, considering the characteristic of the rotation sequence of the laser radar, there may be a situation that the distance between the last core point and the core point in the first cluster also meets the requirement of the second distance value, and if the situation is met, the last cluster and the first cluster can be merged, so that the rationality of the point cloud clustering is improved.
In a second aspect, an apparatus for point cloud clustering according to an embodiment of the present application includes a functional module configured to implement the method for point cloud clustering described in the first aspect and any one of the possible implementation manners of the first aspect.
In a third aspect, an embodiment of the present application provides an unmanned vehicle, including: a vehicle body; a detection device disposed within the vehicle body; the detection apparatus includes: a receiver, a memory, and a processor; the memory is to store computer instructions; the processor is configured to execute the computer instructions stored in the memory to implement the point cloud clustering method in the first aspect and any implementation manner of the first aspect.
In a fourth aspect, this application provides a readable storage medium, on which a computer program is stored, where the computer program is executed by a computer to perform the method described in the first aspect and any one of the implementation manners of the first aspect.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are required to be used in the embodiments of the present application will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered as limiting the scope, and that those skilled in the art can also obtain other related drawings based on the drawings without inventive efforts.
Fig. 1 is a flowchart of a point cloud clustering method according to an embodiment of the present disclosure;
FIG. 2 is a schematic diagram illustrating a principle of pitch calculation according to an embodiment of the present disclosure;
fig. 3 is a schematic diagram of a point cloud clustering effect provided in the embodiment of the present application;
fig. 4 is a functional structure block diagram of a point cloud clustering device provided in the embodiment of the present application;
fig. 5 is a schematic structural diagram of an unmanned vehicle according to an embodiment of the present application;
fig. 6 is a schematic structural diagram of a detection apparatus provided in an embodiment of the present application.
Icon: 200-means for point cloud clustering; 201-an acquisition module; 202-a screening module; 203-clustering module; 30-unmanned vehicle; 31-a vehicle body; 32-a detection device; 320-a receiver; 322-a memory; 324-a processor.
Detailed Description
The technical solutions in the embodiments of the present application will be described below with reference to the drawings in the embodiments of the present application.
The lidar is a sensor capable of detecting obstacles by sensing surrounding environment, and is widely used in the fields of unmanned vehicles (automatic driving automobiles), robots and the like. When the laser radar senses the surrounding environment, the method comprises the main processes of segmentation/detection, tracking, identification, reconstruction and the like, wherein a clustering algorithm is commonly used in the segmentation and detection processes. Clustering is a process of performing aggregation classification on each point detected by a laser radar, and points belonging to the same cluster can be understood as points detected on the same object. Therefore, the clustering of the points is important for the environmental perception of the laser radar, and can directly influence the final positioning result of the obstacle. The technical scheme provided by the embodiment of the application is a point cloud clustering method, and the method can be applied to various scenes in which point cloud data of a laser radar needs to be processed, such as the detection of obstacles in the process of advancing of an unmanned vehicle; the method is applied to the detection of obstacles in the process of the robot; under these scenes, after the obstacle is accurately and quickly detected, the unmanned vehicle and the robot can plan the traveling path as soon as possible according to the obstacle information. Therefore, the method can be applied to a detection system (detection apparatus) of an obstacle of an unmanned vehicle or a robot.
In addition, in the technical scheme provided by the embodiment of the application, the clustering of ordered point cloud data is aimed at, that is, the ordered point cloud data is aimed at point cloud data of the laser radar, so the laser radar in the technical scheme is a common rotary laser radar (with a fixed rotation direction), the detection principle of the rotary laser radar is that the distance of an obstacle when the laser radar reaches different angles is detected by using a Time of flight (TOF) technology when the laser radar rotates, and finally, the direction (angle) and the distance data of the obstacle are acquired. In the 3D imaging, TOF is interpreted as meaning time-of-flight, and a target object distance is obtained by continuously transmitting light pulses to a target, receiving light returned from an object by a laser radar, and detecting the time of flight (round trip) of the light pulses.
Based on the application scenario, referring to fig. 1, a flowchart of a point cloud clustering method provided in the embodiment of the present application is shown, where the method includes:
step 101: and acquiring point cloud data of the laser radar. The point cloud data comprises a plurality of points which are sequentially arranged according to the rotation sequence of the laser radar.
Step 102: calculating the intervals among the multiple points, and screening out core points from the multiple points according to the intervals among the multiple points;
step 103: clustering the plurality of core points according to the distance between two adjacent core points based on the rotation order.
In the embodiment of the application, the acquired point cloud data are a plurality of points which are sequentially arranged according to the rotation sequence of the laser radar, that is, the points are ordered points; when clustering is carried out, core points are screened out firstly, and then only the core points need to be clustered sequentially. Compared with the prior art, the points are ordered, and after the core points are screened out, the core points are also ordered, the clustering is performed without calculating the distances between all the core points, and the clustering is performed only according to the distances between two adjacent core points in sequence, so that the defect that the conventional DBSCAN algorithm needs to traverse all the points for many times and has large calculation amount is overcome, and the point cloud clustering efficiency is improved.
A detailed implementation of steps 101-103 is described next.
In step 101, the point cloud data includes a plurality of points sequentially arranged according to a rotation order of the lidar, and the rotation order of the lidar may be clockwise or counterclockwise. It is understood that in the point cloud data acquired by the lidar, each point is represented by coordinates. Such as: if the laser radar detects 100 points in total, the coordinates of the 100 points are sequentially (in a rotation order) in the point cloud data of the laser radar.
It is noted that for rotational lidar, the resulting coordinate data is also three-dimensional, i.e. polar, coordinates based on time-of-flight 3D imaging. For polar coordinates, a fixed point O, called a pole, is taken in a plane, a ray Ox, called a polar axis, and a length unit and the positive direction of an angle (usually, the counterclockwise direction) are selected. For any point (assumed to be M) in the plane, the length of a line segment OM (which can also be expressed by r) is represented by rho, the angle from Ox to OM is represented by theta, rho is the polar diameter of the point M, and theta is the polar angle of the point M, and the ordered number pair (rho, theta) is called the polar coordinate of the point M. In the description of the subsequent embodiments, the polar coordinates of the points are represented by (r, θ); the fixed point in the plane may be the position where the lidar is located, i.e. a polar coordinate system with the rotating lidar as the fixed point.
Further, after the point cloud data of the laser radar is acquired, step 102 is executed to calculate the distance between the plurality of points, and the core point is screened out from the plurality of points according to the distance between the plurality of points. When calculating the distance between a plurality of points, since the plurality of points are represented by polar coordinates, the embodiment of the present application provides an alternative way of calculating the distance: in calculating the distance between two points, the polar diameter of two points and the polar angle difference between two points are used.
In this implementationIn this manner, the distance between two points can be calculated using the polar diameters of the two points and the polar angle between the two points. For easy understanding, please refer to fig. 2, which is a schematic diagram of the calculation principle of this calculation method, in fig. 2, point 1 and point 2 are included, the polar coordinate fixed point is O, and it is assumed that the polar coordinate of point 1 is (r)1,θ1) The coordinate of point 2 is (r)2,θ2) Then the spacing s between point 1 and point 21-2Can pass through
To obtain an approximation, wherein r1Is the pole diameter of point 1, r2Is the pole diameter of point 2, and Δ θ is the difference in pole angle between point 1 and point 2, i.e., Δ θ is θ1-θ2。
In the embodiment of the application, the point cloud data of the laser radar is polar coordinate points, and compared with the DBSCAN algorithm (which needs to convert polar coordinates into XY coordinates first and then calculate) adopted in the prior art, the distance between two points is directly calculated according to the polar diameter of the two points and the polar angle difference between the two points, so that the conversion of point coordinates is avoided, a large amount of calculation is saved, and the efficiency of point cloud clustering is improved.
It should be noted that, when calculating the distance between multiple points, all the points need to be traversed, and the distance between each point and other points is calculated in turn, such as: assuming a total of 100 points, the first point needs to be spaced from 99 points later, the second point needs to be spaced from 98 points later because the distance from the first point has already been calculated, the third point needs to be spaced from 97 points later because the distance from the first point has already been calculated, and so on.
Further, after calculating the distances between the points, the core points are screened out from the points according to the distances between the points, and for the core points, referring to the definition in the DBSCAN algorithm, the cluster definition is simple: the sample set connected with the maximum density derived from the density reachable relation is a category or a cluster of the final cluster. One or more core objects may be inside a cluster of a DBSCAN. If there is only one core object, the samples of other non-core objects in the cluster are all in the neighborhood of the core object, if there are multiple core objects, there must be one other core object in the neighborhood of any one core object in the cluster, otherwise, the two core objects cannot be reached in density. Therefore, the core points are screened out firstly, and then clustering is carried out based on the core points, so that point cloud clustering with accessible density can be realized.
As an optional implementation, the process of screening the core points includes: aiming at each point, determining the number of adjacent points of which the distance from the point is smaller than a first preset distance value according to the distance between the points; and screening out points with the number of adjacent points larger than a preset value from the plurality of points, and determining the points as core points.
In this embodiment, for each point, the number of neighboring points, which are located at each point and have a distance from the point smaller than a first preset distance value, is determined, and then the points, from which the number of neighboring points is greater than a preset value, are selected and determined as the core point. The core point is equivalent to satisfy two conditions, one is that neighboring points satisfying the first preset distance value condition need to exist around the core point, and the other is that the number of the neighboring points existing around the core point is also larger than a preset value. By these two conditions it can be ensured that the core point is indeed a point on the obstacle, such as: if a point is an isolated point, it may be a point that has been mis-scanned or scanned that is not on an obstacle.
And the first preset distance value and the preset value can be specifically set in advance according to the application scene of the laser radar. Wherein the first preset distance value belongs to an object depending on how close the two points are, for example: for larger obstacles, two points may also belong to the same object when they are far apart; however, for smaller obstacles, two points do not belong to the same object when the distances between the two points are far away, and only belong to the same object when the distances between the two points are near. Therefore, when the first preset distance value is set, the first preset distance value may be set according to the actual situation of an obstacle that may exist in a scene to which the laser radar is applied. The preset value of the number of the adjacent points can be set according to the characteristics of the laser radar. If the performance of the laser radar is poor and stolen points (which can be understood as false detection points) are more, the preset value of the number of the adjacent points can be set higher; if the performance of the laser radar is better and the number of stolen points is less, the preset value of the number of adjacent points can be set to be smaller.
In the embodiment of the application, when the core point is screened, the number of adjacent points which are provided by the point and have a distance with the point smaller than a first preset distance value is determined according to the distance between the points for each point, and then the points with the number of the adjacent points larger than a preset value are screened from the points, so that the core point is rapidly screened, and the point cloud clustering efficiency is improved.
Further, after the core points are screened in step 102, step 103 is executed to cluster the plurality of core points in sequence according to the distance between two adjacent core points based on the rotation sequence.
When the core points are clustered according to the distance between two adjacent core points, if the distance between two adjacent core points is smaller than a second preset distance value, the two adjacent core points are clustered into one class; and if the distance between two adjacent core points is not smaller than the second preset distance value, taking the core point with the next arrangement sequence as the initial core point of the next class. In the whole process, all the adjacent core points are traversed in sequence, and then whether the two adjacent core points can be classified into one class or not is judged according to the second preset distance value. The second preset distance value and the first preset distance value can be set to be the same value or different values; whether the same value or different values are used, the setting modes of the two values may adopt the same mode, that is, the setting mode of the first preset interval value in the foregoing embodiment.
For example, assuming a total of 100 core points, if the distance between the first core point and the second core point is smaller than the second preset distance value, the first core point and the second core point are classified into one class; if the distance between the third core point and the second core point is smaller than the second preset distance value, classifying the third core point into the class of the first core point and the second core point; if the distance between the third core point and the second core point is larger than or equal to the second preset distance value, a new class is created starting from the third core point. And repeating the steps until the clustering of 100 core points is completed, and obtaining a plurality of clusters corresponding to the 100 core points.
In the embodiment of the application, when the core points are clustered, if the distance between two adjacent core points is smaller than a preset distance value, it is indicated that the two adjacent core points can be classified into one class, so that the fast clustering of the core points is realized, and the point cloud clustering efficiency is improved.
Further, after the screening of the core points is performed, the processing of step 103 is not performed for those points that do not belong to the core points.
In addition, if the rotation sequence of the laser radar is counterclockwise or clockwise, there is a case where the last core point and one or more core points in the first cluster also belong to adjacent core points, and in this case, as an optional implementation manner, when the last core point is clustered, it is determined whether the distance between the last core point and the core point in the first cluster is smaller than a second preset distance value; and if the distance between the last core point and the core points in the first cluster is smaller than a second preset distance value, clustering the cluster where the last core point is located and the first cluster into one cluster.
The distance between the last core point and the core point in the first cluster may be a distance between a target core point and the last core point in the first cluster; or may be an average of the distance between the last core point and each core point in the first cluster; or may be other possible implementations. The target core point in the first cluster may be a center core point in the first cluster, where the center core point is a point located between the first core point and the last core point in the first cluster, for example: a total of 5 points in the first cluster, then the 3 rd point in the first cluster is the central core point.
Further, if the distance between the last core point and the core point in the first cluster is smaller than a second preset distance value, the last cluster (i.e. the cluster where the last core point is located) and the first cluster may be grouped into one class; and if the distance between the last core point and the core points in the first cluster is not less than a second preset distance value, the original cluster is kept.
In the embodiment of the application, when the clustering of the last core point is completed, considering the characteristic of the rotation sequence of the laser radar, there may be a situation that the distance between the last core point and the core point in the first cluster also meets the requirement of the second distance value, and if the situation is met, the last cluster and the first cluster can be merged, so that the rationality of the point cloud clustering is improved.
Referring to fig. 3, an exemplary diagram of the effect of point cloud clustering provided in the embodiment of the present application is shown, in fig. 3, two point cloud clusters using the position of the lidar as a reference are obtained, including a left cluster and a right cluster of the position of the lidar, and each cluster includes a plurality of core points.
By the point cloud clustering method provided by the embodiment of the application, when the distance between the points is calculated, calculation is directly carried out based on the polar coordinates, point coordinate conversion is avoided, and a large amount of calculation is saved. When the core points are clustered based on the distance, the clustering range is locked between the sequentially adjacent points by utilizing the orderliness of the point cloud data of the rotary laser radar, and the defect that all the points need to be traversed again in the core point cloud clustering process of the DBSCAN algorithm is avoided. And then can reduce the calculated amount in the clustering process greatly, raise the efficiency of the point cloud clustering.
Further, after the clustering of the points is completed, for the unmanned aerial vehicle or the robot, further processing (such as tracking and positioning) can be performed on the basis of the clustered point cloud data, and finally, the detection of the obstacle is realized.
Based on the same inventive concept, please refer to fig. 4, an embodiment of the present application further provides a point cloud clustering apparatus 200, including: an acquisition module 201, a screening module 202 and a clustering module 203.
The acquisition module 201 is configured to acquire point cloud data of a laser radar; the point cloud data comprises a plurality of points which are sequentially arranged according to the rotation sequence of the laser radar; the screening module 202 is configured to calculate distances between the plurality of points, and screen a core point from the plurality of points according to the distances between the plurality of points; and the clustering module 203 is configured to sequentially cluster the plurality of core points according to the distance between two adjacent core points based on the rotation order.
Optionally, the screening module 202 is specifically configured to: when the distance between two points is calculated, the distance between the two points is calculated according to the polar diameters of the two points and the polar angle difference between the two points, and the distance calculation satisfies the following formula:
wherein s is1-2Is the distance between the two points, r1、r2Δ θ is the polar diameter of the two points and the polar angle difference between the two points.
Optionally, the screening module 202 is specifically configured to: aiming at each point, determining the number of adjacent points of which the distance from the point is smaller than a first preset distance value according to the distance between the plurality of points; and screening out points with the number of the adjacent points larger than a preset value from the plurality of points, and determining the points as the core points.
Optionally, the clustering module 203 is specifically configured to: if the distance between two adjacent core points is smaller than a second preset distance value, the two adjacent core points are gathered into one type; and if the distance between two adjacent core points is not smaller than the second preset distance value, taking the core point with the next arrangement sequence as the initial core point of the next class.
Optionally, the clustering module 203 is further specifically configured to: when the last core point clustering is completed, judging whether the distance between the last core point and the core points in the first clustering is smaller than a second preset distance value or not; and if the distance between the last core point and the core points in the first cluster is smaller than the second preset distance value, the last cluster and the first cluster are classified into one cluster.
The embodiments and specific examples of the method for point cloud clustering in the foregoing embodiments are also applicable to the apparatus 200 for point cloud clustering in fig. 4, and the implementation method of the apparatus 200 for point cloud clustering in fig. 4 can be clearly known by those skilled in the art through the foregoing detailed description of the method for point cloud clustering, so for the brevity of the description, detailed description is omitted here.
Based on the same inventive concept, please refer to fig. 5, the embodiment of the present application further provides an unmanned vehicle 30, which includes a vehicle body 31 and a detection device disposed in the vehicle body 31. Be provided with laser radar on automobile body 31, this laser radar is connected with check out test set, and after laser radar sent point cloud data for check out test set, check out test set can be handled based on point cloud data, realizes the detection of barrier. The laser radar can be arranged at the bottom of the vehicle body 31, of course, also can be arranged at the tail of the vehicle body 31, and the specific arrangement position can be adjusted according to the actual situation.
The detection device may perform processing of point cloud data, including clustering of points, that is, performing a process in the point cloud clustering method in the foregoing embodiment. Referring to fig. 6, which is a schematic structural diagram of the detection device 32 provided in the embodiment of the present application, the detection device 32 includes: a receiver 320, a memory 322, and a processor 324.
The receiver 320, the memory 322, and the processor 324 are electrically connected directly or indirectly to each other to realize data transmission or interaction. For example, electrical connections between these components may be made through one or more communication or signal buses. The method of point cloud clustering comprises at least one software functional module which can be stored in the memory 322 in the form of software or firmware (firmware), for example, a software functional module or a computer program comprised by the apparatus 200 of point cloud clustering, respectively.
The receiver 320 may receive the point cloud data sent by the laser radar and transmit the point cloud data to the processor 324 for processing, or transmit the point cloud data to the memory 322 for storage, so that the processor 324 may call the stored point cloud data for processing at any time.
The memory 322 may store various software programs and modules, such as program instructions/modules corresponding to the method and apparatus for point cloud clustering provided in the embodiments of the present application. The processor 324 executes various functional applications and data processing by executing software programs and modules stored in the memory 322, thereby implementing the methods of the embodiments of the present application.
The Memory 322 may include, but is not limited to, a RAM (Random Access Memory), a ROM (Read Only Memory), a PROM (Programmable Read-Only Memory), an EPROM (Erasable Read-Only Memory), an EEPROM (electrically Erasable Read-Only Memory), and the like.
The processor 324 may be an integrated circuit chip having signal processing capabilities. The Processor 324 may be a general-purpose Processor including a CPU (Central Processing Unit), an NP (Network Processor), etc.; but may also be a digital signal processor, an application specific integrated circuit, an off-the-shelf programmable gate array or other programmable logic device, discrete gate or transistor logic, discrete hardware components. Which may implement or perform the methods, steps, and logic blocks disclosed in the embodiments of the present application. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
It will be appreciated that the configuration shown in fig. 5 and 6 is merely illustrative and that the drone vehicle 30 and detection device 32 may also include more components than shown in fig. 5 and 6, or have a different configuration than shown in fig. 5 and 6. Such as: the unmanned vehicle 30 may further include a display, and the detection device 32 may further include a transmitter, wherein the display is connected to the detection device 32, and when the detection device 32 obtains the detection result of the obstacle, the detection result is transmitted to the display through the transmitter for displaying. The components shown in fig. 5 and 6 may be implemented in hardware, software, or a combination thereof.
Based on the same inventive concept, an embodiment of the present application further provides a readable storage medium, where a computer program is stored on the readable storage medium, and when the computer program is executed by a computer, the method for point cloud clustering according to any of the above embodiments is performed.
In the embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other ways. The above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only one logical division, and there may be other divisions when actually implemented, and for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection of devices or units through some communication interfaces, and may be in an electrical, mechanical or other form.
In addition, units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
Furthermore, the functional modules in the embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.
In this document, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions.
The above description is only an example of the present application and is not intended to limit the scope of the present application, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims (10)
1. A method of point cloud clustering, comprising:
acquiring point cloud data of the laser radar, wherein the point cloud data comprises a plurality of points which are sequentially arranged according to the rotation sequence of the laser radar;
calculating the intervals among the plurality of points, and screening out core points from the plurality of points according to the intervals among the plurality of points;
and sequentially clustering the plurality of core points according to the distance between two adjacent core points based on the rotation sequence.
2. The method of claim 1, wherein the coordinates of the plurality of points are polar coordinates, and wherein calculating the spacing between the plurality of points comprises:
when the distance between two points is calculated, the distance between the two points is calculated according to the polar diameters of the two points and the polar angle difference between the two points, and the distance calculation satisfies the following formula:
3. The method of claim 1, wherein said screening core points from said plurality of points based on a spacing between said plurality of points comprises:
aiming at each point, determining the number of adjacent points of which the distance from the point is smaller than a first preset distance value according to the distance between the plurality of points;
and screening out points with the number of the adjacent points larger than a preset value from the plurality of points, and determining the points as the core points.
4. The method according to claims 1 and 2, wherein said sequentially clustering a plurality of core points according to a pitch between two adjacent core points comprises:
if the distance between two adjacent core points is smaller than a second preset distance value, the two adjacent core points are gathered into one type;
and if the distance between two adjacent core points is not smaller than the second preset distance value, taking the core point with the next arrangement sequence as the initial core point of the next class.
5. The method according to claim 1, wherein said clustering a plurality of core points according to a pitch between two adjacent core points comprises:
when the last core point clustering is completed, judging whether the distance between the last core point and the core points in the first clustering is smaller than a second preset distance value or not;
and if the distance between the last core point and the core points in the first cluster is smaller than a second preset distance value, the last cluster and the first cluster are classified into one cluster.
6. An apparatus for point cloud clustering, comprising:
the acquisition module is used for acquiring point cloud data of the laser radar; the point cloud data comprises a plurality of points which are sequentially arranged according to the rotation sequence of the laser radar;
the screening module is used for calculating the intervals among the plurality of points and screening the core points from the plurality of points according to the intervals among the plurality of points;
and the clustering module is used for sequentially clustering the plurality of core points according to the distance between two adjacent core points based on the rotation sequence.
7. The apparatus of claim 6, wherein the coordinates of the plurality of points are polar coordinates, and wherein the filtering module is specifically configured to:
when the distance between two points is calculated, the distance between the two points is calculated according to the polar diameters of the two points and the polar angle difference between the two points, and the distance calculation satisfies the following formula:
8. The apparatus of claim 6, wherein the clustering module is specifically configured to:
if the distance between two adjacent core points is smaller than a second preset distance value, the two adjacent core points are gathered into one type;
and if the distance between two adjacent core points is not smaller than the second preset distance value, taking the core point with the next arrangement sequence as the initial core point of the next class.
9. An unmanned vehicle, comprising:
a vehicle body; a detection device disposed within the vehicle body;
the detection apparatus includes: a receiver, a memory, and a processor; the memory is to store computer instructions; the processor is configured to execute the computer instructions stored by the memory to implement the method of point cloud clustering of any of claims 1-5.
10. A readable storage medium, having stored thereon a computer program which, when executed by a computer, performs the method of any one of claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010570828.6A CN111723866A (en) | 2020-06-19 | 2020-06-19 | Point cloud clustering method and device, unmanned vehicle and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010570828.6A CN111723866A (en) | 2020-06-19 | 2020-06-19 | Point cloud clustering method and device, unmanned vehicle and readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111723866A true CN111723866A (en) | 2020-09-29 |
Family
ID=72569702
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010570828.6A Pending CN111723866A (en) | 2020-06-19 | 2020-06-19 | Point cloud clustering method and device, unmanned vehicle and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111723866A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2023193567A1 (en) * | 2022-04-08 | 2023-10-12 | 追觅创新科技(苏州)有限公司 | Movement control method and apparatus for robot, and storage medium and electronic apparatus |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107292276A (en) * | 2017-06-28 | 2017-10-24 | 武汉大学 | A kind of vehicle-mounted cloud clustering method and system |
| US20180260636A1 (en) * | 2017-03-13 | 2018-09-13 | Baidu Online Network Technology (Beijing) Co., Ltd. | Obstacle detecting method and apparatus, device and storage medium |
| US20180267166A1 (en) * | 2017-03-16 | 2018-09-20 | Baidu Online Network Technology (Beijing) Co., Ltd. | Obstacle detecting method and apparatus, device and storage medium |
| US20180341019A1 (en) * | 2017-05-26 | 2018-11-29 | Toyota Motor Engineering & Manufacturing North America, Inc. | Publishing lidar cluster data |
| CN109101892A (en) * | 2018-07-17 | 2018-12-28 | 南京航空航天大学 | Laser radar target object detecting method based on grid and density clustering algorithm |
| CN109581312A (en) * | 2018-11-22 | 2019-04-05 | 西安电子科技大学昆山创新研究院 | A kind of high-resolution millimetre-wave radar multi-object clustering method |
| CN110110802A (en) * | 2019-05-14 | 2019-08-09 | 南京林业大学 | Airborne laser point cloud classification method based on high-order condition random field |
| CN110363861A (en) * | 2019-07-14 | 2019-10-22 | 南京林业大学 | Field crop three-dimensional reconstruction method based on laser radar point cloud |
| CN110458055A (en) * | 2019-07-29 | 2019-11-15 | 江苏必得科技股份有限公司 | A kind of obstacle detection method and system |
| CN110647835A (en) * | 2019-09-18 | 2020-01-03 | 合肥中科智驰科技有限公司 | Target detection and classification method and system based on 3D point cloud data |
| CN111192284A (en) * | 2019-12-27 | 2020-05-22 | 吉林大学 | Vehicle-mounted laser point cloud segmentation method and system |
-
2020
- 2020-06-19 CN CN202010570828.6A patent/CN111723866A/en active Pending
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180260636A1 (en) * | 2017-03-13 | 2018-09-13 | Baidu Online Network Technology (Beijing) Co., Ltd. | Obstacle detecting method and apparatus, device and storage medium |
| US20180267166A1 (en) * | 2017-03-16 | 2018-09-20 | Baidu Online Network Technology (Beijing) Co., Ltd. | Obstacle detecting method and apparatus, device and storage medium |
| US20180341019A1 (en) * | 2017-05-26 | 2018-11-29 | Toyota Motor Engineering & Manufacturing North America, Inc. | Publishing lidar cluster data |
| CN107292276A (en) * | 2017-06-28 | 2017-10-24 | 武汉大学 | A kind of vehicle-mounted cloud clustering method and system |
| CN109101892A (en) * | 2018-07-17 | 2018-12-28 | 南京航空航天大学 | Laser radar target object detecting method based on grid and density clustering algorithm |
| CN109581312A (en) * | 2018-11-22 | 2019-04-05 | 西安电子科技大学昆山创新研究院 | A kind of high-resolution millimetre-wave radar multi-object clustering method |
| CN110110802A (en) * | 2019-05-14 | 2019-08-09 | 南京林业大学 | Airborne laser point cloud classification method based on high-order condition random field |
| CN110363861A (en) * | 2019-07-14 | 2019-10-22 | 南京林业大学 | Field crop three-dimensional reconstruction method based on laser radar point cloud |
| CN110458055A (en) * | 2019-07-29 | 2019-11-15 | 江苏必得科技股份有限公司 | A kind of obstacle detection method and system |
| CN110647835A (en) * | 2019-09-18 | 2020-01-03 | 合肥中科智驰科技有限公司 | Target detection and classification method and system based on 3D point cloud data |
| CN111192284A (en) * | 2019-12-27 | 2020-05-22 | 吉林大学 | Vehicle-mounted laser point cloud segmentation method and system |
Non-Patent Citations (1)
| Title |
|---|
| 梁晋等编著: "《数据挖掘原理与应用》", 华中科技大学出版社, pages: 141 - 143 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2023193567A1 (en) * | 2022-04-08 | 2023-10-12 | 追觅创新科技(苏州)有限公司 | Movement control method and apparatus for robot, and storage medium and electronic apparatus |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111222395B (en) | Target detection method and device and electronic equipment | |
| WO2022188663A1 (en) | Target detection method and apparatus | |
| WO2017048383A1 (en) | Systems and methods for non-obstacle area detection | |
| CN109849930B (en) | Method and device for calculating speed of adjacent vehicle of automatic driving automobile | |
| CN111699404B (en) | Driving auxiliary target acquisition method and device, radar, driving system and vehicle | |
| CN113536850B (en) | Target object size testing method and device based on 77G millimeter wave radar | |
| CN113743171A (en) | Target detection method and device | |
| CN112799091A (en) | Algorithm evaluation method, device and storage medium | |
| US11698459B2 (en) | Method and apparatus for determining drivable region information | |
| CN114966591A (en) | Large target detection method, large target detection device and electronic equipment | |
| CN114764885A (en) | Obstacle detection method and device, computer-readable storage medium and processor | |
| CN114627409A (en) | Method and device for detecting abnormal lane change of vehicle | |
| CN111723866A (en) | Point cloud clustering method and device, unmanned vehicle and readable storage medium | |
| CN116824152A (en) | Target detection method and device based on point cloud, readable storage medium and terminal | |
| CN115457506A (en) | Target detection method, device and storage medium | |
| CN115359332A (en) | Data fusion method and device based on vehicle-road cooperation, electronic equipment and system | |
| CN115527187A (en) | Method and device for classifying obstacles | |
| WO2021164514A1 (en) | Target recognition method and apparatus | |
| CN113514825A (en) | Road edge obtaining method and device and terminal equipment | |
| CN113313654A (en) | Laser point cloud filtering and denoising method, system, equipment and storage medium | |
| CN113205059A (en) | Parking space detection method, system, terminal and computer readable storage medium | |
| CN112364888A (en) | Point cloud data processing method and device, computing equipment and computer storage medium | |
| CN113030896B (en) | Radar target clustering method and device and electronic equipment | |
| US20240203107A1 (en) | Obstacle identification method, vehicle-mounted device and storage medium | |
| WO2023178525A1 (en) | Apparatus and method for tracking of extended objects |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |