CN115062683A - 4D millimeter wave radar clustering method based on DBSCAN and storage medium - Google Patents
4D millimeter wave radar clustering method based on DBSCAN and storage medium Download PDFInfo
- Publication number
- CN115062683A CN115062683A CN202210460110.0A CN202210460110A CN115062683A CN 115062683 A CN115062683 A CN 115062683A CN 202210460110 A CN202210460110 A CN 202210460110A CN 115062683 A CN115062683 A CN 115062683A
- Authority
- CN
- China
- Prior art keywords
- point
- dbscan
- millimeter wave
- signal
- 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
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/86—Combinations of radar systems with non-radar systems, e.g. sonar, direction finder
- G01S13/867—Combination of radar systems with cameras
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/88—Radar or analogous systems specially adapted for specific applications
- G01S13/93—Radar or analogous systems specially adapted for specific applications for anti-collision purposes
- G01S13/931—Radar or analogous systems specially adapted for specific applications for anti-collision purposes of land vehicles
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/277—Analysis of motion involving stochastic approaches, e.g. using Kalman filters
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/762—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using clustering, e.g. of similar faces in social networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/80—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level
Landscapes
- Engineering & Computer Science (AREA)
- Remote Sensing (AREA)
- Radar, Positioning & Navigation (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computer Networks & Wireless Communication (AREA)
- Multimedia (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Medical Informatics (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Databases & Information Systems (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Electromagnetism (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention discloses a DBSCAN-based 4D millimeter wave radar clustering method and a storage medium, comprising the following steps of S1, acquiring a vehicle signal, a lane line signal and a point cloud signal of a 4D millimeter wave radar, and analyzing and preprocessing the signals; s2, calculating the local search radius of each point in the point cloud signal, and sequencing the calculated local search radius of each point; and S3, clustering the point cloud signals based on the DBSCAN algorithm. Based on the traditional DBSCAN algorithm, the invention adds the physical attributes based on other sensors and each point for data preprocessing, determines the local search radius epsilon for each point cloud according to the physical quantity of each point cloud, sorts all original detections, clusters the original detections and dynamically judges whether the clustering result points are single-frame observation targets or not, and compares the clustering result of the invention with the traditional DBSCAN clustering mode to find that the clustering effect of the algorithm is better.
Description
Technical Field
The invention belongs to the technical field of intelligent driving of automobiles, and particularly relates to a DBSCAN-based 4D millimeter wave radar clustering method and a storage medium.
Background
In recent two years, multi-chip cascade is adopted, 4D millimeter wave radars with more antenna sending and receiving channels are widely used in the field of automatic driving, compared with common millimeter wave radars, although the detection performance is greatly improved and even comparable with laser radars, the detection principle and the data processing flow are different from those of the traditional radars, and detected point clouds are required to be subjected to clustering processing and then transmitted into a rear end to be subjected to algorithm operation such as Kalman filtering.
The angular resolution of the millimeter wave radar is fixed, so that the point cloud density reflected by the target is determined to be inversely proportional to the distance between the target and the point cloud, and when the target is close to the target, the point cloud density is high; when the target is far away, the point cloud density is low. The density distribution of the point cloud detected by the millimeter wave radar is uneven, and the point cloud at the far position is sparse, and the point cloud at the near position is dense. The traditional DBSCAN clustering algorithm has an unsatisfactory effect when the condition of uneven density distribution is processed. When choosing a smaller epsilon from a higher density set of points, the lower density ones will be split into many different clusters; when choosing a higher epsilon based on a lower density set of points, many clusters at higher densities will be merged into one cluster. Therefore, the traditional DBSCAN algorithm has poor effect under the clustering of radar point clouds with a distance, and missed detection and false detection of radar data can be caused no matter how global parameters are selected.
Chinese patent 201910661302.6 discloses a millimeter wave radar multi-target tracking method, which uses a millimeter wave radar to collect point cloud data, performs clustering processing on the point cloud data to distinguish echo signals of different targets, estimates observed state information of a plurality of targets according to a clustering result, and further realizes multi-target tracking. According to the characteristics of the millimeter wave radar, the DBSCAN clustering algorithm is improved, the target number and the target state estimation accuracy are improved, and prediction and tracking of a plurality of target tracks in a complex environment are realized by using Kalman filtering and data association algorithms. The algorithm from point cloud clustering to multi-target tracking is completed, and from the implementation details, the basic DBSCAN algorithm is used, namely the global epsilon is used, so that the clustering effect cannot be avoided to meet the detection characteristic of millimeter waves.
Chinese patent 202110697622.4 discloses a traffic target identification method based on DBSCAN algorithm, belonging to the technical field of data processing. Firstly, detecting a target to be detected by using a millimeter wave radar in a continuous time period to obtain different position information of the target to be detected; then, clustering the point cloud data by using a DBSCAN clustering algorithm by using the position information as point cloud data to obtain each cluster; and then, identifying and dividing the target types by using the number of scattering points in the clusters to obtain the target types corresponding to the clusters, counting the number of the target types, and finally completing the identification and counting of the traffic targets in the comprehensive traffic environment. The method improves the accuracy of target identification, and the identification process is simple and efficient. The method also uses a millimeter wave radar DBSCAN algorithm and completes the correlation tracking among multi-frame detection data to count the total number of the tracked targets, and the global parameters are still used, so that the clustering result still inevitably and inevitably accords with the measurement characteristics of the millimeter wave radar.
Disclosure of Invention
In order to solve the problems, the invention provides a DBSCAN-based 4D millimeter wave radar clustering method and a storage medium, so that the whole clustering algorithm is more in line with the characteristics of millimeter wave radars.
In order to solve the technical problem, the technical scheme adopted by the invention is as follows: A4D millimeter wave radar clustering method based on DBSCAN comprises the following steps,
s1, acquiring a vehicle signal, a lane line signal and a 4D millimeter wave radar point cloud signal, and analyzing and preprocessing the signals;
s2, calculating local search radiuses of each point in the point cloud signal, and sequencing the calculated local search radiuses of each point;
and S3, clustering the point cloud signals based on the DBSCAN algorithm.
And as optimization, the vehicle signal is obtained from the vehicle CAN signal and comprises the steering wheel angle, the course angular speed, the vehicle speed and the acceleration of the vehicle.
And optimally, the lane line signal is acquired from the front camera CAN signal and comprises the cubic curve coefficient and the length of the lane line.
And as optimization, the point cloud signal is obtained from an Ethernet signal of a 4D millimeter wave radar, wherein the attributes of each point in the point cloud signal comprise a polar diameter, a horizontal angle, a pitch angle, a relative radial velocity and a signal-to-noise ratio.
As an optimization, the point cloud signal is preprocessed, specifically,
s101, converting coordinate information of each point in the point cloud signal into a Cartesian coordinate system;
s102, calculating the absolute radial speed of each point;
s103, eliminating noise points in the point cloud signal according to preset filtering conditions.
As an optimization, the absolute radial velocity is calculated by projecting the vehicle velocity on the corresponding point angle.
As an optimization, step S2 includes,
s201, selecting variables for calculating the distance between two points, wherein the variables comprise x, y, z and vn;
s202, determining the weight coefficient of each variable, calculating the generalized Mahalanobis distance between two points as follows,
dist(p1，p2) 2 ＝0.8(x 1 -x 2 ) 2 +1.3(y 1 -y 2 ) 2 +0.1(z 1 -z 2 ) 2 +(vn 1 -vn 2 ) 2 (1)
wherein x is 1 、y 1 、z 1 Is the coordinate of point p1, x 2 、y 2 、z 2 Is the coordinate of point p2, vn is the absolute radial velocity;
s203, assigning a clustering factor f1 according to the absolute radial velocity of each point, calculating by adopting the following formula,
s204, assigning a clustering factor f2 according to the distance between each point, calculating by adopting the following formula,
s205, setting the basic local search radius ∈ to 1.65, calculating the final local search radius of each point, using the following formula,
ε i ＝ε*f1*f2 (4)
s206, sorting the final local search radiuses of all the points from large to small.
As optimization, in step S3, clustering is performed in a breadth-first search mode, the minimum number of point cloud clusters is adjusted according to a preset strategy, and after the clustering is completed, a clustering result is output.
And as optimization, the preset strategy comprises the steps of taking a larger value for a target point with a short radial distance and taking a smaller value for a target point with a long radial distance.
A storage medium storing one or more programs, which when executed by a processor, perform the steps of the DBSCAN-based 4D millimeter wave radar clustering method.
Compared with the prior art, the invention has the following advantages:
the invention respectively receives a vehicle signal, a lane line signal and a 4D millimeter wave radar point cloud signal which are output by front camera fusion from a vehicle CAN port and a vehicle Ethernet, analyzes and preprocesses the signals and then sends the signals into a point cloud fusion module; clustering the millimeter wave point cloud by using an improved DBSCAN algorithm, and fitting to generate a plurality of single-frame observation targets; the method comprises the following steps of screening and deleting some unimportant single-frame observation clustering targets by utilizing lane line information fused and output by a camera, so that the targets output by a radar are all important targets influencing an automatic driving function; and performing a series of tracking algorithms such as single-sensor multi-target association, Kalman filtering fusion, target creation, deletion, combination and the like on the generated single-frame target. The invention mainly solves the problem that the traditional DBSCAN algorithm is difficult to adapt to the uneven point cloud density of the intelligent driving 4D millimeter wave radar, the whole algorithm is based on the traditional DBSCAN algorithm, and steps of carrying out data preprocessing based on other sensors and the physical properties of each point, determining a local search radius epsilon for each point cloud according to the physical quantity of each point cloud, sequencing all original detections, clustering, dynamically judging whether the point of a clustering result is a single-frame observation target or not and the like are added.
Drawings
FIG. 1 is a diagram illustrating the searching radius of each point in the conventional DBSCAN;
FIG. 2 is a view of local search radii of each point of DBSCAN according to the present invention;
FIG. 3 is a block diagram of the clustering process of DBSCAN algorithm of the present invention;
FIG. 4 is a comparison graph of the clustering effect of the real driving 1 minute data statistics DBSCAN algorithm and the traditional DBSCAN algorithm.
Detailed Description
The invention will be further explained with reference to the drawings and the embodiments.
Example (b): see fig. 1-4.
In particular to a DBSCAN-based 4D millimeter wave radar clustering method, which comprises the following steps,
and S1, acquiring the vehicle signal, the lane line signal and the 4D millimeter wave radar point cloud signal, and analyzing and preprocessing the signals. Specifically, after hardware is connected to an industrial personal computer, a vehicle signal is received from a vehicle CAN signal, and the vehicle signal mainly comprises a vehicle steering wheel turning angle, a course angular velocity yaw rate, a vehicle speed and an acceleration which are used in the algorithm; the automatic driving system is a three-lane model, so that at most six lane lines exist, and cubic curve coefficients and lengths of the six lane lines are received; and receiving point cloud signals from the 4D millimeter wave radar Ethernet signals, wherein the number of the point clouds output most in a single frame is 1024 points. The information contained in each point comprises a polar diameter, a horizontal angle, a pitch angle, a relative radial velocity and a signal-to-noise ratio (SNR), and the received data are analyzed based on a corresponding signal protocol.
The point cloud signal is preprocessed, specifically,
s101, converting coordinate information of each point in the point cloud signal into a Cartesian coordinate system;
s102, calculating the absolute radial speed of each point; the absolute radial velocity is calculated by projecting the vehicle velocity at the corresponding point angle.
S103, eliminating noise points in the point cloud signal according to preset filtering conditions.
Specifically, the point cloud output by the front radar is sent in a polar coordinate mode, in order to match with a subsequent algorithm, the position coordinate information of all points in the point cloud is converted into a Cartesian coordinate from the polar coordinate, the radar only sends the relative radial speed of each point, and in order to achieve a better clustering effect, the clustering algorithm selects the absolute radial speed as a weight variable of clustering, and therefore the absolute radial speed of each point is calculated through projection of the speed of the vehicle on the angle of the corresponding point. In addition, since the ground is used as a reflection source to easily cause noise points of multipath reflection, the noise points are characterized in that the Z-axis position coordinate is a value which is less than zero and unreasonable, and therefore, the filtering conditions are set to eliminate the noise points, for example, the installation height of the radar is 0.5m away from the ground, and all detection points with the height less than-1.5 m are eliminated.
And S2, calculating the local search radius of each point in the point cloud signal, and sequencing the calculated local search radius of each point. Specifically, the flow of the clustering block diagram is shown in fig. 3, and first, preparation work before clustering needs to be performed, that is, parameters for calculating the distance between two points and weights between the parameters are selected. The method selects 4 parameters of x, y, z and vn (absolute radial velocity) as variables for calculating the distance, selects corresponding weights of the four variables for calculating the distance, and obtains the generalized Mahalanobis distance between two points after determining the weight occupied by each component through a large amount of data analysis.
S201, selecting variables for calculating the distance between two points, wherein the variables comprise x, y, z and vn;
s202, determining the weight coefficient of each variable, calculating the generalized Mahalanobis distance between two points, and calculating the following calculation,
dist(p1，p2) 2 ＝0.8(x 1 -x 2 ) 2 +1.3(y 1 -y 2 ) 2 +0.1(z 1 -z 2 ) 2 +(vn 1 -vn 2 ) 2 (1)
wherein x is 1 、y 1 、z 1 Is the coordinate of point p1, x 2 、y 2 、z 2 Is the coordinate of point p2, vn is the absolute radial velocity;
s203, for each detection point, assigning a unique clustering factor f1 to the detection point by referring to the absolute radial velocity of the detection point, wherein the judgment condition is that,
where abs (vn) represents an absolute value function.
S204, in addition, considering that the detection density of the radar is gradually sparse from near to far, which is contrary to the conventional DBSCAN algorithm, a clustering factor f2 is assigned to the distance, and the judgment condition is that,
where px denotes a pixel.
S205, setting the basic local search radius ∈ to 1.65, calculating the final local search radius of each point, using the following formula,
ε i ＝ε*f1*f2 (4)
s206, through the above operations, the local search radius determination of each point is completed, and through the step, the problem that the local details are not easily controlled by global variables in the traditional DBSCAN algorithm is improved, but after the search radius of each point is introduced to be different, the problem of inconsistent results for the same blob cloud clustering algorithm may be caused, as shown in fig. 1 and 2, assuming that there are two probe points A, B, the search radius of probe point a is 2, the search radius of probe point B is 1, and the generalized distance between two points is 1.5, when the point A is used as the core point for searching, the point B is in the clustering range of the point A, when B is used as the core point, point a is not in the search range of B, and if B is not available for further searching, the clustering result is not controlled. Therefore, the DBSCAN algorithm sequence of all points in the point cloud needs to be controlled, all point clouds are sorted once from large to small according to the search radius of the point clouds, and then the clustering algorithm is performed according to the sorted sequence, so that the clustering result of each time can be ensured to be consistent. After the above processing, the point cloud data can be subjected to a subsequent clustering algorithm.
And S3, clustering the point cloud signals based on the DBSCAN algorithm. And performing traditional DBSCAN algorithm clustering processing on the point cloud. Clustering is performed based on a breadth-first search mode, an operation flow chart is shown in figure 3, the flow is the traditional clustering processing of the DBSCAN algorithm, and redundant description is omitted. After a point cloud cluster is found, a traditional DBSCAN algorithm generally defines a global minimum number of the point cloud clusters, but the algorithm adjusts the minimum number of the point cloud clusters to be a non-fixed value based on information of the point cloud clusters, a basic preset strategy can be that the value of a target with a close radial distance is larger, the value of a target with a far radial distance is smaller, the adjusting strategy also accords with the detection characteristic of a front radar, all clustering processes are finished after the processing, and the result can be output to a rear end. The clustering result is compared with the real scene and the statistical index of the traditional DBSCAN algorithm is shown in the attached figure 4.
A storage medium storing one or more programs, which when executed by a processor, perform the steps of the DBSCAN-based 4D millimeter wave radar clustering method.
The invention respectively receives a vehicle signal, a lane line signal and a 4D millimeter wave radar point cloud signal which are output by front camera fusion from a vehicle CAN port and a vehicle Ethernet, analyzes and preprocesses the signals and then sends the signals into a point cloud fusion module; clustering the millimeter wave point cloud by using an improved DBSCAN algorithm, and fitting to generate a plurality of single-frame observation targets; the method comprises the following steps of screening and deleting some unimportant single-frame observation cluster targets by utilizing lane line information fused and output by a camera, so that targets output by a radar are guaranteed to be important targets influencing an automatic driving function; and performing a series of tracking algorithms such as single-sensor multi-target association, Kalman filtering fusion, target creation, deletion, combination and the like on the generated single-frame target. The invention mainly solves the problem that the traditional DBSCAN algorithm is difficult to adapt to the uneven point cloud density of the intelligent driving 4D millimeter wave radar, the whole algorithm is based on the traditional DBSCAN algorithm, and steps of carrying out data preprocessing based on other sensors and the physical properties of each point, determining a local search radius epsilon for each point cloud according to the physical quantity of each point cloud, sequencing all original detections, clustering, dynamically judging whether the point of a clustering result is a single-frame observation target or not and the like are added.
Finally, it should be noted that the above embodiments are only used for illustrating the technical solutions of the present invention and not for limiting the technical solutions, and those skilled in the art should understand that modifications or equivalent substitutions can be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions, and all the modifications and equivalent substitutions should be covered by the claims of the present invention.
Claims (10)
1. A4D millimeter wave radar clustering method based on DBSCAN is characterized by comprising the following steps,
s1, acquiring a vehicle signal, a lane line signal and a 4D millimeter wave radar point cloud signal, and analyzing and preprocessing the signals;
s2, calculating the local search radius of each point in the point cloud signal, and sequencing the calculated local search radius of each point;
and S3, clustering the point cloud signals based on the DBSCAN algorithm.
2. The DBSCAN-based 4D millimeter wave radar clustering method according to claim 1, wherein the vehicle signal is obtained from a vehicle CAN signal and comprises a steering wheel turning angle, a course angular velocity, a vehicle speed and an acceleration of the vehicle.
3. The DBSCAN-based 4D millimeter wave radar clustering method according to claim 1, wherein the lane line signal is obtained from a front camera CAN signal and comprises a cubic curve coefficient and a length of a lane line.
4. The DBSCAN-based 4D millimeter wave radar clustering method according to claim 1, wherein the point cloud signal is obtained from an Ethernet signal of a 4D millimeter wave radar, and attributes of each point in the point cloud signal include a polar diameter, a horizontal angle, a pitch angle, a relative radial velocity and a signal-to-noise ratio.
5. The DBSCAN-based 4D millimeter wave radar clustering method according to any one of claims 1 to 4, wherein the point cloud signal is preprocessed, specifically,
s101, converting coordinate information of each point in the point cloud signal into a Cartesian coordinate system;
s102, calculating the absolute radial speed of each point;
s103, eliminating noise points in the point cloud signal according to preset filtering conditions.
6. The DBSCAN-based 4D millimeter wave radar clustering method according to claim 5, wherein the absolute radial velocity is calculated by projection of the speed of the vehicle on the angle of the corresponding point.
7. The DBSCAN-based 4D millimeter wave radar clustering method according to claim 6, wherein the step S2 comprises,
s201, selecting variables for calculating the distance between two points, wherein the variables comprise x, y, z and vn;
s202, determining the weight coefficient of each variable, calculating the generalized Mahalanobis distance between two points as follows,
dist(p1，p2) 2 ＝0.8(x 1 -x 2 ) 2 +1.3(y 1 -y 2 ) 2 +0.1(z 1 -z 2 ) 2 +(vn 1 -vn 2 ) 2 (1)
wherein x is 1 、y 1 、z 1 Is the coordinate of point p1, x 2 、y 2 、z 2 Is the coordinate of point p2, vn is the absolute radial velocity;
s203, assigning a clustering factor f1 according to the absolute radial velocity of each point, calculating by adopting the following formula,
s204, assigning a clustering factor f2 according to the distance between each point, calculating by adopting the following formula,
s205, setting the basic local search radius ∈ to 1.65, calculating the final local search radius of each point, using the following formula,
ε i ＝ε*f1*f2 (4)
s206, sorting the final local search radiuses of all the points from large to small.
8. The DBSCAN-based 4D millimeter wave radar clustering method according to claim 7, wherein in step S3, clustering is performed in a breadth-first search-based manner, and the minimum number of point cloud clusters is adjusted according to a preset strategy, and after the clustering is completed, a clustering result is output.
9. The DBSCAN-based 4D millimeter wave radar clustering method according to claim 8, wherein the preset strategy includes taking a larger value for a target point with a close radial distance and taking a smaller value for a target point with a far radial distance.
10. A storage medium storing one or more programs which, when executed by a processor, perform the steps of the DBSCAN-based 4D millimeter wave radar clustering method according to any one of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210460110.0A CN115062683A (en) | 2022-04-24 | 2022-04-24 | 4D millimeter wave radar clustering method based on DBSCAN and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210460110.0A CN115062683A (en) | 2022-04-24 | 2022-04-24 | 4D millimeter wave radar clustering method based on DBSCAN and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115062683A true CN115062683A (en) | 2022-09-16 |
Family
ID=83197224
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210460110.0A Pending CN115062683A (en) | 2022-04-24 | 2022-04-24 | 4D millimeter wave radar clustering method based on DBSCAN and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115062683A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115439484A (en) * | 2022-11-10 | 2022-12-06 | 苏州挚途科技有限公司 | Detection method and device based on 4D point cloud, storage medium and processor |
| US20240288569A1 (en) * | 2023-02-20 | 2024-08-29 | Shanghai Geometrical Perception And Learning Co., Ltd. | Method of target feature extraction and multi-target tracking based on 4d millimeter wave radar |
| CN119048787A (en) * | 2024-11-01 | 2024-11-29 | 杭州岸达科技有限公司 | Multi-target point cloud clustering method |
-
2022
- 2022-04-24 CN CN202210460110.0A patent/CN115062683A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115439484A (en) * | 2022-11-10 | 2022-12-06 | 苏州挚途科技有限公司 | Detection method and device based on 4D point cloud, storage medium and processor |
| US20240288569A1 (en) * | 2023-02-20 | 2024-08-29 | Shanghai Geometrical Perception And Learning Co., Ltd. | Method of target feature extraction and multi-target tracking based on 4d millimeter wave radar |
| US12140658B2 (en) * | 2023-02-20 | 2024-11-12 | Shanghai Geometrical Perception And Learning Co., Ltd. | Method of target feature extraction and multi-target tracking based on 4D millimeter wave radar |
| CN119048787A (en) * | 2024-11-01 | 2024-11-29 | 杭州岸达科技有限公司 | Multi-target point cloud clustering method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115062683A (en) | 4D millimeter wave radar clustering method based on DBSCAN and storage medium | |
| CN111260683B (en) | Target detection and tracking method and device for three-dimensional point cloud data | |
| CN112526513B (en) | Millimeter wave radar environment map construction method and device based on clustering algorithm | |
| US11313950B2 (en) | Machine learning based highway radar vehicle classification across multiple lanes and speeds | |
| CN113866742B (en) | Method for point cloud processing and target classification of 4D millimeter wave radar | |
| CN113536850B (en) | Target object size testing method and device based on 77G millimeter wave radar | |
| CN113671481B (en) | 3D multi-target tracking processing method based on millimeter wave radar | |
| CN115236627B (en) | Millimeter wave radar data clustering method based on multi-frame Doppler velocity dimension expansion | |
| CN112731307B (en) | RATM-CFAR detector based on distance-angle joint estimation and detection method | |
| CN114460582B (en) | Millimeter wave radar cart identification method based on point cloud speed | |
| CN114879192A (en) | Decision tree vehicle type classification method based on road side millimeter wave radar and electronic equipment | |
| Wang et al. | DBSCAN clustering algorithm of millimeter wave radar based on multi frame joint | |
| CN111736167B (en) | Method and device for obtaining laser point cloud density | |
| CN115203888A (en) | Method, system and electronic equipment for target association in multi-target tracking | |
| CN115187955A (en) | A target localization method based on post-fusion of visual image and sparse point cloud | |
| CN114609609A (en) | Velocity Estimation Method for Extracting Stationary Point Cloud by Random Sampling of FMCW Lidar | |
| CN114839615B (en) | 4D millimeter wave radar target course angle fitting method and storage medium | |
| CN115980738B (en) | Multi-target tracking method based on dynamic and static separation | |
| CN108983194A (en) | A kind of Objective extraction and condensing method based on ground surveillance radar system | |
| CN116184344A (en) | Self-adaptive vehicle-mounted millimeter wave radar DBSCAN clustering method and device | |
| CN116883707A (en) | Millimeter wave radar point cloud clustering method | |
| CN116935356A (en) | Weak supervision-based automatic driving multi-mode picture and point cloud instance segmentation method | |
| CN111768442B (en) | A Track Initiation Method and System Based on Hierarchical Clustering and Logical Method | |
| CN114137523A (en) | Method and system for noise elimination of original point cloud of vehicle millimeter-wave radar | |
| CN117953201A (en) | False target removing method for millimeter wave Lei Dadian cloud |
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 |